123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473147414751476147714781479148014811482148314841485148614871488148914901491149214931494149514961497149814991500150115021503150415051506150715081509151015111512151315141515151615171518151915201521152215231524152515261527152815291530153115321533153415351536153715381539154015411542154315441545154615471548154915501551155215531554155515561557155815591560156115621563156415651566156715681569157015711572157315741575157615771578157915801581158215831584158515861587158815891590159115921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650165116521653165416551656165716581659166016611662166316641665166616671668166916701671167216731674167516761677167816791680168116821683168416851686168716881689169016911692169316941695169616971698169917001701170217031704170517061707170817091710171117121713171417151716171717181719172017211722172317241725172617271728172917301731173217331734173517361737173817391740174117421743174417451746174717481749175017511752175317541755175617571758175917601761176217631764176517661767176817691770177117721773177417751776177717781779178017811782178317841785178617871788178917901791179217931794179517961797179817991800180118021803180418051806180718081809181018111812181318141815181618171818181918201821182218231824182518261827182818291830183118321833183418351836183718381839184018411842184318441845184618471848184918501851185218531854185518561857185818591860186118621863186418651866186718681869187018711872187318741875187618771878187918801881188218831884188518861887188818891890189118921893189418951896189718981899190019011902190319041905190619071908190919101911191219131914191519161917191819191920192119221923192419251926192719281929193019311932193319341935193619371938193919401941194219431944194519461947194819491950195119521953195419551956195719581959196019611962196319641965196619671968196919701971197219731974197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010201120122013201420152016201720182019202020212022202320242025202620272028202920302031203220332034203520362037203820392040204120422043204420452046204720482049205020512052205320542055205620572058205920602061206220632064206520662067206820692070207120722073207420752076207720782079208020812082208320842085208620872088208920902091209220932094209520962097209820992100210121022103210421052106210721082109211021112112211321142115211621172118211921202121212221232124212521262127212821292130213121322133213421352136213721382139214021412142214321442145214621472148214921502151215221532154215521562157215821592160216121622163216421652166216721682169217021712172217321742175217621772178217921802181218221832184218521862187218821892190219121922193219421952196219721982199220022012202220322042205220622072208220922102211221222132214221522162217221822192220222122222223222422252226222722282229223022312232223322342235223622372238223922402241224222432244224522462247224822492250225122522253225422552256225722582259226022612262226322642265226622672268226922702271227222732274227522762277227822792280228122822283228422852286228722882289229022912292229322942295229622972298229923002301230223032304230523062307230823092310231123122313231423152316231723182319232023212322232323242325232623272328232923302331233223332334233523362337233823392340234123422343234423452346234723482349235023512352235323542355235623572358235923602361236223632364236523662367236823692370237123722373237423752376237723782379238023812382238323842385238623872388238923902391239223932394239523962397239823992400240124022403240424052406240724082409241024112412241324142415241624172418241924202421242224232424242524262427242824292430243124322433243424352436243724382439244024412442244324442445244624472448244924502451245224532454245524562457245824592460246124622463246424652466246724682469247024712472247324742475247624772478247924802481248224832484248524862487248824892490249124922493249424952496249724982499250025012502250325042505250625072508250925102511251225132514251525162517251825192520252125222523252425252526252725282529253025312532253325342535253625372538253925402541254225432544254525462547254825492550255125522553255425552556255725582559256025612562256325642565256625672568256925702571257225732574257525762577257825792580258125822583258425852586258725882589259025912592259325942595259625972598259926002601260226032604260526062607260826092610261126122613261426152616261726182619262026212622262326242625262626272628262926302631263226332634263526362637263826392640264126422643264426452646264726482649265026512652265326542655265626572658265926602661266226632664266526662667266826692670267126722673267426752676267726782679268026812682268326842685268626872688268926902691269226932694269526962697269826992700270127022703270427052706270727082709271027112712271327142715271627172718271927202721272227232724272527262727272827292730273127322733273427352736273727382739274027412742274327442745274627472748274927502751275227532754275527562757275827592760276127622763276427652766276727682769277027712772277327742775277627772778277927802781278227832784278527862787278827892790279127922793279427952796279727982799280028012802280328042805280628072808280928102811281228132814281528162817281828192820282128222823282428252826282728282829283028312832283328342835283628372838283928402841284228432844284528462847284828492850285128522853285428552856285728582859286028612862286328642865286628672868286928702871287228732874287528762877287828792880288128822883288428852886288728882889289028912892289328942895289628972898289929002901290229032904290529062907290829092910291129122913291429152916291729182919292029212922292329242925292629272928292929302931293229332934293529362937293829392940294129422943294429452946294729482949295029512952295329542955295629572958295929602961296229632964296529662967296829692970297129722973297429752976297729782979298029812982298329842985298629872988298929902991299229932994299529962997299829993000300130023003300430053006300730083009301030113012301330143015301630173018301930203021302230233024302530263027302830293030303130323033303430353036303730383039304030413042304330443045304630473048304930503051305230533054305530563057305830593060306130623063306430653066306730683069307030713072307330743075307630773078307930803081308230833084308530863087308830893090309130923093309430953096309730983099310031013102310331043105310631073108310931103111311231133114311531163117311831193120312131223123312431253126312731283129313031313132313331343135313631373138313931403141314231433144314531463147314831493150315131523153315431553156315731583159316031613162316331643165316631673168316931703171317231733174317531763177317831793180318131823183318431853186318731883189319031913192319331943195319631973198319932003201320232033204320532063207320832093210321132123213321432153216321732183219322032213222322332243225322632273228322932303231323232333234323532363237323832393240324132423243324432453246324732483249325032513252325332543255325632573258325932603261326232633264326532663267326832693270327132723273327432753276327732783279328032813282328332843285328632873288328932903291329232933294329532963297329832993300330133023303330433053306330733083309331033113312331333143315331633173318331933203321332233233324332533263327332833293330333133323333333433353336333733383339334033413342334333443345334633473348334933503351335233533354335533563357335833593360336133623363336433653366336733683369337033713372337333743375337633773378337933803381338233833384338533863387338833893390339133923393339433953396339733983399340034013402340334043405340634073408340934103411341234133414341534163417341834193420342134223423342434253426342734283429343034313432343334343435343634373438343934403441344234433444344534463447344834493450345134523453345434553456345734583459346034613462346334643465346634673468346934703471347234733474347534763477347834793480348134823483348434853486348734883489349034913492349334943495349634973498349935003501350235033504350535063507350835093510351135123513351435153516351735183519352035213522352335243525352635273528352935303531353235333534353535363537353835393540354135423543354435453546354735483549355035513552355335543555355635573558355935603561356235633564356535663567356835693570357135723573357435753576357735783579358035813582358335843585358635873588358935903591359235933594359535963597359835993600360136023603360436053606360736083609361036113612361336143615361636173618361936203621362236233624362536263627362836293630363136323633363436353636363736383639364036413642364336443645364636473648364936503651365236533654365536563657365836593660366136623663366436653666366736683669367036713672367336743675367636773678367936803681368236833684368536863687368836893690369136923693369436953696369736983699370037013702370337043705370637073708370937103711371237133714371537163717371837193720372137223723372437253726372737283729373037313732373337343735373637373738373937403741374237433744374537463747374837493750375137523753375437553756375737583759376037613762376337643765376637673768376937703771377237733774377537763777377837793780378137823783378437853786378737883789379037913792379337943795379637973798379938003801380238033804380538063807380838093810381138123813381438153816381738183819382038213822382338243825382638273828382938303831383238333834383538363837383838393840384138423843384438453846384738483849385038513852385338543855385638573858385938603861386238633864386538663867386838693870387138723873387438753876387738783879388038813882388338843885388638873888388938903891389238933894389538963897389838993900390139023903390439053906390739083909391039113912391339143915391639173918391939203921392239233924392539263927392839293930393139323933393439353936393739383939394039413942394339443945394639473948394939503951395239533954395539563957395839593960396139623963396439653966396739683969397039713972397339743975397639773978397939803981398239833984398539863987398839893990399139923993399439953996399739983999400040014002400340044005400640074008400940104011401240134014401540164017401840194020402140224023402440254026402740284029403040314032403340344035403640374038403940404041404240434044404540464047404840494050405140524053405440554056405740584059406040614062406340644065406640674068406940704071407240734074407540764077407840794080408140824083408440854086408740884089409040914092409340944095409640974098409941004101410241034104410541064107410841094110411141124113411441154116411741184119412041214122412341244125412641274128412941304131413241334134413541364137413841394140414141424143414441454146414741484149415041514152415341544155415641574158415941604161416241634164416541664167416841694170417141724173417441754176417741784179418041814182418341844185418641874188418941904191419241934194419541964197419841994200420142024203420442054206420742084209421042114212421342144215421642174218421942204221422242234224422542264227422842294230423142324233423442354236423742384239424042414242424342444245424642474248424942504251425242534254425542564257425842594260426142624263426442654266426742684269427042714272427342744275427642774278427942804281428242834284428542864287428842894290429142924293429442954296429742984299430043014302430343044305430643074308430943104311431243134314431543164317431843194320432143224323432443254326432743284329433043314332433343344335433643374338433943404341434243434344434543464347434843494350435143524353435443554356435743584359436043614362436343644365436643674368436943704371437243734374437543764377437843794380438143824383438443854386438743884389439043914392439343944395439643974398439944004401440244034404440544064407440844094410441144124413441444154416441744184419442044214422442344244425442644274428442944304431443244334434443544364437443844394440444144424443444444454446444744484449445044514452445344544455445644574458445944604461446244634464446544664467446844694470447144724473447444754476447744784479448044814482448344844485448644874488448944904491449244934494449544964497449844994500450145024503450445054506450745084509451045114512451345144515451645174518451945204521452245234524452545264527452845294530453145324533453445354536453745384539454045414542454345444545454645474548454945504551455245534554455545564557455845594560456145624563456445654566456745684569457045714572457345744575457645774578457945804581458245834584458545864587458845894590459145924593459445954596459745984599460046014602460346044605460646074608460946104611461246134614461546164617461846194620462146224623462446254626462746284629463046314632463346344635463646374638463946404641464246434644464546464647464846494650465146524653465446554656465746584659466046614662466346644665466646674668466946704671467246734674467546764677467846794680468146824683468446854686468746884689469046914692469346944695469646974698469947004701470247034704470547064707470847094710471147124713471447154716471747184719472047214722472347244725472647274728472947304731473247334734473547364737473847394740474147424743474447454746474747484749475047514752475347544755475647574758475947604761476247634764476547664767476847694770477147724773477447754776477747784779478047814782478347844785478647874788478947904791479247934794479547964797479847994800480148024803480448054806480748084809481048114812481348144815481648174818481948204821482248234824482548264827482848294830483148324833483448354836483748384839484048414842484348444845484648474848484948504851485248534854485548564857485848594860486148624863486448654866486748684869487048714872487348744875487648774878487948804881488248834884488548864887488848894890489148924893489448954896489748984899490049014902490349044905490649074908490949104911491249134914491549164917491849194920492149224923492449254926492749284929493049314932493349344935493649374938493949404941494249434944494549464947494849494950495149524953495449554956495749584959496049614962496349644965496649674968496949704971497249734974497549764977497849794980498149824983498449854986498749884989499049914992499349944995499649974998499950005001500250035004500550065007500850095010501150125013501450155016501750185019502050215022502350245025502650275028502950305031503250335034503550365037503850395040504150425043504450455046504750485049505050515052505350545055505650575058505950605061506250635064506550665067506850695070507150725073507450755076507750785079508050815082508350845085508650875088508950905091509250935094509550965097509850995100510151025103510451055106510751085109511051115112511351145115511651175118511951205121512251235124512551265127512851295130513151325133513451355136513751385139514051415142514351445145514651475148514951505151515251535154515551565157515851595160516151625163516451655166516751685169517051715172517351745175517651775178517951805181518251835184518551865187518851895190519151925193519451955196519751985199520052015202520352045205520652075208520952105211521252135214521552165217521852195220522152225223522452255226522752285229523052315232523352345235523652375238523952405241524252435244524552465247524852495250525152525253525452555256525752585259526052615262526352645265526652675268526952705271527252735274527552765277527852795280528152825283528452855286528752885289529052915292529352945295529652975298529953005301530253035304530553065307530853095310531153125313531453155316531753185319532053215322532353245325532653275328532953305331533253335334533553365337533853395340534153425343534453455346534753485349535053515352535353545355535653575358535953605361536253635364536553665367536853695370537153725373537453755376537753785379538053815382538353845385538653875388538953905391539253935394539553965397539853995400540154025403540454055406540754085409541054115412541354145415541654175418541954205421542254235424542554265427542854295430543154325433543454355436543754385439544054415442544354445445544654475448544954505451545254535454545554565457545854595460546154625463546454655466546754685469547054715472547354745475547654775478547954805481548254835484548554865487548854895490549154925493549454955496549754985499550055015502550355045505550655075508550955105511551255135514551555165517551855195520552155225523552455255526552755285529553055315532553355345535553655375538553955405541554255435544554555465547554855495550555155525553555455555556555755585559556055615562556355645565556655675568556955705571557255735574557555765577557855795580558155825583558455855586558755885589559055915592559355945595559655975598559956005601560256035604560556065607560856095610561156125613561456155616561756185619562056215622562356245625562656275628562956305631563256335634563556365637563856395640564156425643564456455646564756485649565056515652565356545655565656575658565956605661566256635664566556665667566856695670567156725673567456755676567756785679568056815682568356845685568656875688568956905691569256935694569556965697569856995700570157025703570457055706570757085709571057115712571357145715571657175718571957205721572257235724572557265727572857295730573157325733573457355736573757385739574057415742574357445745574657475748574957505751575257535754575557565757575857595760576157625763576457655766576757685769577057715772577357745775577657775778577957805781578257835784578557865787578857895790579157925793579457955796579757985799580058015802580358045805580658075808580958105811581258135814581558165817581858195820582158225823582458255826582758285829583058315832583358345835583658375838583958405841584258435844584558465847584858495850585158525853585458555856585758585859586058615862586358645865586658675868586958705871587258735874587558765877587858795880588158825883588458855886588758885889589058915892589358945895589658975898589959005901590259035904590559065907590859095910591159125913591459155916591759185919592059215922592359245925592659275928592959305931593259335934593559365937593859395940594159425943594459455946594759485949595059515952595359545955595659575958595959605961596259635964596559665967596859695970597159725973597459755976597759785979598059815982598359845985598659875988598959905991599259935994599559965997599859996000600160026003600460056006600760086009601060116012601360146015601660176018601960206021602260236024602560266027602860296030603160326033603460356036603760386039604060416042604360446045604660476048604960506051605260536054605560566057605860596060606160626063606460656066606760686069607060716072607360746075607660776078607960806081608260836084608560866087608860896090609160926093609460956096609760986099610061016102610361046105610661076108610961106111611261136114611561166117611861196120612161226123612461256126612761286129613061316132613361346135613661376138613961406141614261436144614561466147614861496150615161526153615461556156615761586159616061616162616361646165616661676168616961706171617261736174617561766177617861796180618161826183618461856186618761886189619061916192619361946195619661976198619962006201620262036204620562066207620862096210621162126213621462156216621762186219622062216222622362246225622662276228622962306231623262336234623562366237623862396240624162426243624462456246624762486249625062516252625362546255625662576258625962606261626262636264626562666267626862696270627162726273627462756276627762786279628062816282628362846285628662876288628962906291629262936294629562966297629862996300630163026303630463056306630763086309631063116312631363146315631663176318631963206321632263236324632563266327632863296330633163326333633463356336633763386339634063416342634363446345634663476348634963506351635263536354635563566357635863596360636163626363636463656366636763686369637063716372637363746375637663776378637963806381638263836384638563866387638863896390639163926393639463956396639763986399640064016402640364046405640664076408640964106411641264136414641564166417641864196420642164226423642464256426642764286429643064316432643364346435643664376438643964406441644264436444644564466447644864496450645164526453645464556456645764586459646064616462646364646465646664676468646964706471647264736474647564766477647864796480648164826483648464856486648764886489649064916492649364946495649664976498649965006501650265036504650565066507650865096510651165126513651465156516651765186519652065216522652365246525652665276528652965306531653265336534653565366537653865396540654165426543654465456546654765486549655065516552655365546555655665576558655965606561656265636564656565666567656865696570657165726573657465756576657765786579658065816582658365846585658665876588658965906591659265936594659565966597659865996600660166026603660466056606660766086609661066116612661366146615661666176618661966206621662266236624662566266627662866296630663166326633663466356636663766386639664066416642664366446645664666476648664966506651665266536654665566566657665866596660666166626663666466656666666766686669667066716672667366746675667666776678667966806681668266836684668566866687668866896690669166926693669466956696669766986699670067016702670367046705670667076708670967106711671267136714671567166717671867196720672167226723672467256726672767286729673067316732673367346735673667376738673967406741674267436744674567466747674867496750675167526753675467556756675767586759676067616762676367646765676667676768676967706771677267736774677567766777677867796780678167826783678467856786678767886789679067916792679367946795679667976798679968006801680268036804680568066807680868096810681168126813681468156816681768186819682068216822682368246825682668276828682968306831683268336834683568366837683868396840684168426843684468456846684768486849685068516852685368546855685668576858685968606861686268636864686568666867686868696870687168726873687468756876687768786879688068816882688368846885688668876888688968906891689268936894689568966897689868996900690169026903690469056906690769086909691069116912691369146915691669176918691969206921692269236924692569266927692869296930693169326933693469356936693769386939694069416942694369446945694669476948694969506951695269536954695569566957695869596960696169626963696469656966696769686969697069716972697369746975697669776978697969806981698269836984698569866987698869896990699169926993699469956996699769986999700070017002700370047005700670077008700970107011701270137014701570167017701870197020702170227023702470257026702770287029703070317032703370347035703670377038703970407041704270437044704570467047704870497050705170527053705470557056705770587059706070617062706370647065706670677068706970707071707270737074707570767077707870797080708170827083708470857086708770887089709070917092709370947095709670977098709971007101710271037104710571067107710871097110711171127113711471157116711771187119712071217122712371247125712671277128712971307131713271337134713571367137713871397140714171427143714471457146714771487149715071517152715371547155715671577158715971607161716271637164716571667167716871697170717171727173717471757176717771787179718071817182718371847185718671877188718971907191719271937194719571967197719871997200720172027203720472057206720772087209721072117212721372147215721672177218721972207221722272237224722572267227722872297230723172327233723472357236723772387239724072417242724372447245724672477248724972507251725272537254725572567257725872597260726172627263726472657266726772687269727072717272727372747275727672777278727972807281728272837284728572867287728872897290729172927293729472957296729772987299730073017302730373047305730673077308730973107311731273137314731573167317731873197320732173227323732473257326732773287329733073317332733373347335733673377338733973407341734273437344734573467347734873497350735173527353735473557356735773587359736073617362736373647365736673677368736973707371737273737374737573767377737873797380738173827383738473857386738773887389739073917392739373947395739673977398739974007401740274037404740574067407740874097410741174127413741474157416741774187419742074217422742374247425742674277428742974307431743274337434743574367437743874397440744174427443744474457446744774487449745074517452745374547455745674577458745974607461746274637464746574667467746874697470747174727473747474757476747774787479748074817482748374847485748674877488748974907491749274937494749574967497749874997500750175027503750475057506750775087509751075117512751375147515751675177518751975207521752275237524752575267527752875297530753175327533753475357536753775387539754075417542754375447545754675477548754975507551755275537554755575567557755875597560756175627563756475657566756775687569757075717572757375747575757675777578757975807581758275837584758575867587758875897590759175927593759475957596759775987599760076017602760376047605760676077608760976107611761276137614761576167617761876197620762176227623762476257626762776287629763076317632763376347635763676377638763976407641764276437644764576467647764876497650765176527653765476557656765776587659766076617662766376647665766676677668766976707671767276737674767576767677767876797680768176827683768476857686768776887689769076917692769376947695769676977698769977007701770277037704770577067707770877097710771177127713771477157716771777187719772077217722772377247725772677277728772977307731773277337734773577367737773877397740774177427743774477457746774777487749775077517752775377547755775677577758775977607761776277637764776577667767776877697770777177727773777477757776777777787779778077817782778377847785778677877788778977907791779277937794779577967797779877997800780178027803780478057806780778087809781078117812781378147815781678177818781978207821782278237824782578267827782878297830783178327833783478357836783778387839784078417842784378447845784678477848784978507851785278537854785578567857785878597860786178627863786478657866786778687869787078717872787378747875787678777878787978807881788278837884788578867887788878897890789178927893789478957896789778987899790079017902790379047905790679077908790979107911791279137914791579167917791879197920792179227923792479257926792779287929793079317932793379347935793679377938793979407941794279437944794579467947794879497950795179527953795479557956795779587959796079617962796379647965796679677968796979707971797279737974797579767977797879797980798179827983798479857986798779887989799079917992799379947995799679977998799980008001800280038004800580068007800880098010801180128013801480158016801780188019802080218022802380248025802680278028802980308031803280338034803580368037803880398040804180428043804480458046804780488049805080518052805380548055805680578058805980608061806280638064806580668067806880698070807180728073807480758076807780788079808080818082808380848085808680878088808980908091809280938094809580968097809880998100810181028103810481058106810781088109811081118112811381148115811681178118811981208121812281238124812581268127812881298130813181328133813481358136813781388139814081418142814381448145814681478148814981508151815281538154815581568157815881598160816181628163816481658166816781688169817081718172817381748175817681778178817981808181818281838184818581868187818881898190819181928193819481958196819781988199820082018202820382048205820682078208820982108211821282138214821582168217821882198220822182228223822482258226822782288229823082318232823382348235823682378238823982408241824282438244824582468247824882498250825182528253825482558256825782588259826082618262826382648265826682678268826982708271827282738274827582768277827882798280828182828283828482858286828782888289829082918292829382948295829682978298829983008301830283038304830583068307830883098310831183128313831483158316831783188319832083218322832383248325832683278328832983308331833283338334833583368337833883398340834183428343834483458346834783488349835083518352835383548355835683578358835983608361836283638364836583668367836883698370837183728373837483758376837783788379838083818382838383848385838683878388838983908391839283938394839583968397839883998400840184028403840484058406840784088409841084118412841384148415841684178418841984208421842284238424842584268427842884298430843184328433843484358436843784388439844084418442844384448445844684478448844984508451845284538454845584568457845884598460846184628463846484658466846784688469847084718472847384748475847684778478847984808481848284838484848584868487848884898490849184928493849484958496849784988499850085018502850385048505850685078508850985108511851285138514851585168517851885198520852185228523852485258526852785288529853085318532853385348535853685378538853985408541854285438544854585468547854885498550855185528553855485558556855785588559856085618562856385648565856685678568856985708571857285738574857585768577857885798580858185828583858485858586858785888589859085918592859385948595859685978598859986008601860286038604860586068607860886098610861186128613861486158616861786188619862086218622862386248625862686278628862986308631863286338634863586368637863886398640864186428643864486458646864786488649865086518652865386548655865686578658865986608661866286638664866586668667866886698670867186728673867486758676867786788679868086818682868386848685868686878688868986908691869286938694869586968697869886998700870187028703870487058706870787088709871087118712871387148715871687178718871987208721872287238724872587268727872887298730873187328733873487358736873787388739874087418742874387448745874687478748874987508751875287538754875587568757875887598760876187628763876487658766876787688769877087718772877387748775877687778778877987808781878287838784878587868787878887898790879187928793879487958796879787988799880088018802880388048805880688078808880988108811881288138814881588168817881888198820882188228823882488258826882788288829883088318832883388348835883688378838883988408841884288438844884588468847884888498850885188528853885488558856885788588859886088618862886388648865886688678868886988708871887288738874887588768877887888798880888188828883888488858886888788888889889088918892889388948895889688978898889989008901890289038904890589068907890889098910891189128913891489158916891789188919892089218922892389248925892689278928892989308931893289338934893589368937893889398940894189428943894489458946894789488949895089518952895389548955895689578958895989608961896289638964896589668967896889698970897189728973897489758976897789788979898089818982898389848985898689878988898989908991899289938994899589968997899889999000900190029003900490059006900790089009901090119012901390149015901690179018901990209021902290239024902590269027902890299030903190329033903490359036903790389039904090419042904390449045904690479048904990509051905290539054905590569057905890599060906190629063906490659066906790689069907090719072907390749075907690779078907990809081908290839084908590869087908890899090909190929093909490959096909790989099910091019102910391049105910691079108910991109111911291139114911591169117911891199120912191229123912491259126912791289129913091319132913391349135913691379138913991409141914291439144914591469147914891499150915191529153915491559156915791589159916091619162916391649165916691679168916991709171917291739174917591769177917891799180918191829183918491859186918791889189919091919192919391949195919691979198919992009201920292039204920592069207920892099210921192129213921492159216921792189219922092219222922392249225922692279228922992309231923292339234923592369237923892399240924192429243924492459246924792489249925092519252925392549255925692579258925992609261926292639264926592669267926892699270927192729273927492759276927792789279928092819282928392849285928692879288928992909291929292939294929592969297929892999300930193029303930493059306930793089309931093119312931393149315931693179318931993209321932293239324932593269327932893299330933193329333933493359336933793389339934093419342934393449345934693479348934993509351935293539354935593569357935893599360936193629363936493659366936793689369937093719372937393749375937693779378937993809381938293839384938593869387938893899390939193929393939493959396939793989399940094019402940394049405940694079408940994109411941294139414941594169417941894199420942194229423942494259426942794289429943094319432943394349435943694379438943994409441944294439444944594469447944894499450945194529453945494559456945794589459946094619462946394649465946694679468946994709471947294739474947594769477947894799480948194829483948494859486948794889489949094919492949394949495949694979498949995009501950295039504950595069507950895099510951195129513951495159516951795189519952095219522952395249525952695279528952995309531953295339534953595369537953895399540954195429543954495459546954795489549955095519552955395549555955695579558955995609561956295639564956595669567956895699570957195729573957495759576957795789579958095819582958395849585958695879588958995909591959295939594959595969597959895999600960196029603960496059606960796089609961096119612961396149615961696179618961996209621962296239624962596269627962896299630963196329633963496359636963796389639964096419642964396449645964696479648964996509651965296539654965596569657965896599660966196629663966496659666966796689669967096719672967396749675967696779678967996809681968296839684968596869687968896899690969196929693969496959696969796989699970097019702970397049705970697079708970997109711971297139714971597169717971897199720972197229723972497259726972797289729973097319732973397349735973697379738973997409741974297439744974597469747974897499750975197529753975497559756975797589759976097619762976397649765976697679768976997709771977297739774977597769777977897799780978197829783978497859786978797889789979097919792979397949795979697979798979998009801980298039804980598069807980898099810981198129813981498159816981798189819982098219822982398249825982698279828982998309831983298339834983598369837983898399840984198429843984498459846984798489849985098519852985398549855985698579858985998609861986298639864986598669867986898699870987198729873987498759876987798789879988098819882988398849885988698879888988998909891989298939894989598969897989898999900990199029903990499059906990799089909991099119912991399149915991699179918991999209921992299239924992599269927992899299930993199329933993499359936993799389939994099419942994399449945994699479948994999509951995299539954995599569957995899599960996199629963996499659966996799689969997099719972997399749975997699779978997999809981998299839984998599869987998899899990999199929993999499959996999799989999100001000110002100031000410005100061000710008100091001010011100121001310014100151001610017100181001910020100211002210023100241002510026100271002810029100301003110032100331003410035100361003710038100391004010041100421004310044100451004610047100481004910050100511005210053100541005510056100571005810059100601006110062100631006410065100661006710068100691007010071100721007310074100751007610077100781007910080100811008210083100841008510086100871008810089100901009110092100931009410095100961009710098100991010010101101021010310104101051010610107101081010910110101111011210113101141011510116101171011810119101201012110122101231012410125101261012710128101291013010131101321013310134101351013610137101381013910140101411014210143101441014510146101471014810149101501015110152101531015410155101561015710158101591016010161101621016310164101651016610167101681016910170101711017210173101741017510176101771017810179101801018110182101831018410185101861018710188101891019010191101921019310194101951019610197101981019910200102011020210203102041020510206102071020810209102101021110212102131021410215102161021710218102191022010221102221022310224102251022610227102281022910230102311023210233102341023510236102371023810239102401024110242102431024410245102461024710248102491025010251102521025310254102551025610257102581025910260102611026210263102641026510266102671026810269102701027110272102731027410275102761027710278102791028010281102821028310284102851028610287102881028910290102911029210293102941029510296102971029810299103001030110302103031030410305103061030710308103091031010311103121031310314103151031610317103181031910320103211032210323103241032510326103271032810329103301033110332103331033410335103361033710338103391034010341103421034310344103451034610347103481034910350103511035210353103541035510356103571035810359103601036110362103631036410365103661036710368103691037010371103721037310374103751037610377103781037910380103811038210383103841038510386103871038810389103901039110392103931039410395103961039710398103991040010401104021040310404104051040610407104081040910410104111041210413104141041510416104171041810419104201042110422104231042410425104261042710428104291043010431104321043310434104351043610437104381043910440104411044210443104441044510446104471044810449104501045110452104531045410455104561045710458104591046010461104621046310464104651046610467104681046910470104711047210473104741047510476104771047810479104801048110482104831048410485104861048710488104891049010491104921049310494104951049610497104981049910500105011050210503105041050510506105071050810509105101051110512105131051410515105161051710518105191052010521105221052310524105251052610527105281052910530105311053210533105341053510536105371053810539105401054110542105431054410545105461054710548105491055010551105521055310554105551055610557105581055910560105611056210563105641056510566105671056810569105701057110572105731057410575105761057710578105791058010581105821058310584105851058610587105881058910590105911059210593105941059510596105971059810599106001060110602106031060410605106061060710608106091061010611106121061310614106151061610617106181061910620106211062210623106241062510626106271062810629106301063110632106331063410635106361063710638106391064010641106421064310644106451064610647106481064910650106511065210653106541065510656106571065810659106601066110662106631066410665106661066710668106691067010671106721067310674106751067610677106781067910680106811068210683106841068510686106871068810689106901069110692106931069410695106961069710698106991070010701107021070310704107051070610707107081070910710107111071210713107141071510716107171071810719107201072110722107231072410725107261072710728107291073010731107321073310734107351073610737107381073910740107411074210743107441074510746107471074810749107501075110752107531075410755107561075710758107591076010761107621076310764107651076610767107681076910770107711077210773107741077510776107771077810779107801078110782107831078410785107861078710788107891079010791107921079310794107951079610797107981079910800108011080210803108041080510806108071080810809108101081110812108131081410815108161081710818108191082010821108221082310824108251082610827108281082910830108311083210833108341083510836108371083810839108401084110842108431084410845108461084710848108491085010851108521085310854108551085610857108581085910860108611086210863108641086510866108671086810869108701087110872108731087410875108761087710878108791088010881108821088310884108851088610887108881088910890108911089210893108941089510896108971089810899109001090110902109031090410905109061090710908109091091010911109121091310914109151091610917109181091910920109211092210923109241092510926109271092810929109301093110932109331093410935109361093710938109391094010941109421094310944109451094610947109481094910950109511095210953109541095510956109571095810959109601096110962109631096410965109661096710968109691097010971109721097310974109751097610977109781097910980109811098210983109841098510986109871098810989109901099110992109931099410995109961099710998109991100011001110021100311004110051100611007110081100911010110111101211013110141101511016110171101811019110201102111022110231102411025110261102711028110291103011031110321103311034110351103611037110381103911040110411104211043110441104511046110471104811049110501105111052110531105411055110561105711058110591106011061110621106311064110651106611067110681106911070110711107211073110741107511076110771107811079110801108111082110831108411085110861108711088110891109011091110921109311094110951109611097110981109911100111011110211103111041110511106111071110811109111101111111112111131111411115111161111711118111191112011121111221112311124111251112611127111281112911130111311113211133111341113511136111371113811139111401114111142111431114411145111461114711148111491115011151111521115311154111551115611157111581115911160111611116211163111641116511166111671116811169111701117111172111731117411175111761117711178111791118011181111821118311184111851118611187111881118911190111911119211193111941119511196111971119811199112001120111202112031120411205112061120711208112091121011211112121121311214112151121611217112181121911220112211122211223112241122511226112271122811229112301123111232112331123411235112361123711238112391124011241112421124311244112451124611247112481124911250112511125211253112541125511256112571125811259112601126111262112631126411265112661126711268112691127011271112721127311274112751127611277112781127911280112811128211283112841128511286112871128811289112901129111292112931129411295112961129711298112991130011301113021130311304113051130611307113081130911310113111131211313113141131511316113171131811319113201132111322113231132411325113261132711328113291133011331113321133311334113351133611337113381133911340113411134211343113441134511346113471134811349113501135111352113531135411355113561135711358113591136011361113621136311364113651136611367113681136911370113711137211373113741137511376113771137811379113801138111382113831138411385113861138711388113891139011391113921139311394113951139611397113981139911400114011140211403114041140511406114071140811409114101141111412114131141411415114161141711418114191142011421114221142311424114251142611427114281142911430114311143211433114341143511436114371143811439114401144111442114431144411445114461144711448114491145011451114521145311454114551145611457114581145911460114611146211463114641146511466114671146811469114701147111472114731147411475114761147711478114791148011481114821148311484114851148611487114881148911490114911149211493114941149511496114971149811499115001150111502115031150411505115061150711508115091151011511115121151311514115151151611517115181151911520115211152211523115241152511526115271152811529115301153111532115331153411535115361153711538115391154011541115421154311544115451154611547115481154911550115511155211553115541155511556115571155811559115601156111562115631156411565115661156711568115691157011571115721157311574115751157611577115781157911580115811158211583115841158511586115871158811589115901159111592115931159411595115961159711598115991160011601116021160311604116051160611607116081160911610116111161211613116141161511616116171161811619116201162111622116231162411625116261162711628116291163011631116321163311634116351163611637116381163911640116411164211643116441164511646116471164811649116501165111652116531165411655116561165711658116591166011661116621166311664116651166611667116681166911670116711167211673116741167511676116771167811679116801168111682116831168411685116861168711688116891169011691116921169311694116951169611697116981169911700117011170211703117041170511706117071170811709117101171111712117131171411715117161171711718117191172011721117221172311724117251172611727117281172911730117311173211733117341173511736117371173811739117401174111742117431174411745117461174711748117491175011751117521175311754117551175611757117581175911760117611176211763117641176511766117671176811769117701177111772117731177411775117761177711778117791178011781117821178311784117851178611787117881178911790117911179211793117941179511796117971179811799118001180111802118031180411805118061180711808118091181011811118121181311814118151181611817118181181911820118211182211823118241182511826118271182811829118301183111832118331183411835118361183711838118391184011841118421184311844118451184611847118481184911850118511185211853118541185511856118571185811859118601186111862118631186411865118661186711868118691187011871118721187311874118751187611877118781187911880118811188211883118841188511886118871188811889118901189111892118931189411895118961189711898118991190011901119021190311904119051190611907119081190911910119111191211913119141191511916119171191811919119201192111922119231192411925119261192711928119291193011931119321193311934119351193611937119381193911940119411194211943119441194511946119471194811949119501195111952119531195411955119561195711958119591196011961119621196311964119651196611967119681196911970119711197211973119741197511976119771197811979119801198111982119831198411985119861198711988119891199011991119921199311994119951199611997119981199912000120011200212003120041200512006120071200812009120101201112012120131201412015120161201712018120191202012021120221202312024120251202612027120281202912030120311203212033120341203512036120371203812039120401204112042120431204412045120461204712048120491205012051120521205312054120551205612057120581205912060120611206212063120641206512066120671206812069120701207112072120731207412075120761207712078120791208012081120821208312084120851208612087120881208912090120911209212093120941209512096120971209812099121001210112102121031210412105121061210712108121091211012111121121211312114121151211612117121181211912120121211212212123121241212512126121271212812129121301213112132121331213412135121361213712138121391214012141121421214312144121451214612147121481214912150121511215212153121541215512156121571215812159121601216112162121631216412165121661216712168121691217012171121721217312174121751217612177121781217912180121811218212183121841218512186121871218812189121901219112192121931219412195121961219712198121991220012201122021220312204122051220612207122081220912210122111221212213122141221512216122171221812219122201222112222122231222412225122261222712228122291223012231122321223312234122351223612237122381223912240122411224212243122441224512246122471224812249122501225112252122531225412255122561225712258122591226012261122621226312264122651226612267122681226912270122711227212273122741227512276122771227812279122801228112282122831228412285122861228712288122891229012291122921229312294122951229612297122981229912300123011230212303123041230512306123071230812309123101231112312123131231412315123161231712318123191232012321123221232312324123251232612327123281232912330123311233212333123341233512336123371233812339123401234112342123431234412345123461234712348123491235012351123521235312354123551235612357123581235912360123611236212363123641236512366123671236812369123701237112372123731237412375123761237712378123791238012381123821238312384123851238612387123881238912390123911239212393123941239512396123971239812399124001240112402124031240412405124061240712408124091241012411124121241312414124151241612417124181241912420124211242212423124241242512426124271242812429124301243112432124331243412435124361243712438124391244012441124421244312444124451244612447124481244912450124511245212453124541245512456124571245812459124601246112462124631246412465124661246712468124691247012471124721247312474124751247612477124781247912480124811248212483124841248512486124871248812489124901249112492124931249412495124961249712498124991250012501125021250312504125051250612507125081250912510125111251212513125141251512516125171251812519125201252112522125231252412525125261252712528125291253012531125321253312534125351253612537125381253912540125411254212543125441254512546125471254812549125501255112552125531255412555125561255712558125591256012561125621256312564125651256612567125681256912570125711257212573125741257512576125771257812579125801258112582125831258412585125861258712588125891259012591125921259312594125951259612597125981259912600126011260212603126041260512606126071260812609126101261112612126131261412615126161261712618126191262012621126221262312624126251262612627126281262912630126311263212633126341263512636126371263812639126401264112642126431264412645126461264712648126491265012651126521265312654126551265612657126581265912660126611266212663126641266512666126671266812669126701267112672126731267412675126761267712678126791268012681126821268312684126851268612687126881268912690126911269212693126941269512696126971269812699127001270112702127031270412705127061270712708127091271012711127121271312714127151271612717127181271912720127211272212723127241272512726127271272812729127301273112732127331273412735127361273712738127391274012741127421274312744127451274612747127481274912750127511275212753127541275512756127571275812759127601276112762127631276412765127661276712768127691277012771127721277312774127751277612777127781277912780127811278212783127841278512786127871278812789127901279112792127931279412795127961279712798127991280012801128021280312804128051280612807128081280912810128111281212813128141281512816128171281812819128201282112822128231282412825128261282712828128291283012831128321283312834128351283612837128381283912840128411284212843128441284512846128471284812849128501285112852128531285412855128561285712858128591286012861128621286312864128651286612867128681286912870128711287212873128741287512876128771287812879128801288112882128831288412885128861288712888128891289012891128921289312894128951289612897128981289912900129011290212903129041290512906129071290812909129101291112912129131291412915129161291712918129191292012921129221292312924129251292612927129281292912930129311293212933129341293512936129371293812939129401294112942129431294412945129461294712948129491295012951129521295312954129551295612957129581295912960129611296212963129641296512966129671296812969129701297112972129731297412975129761297712978129791298012981129821298312984129851298612987129881298912990129911299212993129941299512996129971299812999130001300113002130031300413005130061300713008130091301013011130121301313014130151301613017130181301913020130211302213023130241302513026130271302813029130301303113032130331303413035130361303713038130391304013041130421304313044130451304613047130481304913050130511305213053130541305513056130571305813059130601306113062130631306413065130661306713068130691307013071130721307313074130751307613077130781307913080130811308213083130841308513086130871308813089130901309113092130931309413095130961309713098130991310013101131021310313104131051310613107131081310913110131111311213113131141311513116131171311813119131201312113122131231312413125131261312713128131291313013131131321313313134131351313613137131381313913140131411314213143131441314513146131471314813149131501315113152131531315413155131561315713158131591316013161131621316313164131651316613167131681316913170131711317213173131741317513176131771317813179131801318113182131831318413185131861318713188131891319013191131921319313194131951319613197131981319913200132011320213203132041320513206132071320813209132101321113212132131321413215132161321713218132191322013221132221322313224132251322613227132281322913230132311323213233132341323513236132371323813239132401324113242132431324413245132461324713248132491325013251132521325313254132551325613257132581325913260132611326213263132641326513266132671326813269132701327113272132731327413275132761327713278132791328013281132821328313284132851328613287132881328913290132911329213293132941329513296132971329813299133001330113302133031330413305133061330713308133091331013311133121331313314133151331613317133181331913320133211332213323133241332513326133271332813329133301333113332133331333413335133361333713338133391334013341133421334313344133451334613347133481334913350133511335213353133541335513356133571335813359133601336113362133631336413365133661336713368133691337013371133721337313374133751337613377133781337913380133811338213383133841338513386133871338813389133901339113392133931339413395133961339713398133991340013401134021340313404134051340613407134081340913410134111341213413134141341513416134171341813419134201342113422134231342413425134261342713428134291343013431134321343313434134351343613437134381343913440134411344213443134441344513446134471344813449134501345113452134531345413455134561345713458134591346013461134621346313464134651346613467134681346913470134711347213473134741347513476134771347813479134801348113482134831348413485134861348713488134891349013491134921349313494134951349613497134981349913500135011350213503135041350513506135071350813509135101351113512135131351413515135161351713518135191352013521135221352313524135251352613527135281352913530135311353213533135341353513536135371353813539135401354113542135431354413545135461354713548135491355013551135521355313554135551355613557135581355913560135611356213563135641356513566135671356813569135701357113572135731357413575135761357713578135791358013581135821358313584135851358613587135881358913590135911359213593135941359513596135971359813599136001360113602136031360413605136061360713608136091361013611136121361313614136151361613617136181361913620136211362213623136241362513626136271362813629136301363113632136331363413635136361363713638136391364013641136421364313644136451364613647136481364913650136511365213653136541365513656136571365813659136601366113662136631366413665136661366713668136691367013671136721367313674136751367613677136781367913680136811368213683136841368513686136871368813689136901369113692136931369413695136961369713698136991370013701137021370313704137051370613707137081370913710137111371213713137141371513716137171371813719137201372113722137231372413725137261372713728137291373013731137321373313734137351373613737137381373913740137411374213743137441374513746137471374813749137501375113752137531375413755137561375713758137591376013761137621376313764137651376613767137681376913770137711377213773137741377513776137771377813779137801378113782137831378413785137861378713788137891379013791137921379313794137951379613797137981379913800138011380213803138041380513806138071380813809138101381113812138131381413815138161381713818138191382013821138221382313824138251382613827138281382913830138311383213833138341383513836138371383813839138401384113842138431384413845138461384713848138491385013851138521385313854138551385613857138581385913860138611386213863138641386513866138671386813869138701387113872138731387413875138761387713878138791388013881138821388313884138851388613887138881388913890138911389213893138941389513896138971389813899139001390113902139031390413905139061390713908139091391013911139121391313914139151391613917139181391913920139211392213923139241392513926139271392813929139301393113932139331393413935139361393713938139391394013941139421394313944139451394613947139481394913950139511395213953139541395513956139571395813959139601396113962139631396413965139661396713968139691397013971139721397313974139751397613977139781397913980139811398213983139841398513986139871398813989139901399113992139931399413995139961399713998139991400014001140021400314004140051400614007140081400914010140111401214013140141401514016140171401814019140201402114022140231402414025140261402714028140291403014031140321403314034140351403614037140381403914040140411404214043140441404514046140471404814049140501405114052140531405414055140561405714058140591406014061140621406314064140651406614067140681406914070140711407214073140741407514076140771407814079140801408114082140831408414085140861408714088140891409014091140921409314094140951409614097140981409914100141011410214103141041410514106141071410814109141101411114112141131411414115141161411714118141191412014121141221412314124141251412614127141281412914130141311413214133141341413514136141371413814139141401414114142141431414414145141461414714148141491415014151141521415314154141551415614157141581415914160141611416214163141641416514166141671416814169141701417114172141731417414175141761417714178141791418014181141821418314184141851418614187141881418914190141911419214193141941419514196141971419814199142001420114202142031420414205142061420714208142091421014211142121421314214142151421614217142181421914220142211422214223142241422514226142271422814229142301423114232142331423414235142361423714238142391424014241142421424314244142451424614247142481424914250142511425214253142541425514256142571425814259142601426114262142631426414265142661426714268142691427014271142721427314274142751427614277142781427914280142811428214283142841428514286142871428814289142901429114292142931429414295142961429714298142991430014301143021430314304143051430614307143081430914310143111431214313143141431514316143171431814319143201432114322143231432414325143261432714328143291433014331143321433314334143351433614337143381433914340143411434214343143441434514346143471434814349143501435114352143531435414355143561435714358143591436014361143621436314364143651436614367143681436914370143711437214373143741437514376143771437814379143801438114382143831438414385143861438714388143891439014391143921439314394143951439614397143981439914400144011440214403144041440514406144071440814409144101441114412144131441414415144161441714418144191442014421144221442314424144251442614427144281442914430144311443214433144341443514436144371443814439144401444114442144431444414445144461444714448144491445014451144521445314454144551445614457144581445914460144611446214463144641446514466144671446814469144701447114472144731447414475144761447714478144791448014481144821448314484144851448614487144881448914490144911449214493144941449514496144971449814499145001450114502145031450414505145061450714508145091451014511145121451314514145151451614517145181451914520145211452214523145241452514526145271452814529145301453114532145331453414535145361453714538145391454014541145421454314544145451454614547145481454914550145511455214553145541455514556145571455814559145601456114562145631456414565145661456714568145691457014571145721457314574145751457614577145781457914580145811458214583145841458514586145871458814589145901459114592145931459414595145961459714598145991460014601146021460314604146051460614607146081460914610146111461214613146141461514616146171461814619146201462114622146231462414625146261462714628146291463014631146321463314634146351463614637146381463914640146411464214643146441464514646146471464814649146501465114652146531465414655146561465714658146591466014661146621466314664146651466614667146681466914670146711467214673146741467514676146771467814679146801468114682146831468414685146861468714688146891469014691146921469314694146951469614697146981469914700147011470214703147041470514706147071470814709147101471114712147131471414715147161471714718147191472014721147221472314724147251472614727147281472914730147311473214733147341473514736147371473814739147401474114742147431474414745147461474714748147491475014751147521475314754147551475614757147581475914760147611476214763147641476514766147671476814769147701477114772147731477414775147761477714778147791478014781147821478314784147851478614787147881478914790147911479214793147941479514796147971479814799148001480114802148031480414805148061480714808148091481014811148121481314814148151481614817148181481914820148211482214823148241482514826148271482814829148301483114832148331483414835148361483714838148391484014841148421484314844148451484614847148481484914850148511485214853148541485514856148571485814859148601486114862148631486414865148661486714868148691487014871148721487314874148751487614877148781487914880148811488214883148841488514886148871488814889148901489114892148931489414895148961489714898148991490014901149021490314904149051490614907149081490914910149111491214913149141491514916149171491814919149201492114922149231492414925149261492714928149291493014931149321493314934149351493614937149381493914940149411494214943149441494514946149471494814949149501495114952149531495414955149561495714958149591496014961149621496314964149651496614967149681496914970149711497214973149741497514976149771497814979149801498114982149831498414985149861498714988149891499014991149921499314994149951499614997149981499915000150011500215003150041500515006150071500815009150101501115012150131501415015150161501715018150191502015021150221502315024150251502615027150281502915030150311503215033150341503515036150371503815039150401504115042150431504415045150461504715048150491505015051150521505315054150551505615057150581505915060150611506215063150641506515066150671506815069150701507115072150731507415075150761507715078150791508015081150821508315084150851508615087150881508915090150911509215093150941509515096150971509815099151001510115102151031510415105151061510715108151091511015111151121511315114151151511615117151181511915120151211512215123151241512515126151271512815129151301513115132151331513415135151361513715138151391514015141151421514315144151451514615147151481514915150151511515215153151541515515156151571515815159151601516115162151631516415165151661516715168151691517015171151721517315174151751517615177151781517915180151811518215183151841518515186151871518815189151901519115192151931519415195151961519715198151991520015201152021520315204152051520615207152081520915210152111521215213152141521515216152171521815219152201522115222152231522415225152261522715228152291523015231152321523315234152351523615237152381523915240152411524215243152441524515246152471524815249152501525115252152531525415255152561525715258152591526015261152621526315264152651526615267152681526915270152711527215273152741527515276152771527815279152801528115282152831528415285152861528715288152891529015291152921529315294152951529615297152981529915300153011530215303153041530515306153071530815309153101531115312153131531415315153161531715318153191532015321153221532315324153251532615327153281532915330153311533215333153341533515336153371533815339153401534115342153431534415345153461534715348153491535015351153521535315354153551535615357153581535915360153611536215363153641536515366153671536815369153701537115372153731537415375153761537715378153791538015381153821538315384153851538615387153881538915390153911539215393153941539515396153971539815399154001540115402154031540415405154061540715408154091541015411154121541315414154151541615417154181541915420154211542215423154241542515426154271542815429154301543115432154331543415435154361543715438154391544015441154421544315444154451544615447154481544915450154511545215453154541545515456154571545815459154601546115462154631546415465154661546715468154691547015471154721547315474154751547615477154781547915480154811548215483154841548515486154871548815489154901549115492154931549415495154961549715498154991550015501155021550315504155051550615507155081550915510155111551215513155141551515516155171551815519155201552115522155231552415525155261552715528155291553015531155321553315534155351553615537155381553915540155411554215543155441554515546155471554815549155501555115552 |
- /*
- * parser.c : an XML 1.0 parser, namespaces and validity support are mostly
- * implemented on top of the SAX interfaces
- *
- * References:
- * The XML specification:
- * http://www.w3.org/TR/REC-xml
- * Original 1.0 version:
- * http://www.w3.org/TR/1998/REC-xml-19980210
- * XML second edition working draft
- * http://www.w3.org/TR/2000/WD-xml-2e-20000814
- *
- * Okay this is a big file, the parser core is around 7000 lines, then it
- * is followed by the progressive parser top routines, then the various
- * high level APIs to call the parser and a few miscellaneous functions.
- * A number of helper functions and deprecated ones have been moved to
- * parserInternals.c to reduce this file size.
- * As much as possible the functions are associated with their relative
- * production in the XML specification. A few productions defining the
- * different ranges of character are actually implanted either in
- * parserInternals.h or parserInternals.c
- * The DOM tree build is realized from the default SAX callbacks in
- * the module SAX.c.
- * The routines doing the validation checks are in valid.c and called either
- * from the SAX callbacks or as standalone functions using a preparsed
- * document.
- *
- * See Copyright for the status of this software.
- *
- * daniel@veillard.com
- */
- /* To avoid EBCDIC trouble when parsing on zOS */
- #if defined(__MVS__)
- #pragma convert("ISO8859-1")
- #endif
- #define IN_LIBXML
- #include "libxml.h"
- #if defined(_WIN32) && !defined (__CYGWIN__)
- #define XML_DIR_SEP '\\'
- #else
- #define XML_DIR_SEP '/'
- #endif
- #include <stdlib.h>
- #include <limits.h>
- #include <string.h>
- #include <stdarg.h>
- #include <stddef.h>
- #include <libxml/xmlmemory.h>
- #include <libxml/threads.h>
- #include <libxml/globals.h>
- #include <libxml/tree.h>
- #include <libxml/parser.h>
- #include <libxml/parserInternals.h>
- #include <libxml/valid.h>
- #include <libxml/entities.h>
- #include <libxml/xmlerror.h>
- #include <libxml/encoding.h>
- #include <libxml/xmlIO.h>
- #include <libxml/uri.h>
- #ifdef LIBXML_CATALOG_ENABLED
- #include <libxml/catalog.h>
- #endif
- #ifdef LIBXML_SCHEMAS_ENABLED
- #include <libxml/xmlschemastypes.h>
- #include <libxml/relaxng.h>
- #endif
- #ifdef HAVE_CTYPE_H
- #include <ctype.h>
- #endif
- #ifdef HAVE_STDLIB_H
- #include <stdlib.h>
- #endif
- #ifdef HAVE_SYS_STAT_H
- #include <sys/stat.h>
- #endif
- #ifdef HAVE_FCNTL_H
- #include <fcntl.h>
- #endif
- #ifdef HAVE_UNISTD_H
- #include <unistd.h>
- #endif
- #include "buf.h"
- #include "enc.h"
- struct _xmlStartTag {
- const xmlChar *prefix;
- const xmlChar *URI;
- int line;
- int nsNr;
- };
- static void
- xmlFatalErr(xmlParserCtxtPtr ctxt, xmlParserErrors error, const char *info);
- static xmlParserCtxtPtr
- xmlCreateEntityParserCtxtInternal(const xmlChar *URL, const xmlChar *ID,
- const xmlChar *base, xmlParserCtxtPtr pctx);
- static void xmlHaltParser(xmlParserCtxtPtr ctxt);
- static int
- xmlParseElementStart(xmlParserCtxtPtr ctxt);
- static void
- xmlParseElementEnd(xmlParserCtxtPtr ctxt);
- /************************************************************************
- * *
- * Arbitrary limits set in the parser. See XML_PARSE_HUGE *
- * *
- ************************************************************************/
- #define XML_PARSER_BIG_ENTITY 1000
- #define XML_PARSER_LOT_ENTITY 5000
- /*
- * XML_PARSER_NON_LINEAR is the threshold where the ratio of parsed entity
- * replacement over the size in byte of the input indicates that you have
- * and exponential behaviour. A value of 10 correspond to at least 3 entity
- * replacement per byte of input.
- */
- #define XML_PARSER_NON_LINEAR 10
- /*
- * xmlParserEntityCheck
- *
- * Function to check non-linear entity expansion behaviour
- * This is here to detect and stop exponential linear entity expansion
- * This is not a limitation of the parser but a safety
- * boundary feature. It can be disabled with the XML_PARSE_HUGE
- * parser option.
- */
- static int
- xmlParserEntityCheck(xmlParserCtxtPtr ctxt, size_t size,
- xmlEntityPtr ent, size_t replacement)
- {
- size_t consumed = 0;
- int i;
- if ((ctxt == NULL) || (ctxt->options & XML_PARSE_HUGE))
- return (0);
- if (ctxt->lastError.code == XML_ERR_ENTITY_LOOP)
- return (1);
- /*
- * This may look absurd but is needed to detect
- * entities problems
- */
- if ((ent != NULL) && (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
- (ent->content != NULL) && (ent->checked == 0) &&
- (ctxt->errNo != XML_ERR_ENTITY_LOOP)) {
- unsigned long oldnbent = ctxt->nbentities, diff;
- xmlChar *rep;
- ent->checked = 1;
- ++ctxt->depth;
- rep = xmlStringDecodeEntities(ctxt, ent->content,
- XML_SUBSTITUTE_REF, 0, 0, 0);
- --ctxt->depth;
- if ((rep == NULL) || (ctxt->errNo == XML_ERR_ENTITY_LOOP)) {
- ent->content[0] = 0;
- }
- diff = ctxt->nbentities - oldnbent + 1;
- if (diff > INT_MAX / 2)
- diff = INT_MAX / 2;
- ent->checked = diff * 2;
- if (rep != NULL) {
- if (xmlStrchr(rep, '<'))
- ent->checked |= 1;
- xmlFree(rep);
- rep = NULL;
- }
- }
- /*
- * Prevent entity exponential check, not just replacement while
- * parsing the DTD
- * The check is potentially costly so do that only once in a thousand
- */
- if ((ctxt->instate == XML_PARSER_DTD) && (ctxt->nbentities > 10000) &&
- (ctxt->nbentities % 1024 == 0)) {
- for (i = 0;i < ctxt->inputNr;i++) {
- consumed += ctxt->inputTab[i]->consumed +
- (ctxt->inputTab[i]->cur - ctxt->inputTab[i]->base);
- }
- if (ctxt->nbentities > consumed * XML_PARSER_NON_LINEAR) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- ctxt->instate = XML_PARSER_EOF;
- return (1);
- }
- consumed = 0;
- }
- if (replacement != 0) {
- if (replacement < XML_MAX_TEXT_LENGTH)
- return(0);
- /*
- * If the volume of entity copy reaches 10 times the
- * amount of parsed data and over the large text threshold
- * then that's very likely to be an abuse.
- */
- if (ctxt->input != NULL) {
- consumed = ctxt->input->consumed +
- (ctxt->input->cur - ctxt->input->base);
- }
- consumed += ctxt->sizeentities;
- if (replacement < XML_PARSER_NON_LINEAR * consumed)
- return(0);
- } else if (size != 0) {
- /*
- * Do the check based on the replacement size of the entity
- */
- if (size < XML_PARSER_BIG_ENTITY)
- return(0);
- /*
- * A limit on the amount of text data reasonably used
- */
- if (ctxt->input != NULL) {
- consumed = ctxt->input->consumed +
- (ctxt->input->cur - ctxt->input->base);
- }
- consumed += ctxt->sizeentities;
- if ((size < XML_PARSER_NON_LINEAR * consumed) &&
- (ctxt->nbentities * 3 < XML_PARSER_NON_LINEAR * consumed))
- return (0);
- } else if (ent != NULL) {
- /*
- * use the number of parsed entities in the replacement
- */
- size = ent->checked / 2;
- /*
- * The amount of data parsed counting entities size only once
- */
- if (ctxt->input != NULL) {
- consumed = ctxt->input->consumed +
- (ctxt->input->cur - ctxt->input->base);
- }
- consumed += ctxt->sizeentities;
- /*
- * Check the density of entities for the amount of data
- * knowing an entity reference will take at least 3 bytes
- */
- if (size * 3 < consumed * XML_PARSER_NON_LINEAR)
- return (0);
- } else {
- /*
- * strange we got no data for checking
- */
- if (((ctxt->lastError.code != XML_ERR_UNDECLARED_ENTITY) &&
- (ctxt->lastError.code != XML_WAR_UNDECLARED_ENTITY)) ||
- (ctxt->nbentities <= 10000))
- return (0);
- }
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- return (1);
- }
- /**
- * xmlParserMaxDepth:
- *
- * arbitrary depth limit for the XML documents that we allow to
- * process. This is not a limitation of the parser but a safety
- * boundary feature. It can be disabled with the XML_PARSE_HUGE
- * parser option.
- */
- unsigned int xmlParserMaxDepth = 256;
- #define SAX2 1
- #define XML_PARSER_BIG_BUFFER_SIZE 300
- #define XML_PARSER_BUFFER_SIZE 100
- #define SAX_COMPAT_MODE BAD_CAST "SAX compatibility mode document"
- /**
- * XML_PARSER_CHUNK_SIZE
- *
- * When calling GROW that's the minimal amount of data
- * the parser expected to have received. It is not a hard
- * limit but an optimization when reading strings like Names
- * It is not strictly needed as long as inputs available characters
- * are followed by 0, which should be provided by the I/O level
- */
- #define XML_PARSER_CHUNK_SIZE 100
- /*
- * List of XML prefixed PI allowed by W3C specs
- */
- static const char *xmlW3CPIs[] = {
- "xml-stylesheet",
- "xml-model",
- NULL
- };
- /* DEPR void xmlParserHandleReference(xmlParserCtxtPtr ctxt); */
- static xmlEntityPtr xmlParseStringPEReference(xmlParserCtxtPtr ctxt,
- const xmlChar **str);
- static xmlParserErrors
- xmlParseExternalEntityPrivate(xmlDocPtr doc, xmlParserCtxtPtr oldctxt,
- xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *URL,
- const xmlChar *ID, xmlNodePtr *list);
- static int
- xmlCtxtUseOptionsInternal(xmlParserCtxtPtr ctxt, int options,
- const char *encoding);
- #ifdef LIBXML_LEGACY_ENABLED
- static void
- xmlAddEntityReference(xmlEntityPtr ent, xmlNodePtr firstNode,
- xmlNodePtr lastNode);
- #endif /* LIBXML_LEGACY_ENABLED */
- static xmlParserErrors
- xmlParseBalancedChunkMemoryInternal(xmlParserCtxtPtr oldctxt,
- const xmlChar *string, void *user_data, xmlNodePtr *lst);
- static int
- xmlLoadEntityContent(xmlParserCtxtPtr ctxt, xmlEntityPtr entity);
- /************************************************************************
- * *
- * Some factorized error routines *
- * *
- ************************************************************************/
- /**
- * xmlErrAttributeDup:
- * @ctxt: an XML parser context
- * @prefix: the attribute prefix
- * @localname: the attribute localname
- *
- * Handle a redefinition of attribute error
- */
- static void
- xmlErrAttributeDup(xmlParserCtxtPtr ctxt, const xmlChar * prefix,
- const xmlChar * localname)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = XML_ERR_ATTRIBUTE_REDEFINED;
- if (prefix == NULL)
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER,
- XML_ERR_ATTRIBUTE_REDEFINED, XML_ERR_FATAL, NULL, 0,
- (const char *) localname, NULL, NULL, 0, 0,
- "Attribute %s redefined\n", localname);
- else
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER,
- XML_ERR_ATTRIBUTE_REDEFINED, XML_ERR_FATAL, NULL, 0,
- (const char *) prefix, (const char *) localname,
- NULL, 0, 0, "Attribute %s:%s redefined\n", prefix,
- localname);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @extra: extra information string
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void
- xmlFatalErr(xmlParserCtxtPtr ctxt, xmlParserErrors error, const char *info)
- {
- const char *errmsg;
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- switch (error) {
- case XML_ERR_INVALID_HEX_CHARREF:
- errmsg = "CharRef: invalid hexadecimal value";
- break;
- case XML_ERR_INVALID_DEC_CHARREF:
- errmsg = "CharRef: invalid decimal value";
- break;
- case XML_ERR_INVALID_CHARREF:
- errmsg = "CharRef: invalid value";
- break;
- case XML_ERR_INTERNAL_ERROR:
- errmsg = "internal error";
- break;
- case XML_ERR_PEREF_AT_EOF:
- errmsg = "PEReference at end of document";
- break;
- case XML_ERR_PEREF_IN_PROLOG:
- errmsg = "PEReference in prolog";
- break;
- case XML_ERR_PEREF_IN_EPILOG:
- errmsg = "PEReference in epilog";
- break;
- case XML_ERR_PEREF_NO_NAME:
- errmsg = "PEReference: no name";
- break;
- case XML_ERR_PEREF_SEMICOL_MISSING:
- errmsg = "PEReference: expecting ';'";
- break;
- case XML_ERR_ENTITY_LOOP:
- errmsg = "Detected an entity reference loop";
- break;
- case XML_ERR_ENTITY_NOT_STARTED:
- errmsg = "EntityValue: \" or ' expected";
- break;
- case XML_ERR_ENTITY_PE_INTERNAL:
- errmsg = "PEReferences forbidden in internal subset";
- break;
- case XML_ERR_ENTITY_NOT_FINISHED:
- errmsg = "EntityValue: \" or ' expected";
- break;
- case XML_ERR_ATTRIBUTE_NOT_STARTED:
- errmsg = "AttValue: \" or ' expected";
- break;
- case XML_ERR_LT_IN_ATTRIBUTE:
- errmsg = "Unescaped '<' not allowed in attributes values";
- break;
- case XML_ERR_LITERAL_NOT_STARTED:
- errmsg = "SystemLiteral \" or ' expected";
- break;
- case XML_ERR_LITERAL_NOT_FINISHED:
- errmsg = "Unfinished System or Public ID \" or ' expected";
- break;
- case XML_ERR_MISPLACED_CDATA_END:
- errmsg = "Sequence ']]>' not allowed in content";
- break;
- case XML_ERR_URI_REQUIRED:
- errmsg = "SYSTEM or PUBLIC, the URI is missing";
- break;
- case XML_ERR_PUBID_REQUIRED:
- errmsg = "PUBLIC, the Public Identifier is missing";
- break;
- case XML_ERR_HYPHEN_IN_COMMENT:
- errmsg = "Comment must not contain '--' (double-hyphen)";
- break;
- case XML_ERR_PI_NOT_STARTED:
- errmsg = "xmlParsePI : no target name";
- break;
- case XML_ERR_RESERVED_XML_NAME:
- errmsg = "Invalid PI name";
- break;
- case XML_ERR_NOTATION_NOT_STARTED:
- errmsg = "NOTATION: Name expected here";
- break;
- case XML_ERR_NOTATION_NOT_FINISHED:
- errmsg = "'>' required to close NOTATION declaration";
- break;
- case XML_ERR_VALUE_REQUIRED:
- errmsg = "Entity value required";
- break;
- case XML_ERR_URI_FRAGMENT:
- errmsg = "Fragment not allowed";
- break;
- case XML_ERR_ATTLIST_NOT_STARTED:
- errmsg = "'(' required to start ATTLIST enumeration";
- break;
- case XML_ERR_NMTOKEN_REQUIRED:
- errmsg = "NmToken expected in ATTLIST enumeration";
- break;
- case XML_ERR_ATTLIST_NOT_FINISHED:
- errmsg = "')' required to finish ATTLIST enumeration";
- break;
- case XML_ERR_MIXED_NOT_STARTED:
- errmsg = "MixedContentDecl : '|' or ')*' expected";
- break;
- case XML_ERR_PCDATA_REQUIRED:
- errmsg = "MixedContentDecl : '#PCDATA' expected";
- break;
- case XML_ERR_ELEMCONTENT_NOT_STARTED:
- errmsg = "ContentDecl : Name or '(' expected";
- break;
- case XML_ERR_ELEMCONTENT_NOT_FINISHED:
- errmsg = "ContentDecl : ',' '|' or ')' expected";
- break;
- case XML_ERR_PEREF_IN_INT_SUBSET:
- errmsg =
- "PEReference: forbidden within markup decl in internal subset";
- break;
- case XML_ERR_GT_REQUIRED:
- errmsg = "expected '>'";
- break;
- case XML_ERR_CONDSEC_INVALID:
- errmsg = "XML conditional section '[' expected";
- break;
- case XML_ERR_EXT_SUBSET_NOT_FINISHED:
- errmsg = "Content error in the external subset";
- break;
- case XML_ERR_CONDSEC_INVALID_KEYWORD:
- errmsg =
- "conditional section INCLUDE or IGNORE keyword expected";
- break;
- case XML_ERR_CONDSEC_NOT_FINISHED:
- errmsg = "XML conditional section not closed";
- break;
- case XML_ERR_XMLDECL_NOT_STARTED:
- errmsg = "Text declaration '<?xml' required";
- break;
- case XML_ERR_XMLDECL_NOT_FINISHED:
- errmsg = "parsing XML declaration: '?>' expected";
- break;
- case XML_ERR_EXT_ENTITY_STANDALONE:
- errmsg = "external parsed entities cannot be standalone";
- break;
- case XML_ERR_ENTITYREF_SEMICOL_MISSING:
- errmsg = "EntityRef: expecting ';'";
- break;
- case XML_ERR_DOCTYPE_NOT_FINISHED:
- errmsg = "DOCTYPE improperly terminated";
- break;
- case XML_ERR_LTSLASH_REQUIRED:
- errmsg = "EndTag: '</' not found";
- break;
- case XML_ERR_EQUAL_REQUIRED:
- errmsg = "expected '='";
- break;
- case XML_ERR_STRING_NOT_CLOSED:
- errmsg = "String not closed expecting \" or '";
- break;
- case XML_ERR_STRING_NOT_STARTED:
- errmsg = "String not started expecting ' or \"";
- break;
- case XML_ERR_ENCODING_NAME:
- errmsg = "Invalid XML encoding name";
- break;
- case XML_ERR_STANDALONE_VALUE:
- errmsg = "standalone accepts only 'yes' or 'no'";
- break;
- case XML_ERR_DOCUMENT_EMPTY:
- errmsg = "Document is empty";
- break;
- case XML_ERR_DOCUMENT_END:
- errmsg = "Extra content at the end of the document";
- break;
- case XML_ERR_NOT_WELL_BALANCED:
- errmsg = "chunk is not well balanced";
- break;
- case XML_ERR_EXTRA_CONTENT:
- errmsg = "extra content at the end of well balanced chunk";
- break;
- case XML_ERR_VERSION_MISSING:
- errmsg = "Malformed declaration expecting version";
- break;
- case XML_ERR_NAME_TOO_LONG:
- errmsg = "Name too long use XML_PARSE_HUGE option";
- break;
- #if 0
- case:
- errmsg = "";
- break;
- #endif
- default:
- errmsg = "Unregistered error message";
- }
- if (ctxt != NULL)
- ctxt->errNo = error;
- if (info == NULL) {
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_FATAL, NULL, 0, info, NULL, NULL, 0, 0, "%s\n",
- errmsg);
- } else {
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_FATAL, NULL, 0, info, NULL, NULL, 0, 0, "%s: %s\n",
- errmsg, info);
- }
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErrMsg:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsg(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_FATAL, NULL, 0, NULL, NULL, NULL, 0, 0, "%s", msg);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlWarningMsg:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @str1: extra data
- * @str2: extra data
- *
- * Handle a warning.
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlWarningMsg(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar *str1, const xmlChar *str2)
- {
- xmlStructuredErrorFunc schannel = NULL;
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if ((ctxt != NULL) && (ctxt->sax != NULL) &&
- (ctxt->sax->initialized == XML_SAX2_MAGIC))
- schannel = ctxt->sax->serror;
- if (ctxt != NULL) {
- __xmlRaiseError(schannel,
- (ctxt->sax) ? ctxt->sax->warning : NULL,
- ctxt->userData,
- ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_WARNING, NULL, 0,
- (const char *) str1, (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- } else {
- __xmlRaiseError(schannel, NULL, NULL,
- ctxt, NULL, XML_FROM_PARSER, error,
- XML_ERR_WARNING, NULL, 0,
- (const char *) str1, (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- }
- }
- /**
- * xmlValidityError:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @str1: extra data
- *
- * Handle a validity error.
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlValidityError(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar *str1, const xmlChar *str2)
- {
- xmlStructuredErrorFunc schannel = NULL;
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL) {
- ctxt->errNo = error;
- if ((ctxt->sax != NULL) && (ctxt->sax->initialized == XML_SAX2_MAGIC))
- schannel = ctxt->sax->serror;
- }
- if (ctxt != NULL) {
- __xmlRaiseError(schannel,
- ctxt->vctxt.error, ctxt->vctxt.userData,
- ctxt, NULL, XML_FROM_DTD, error,
- XML_ERR_ERROR, NULL, 0, (const char *) str1,
- (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- ctxt->valid = 0;
- } else {
- __xmlRaiseError(schannel, NULL, NULL,
- ctxt, NULL, XML_FROM_DTD, error,
- XML_ERR_ERROR, NULL, 0, (const char *) str1,
- (const char *) str2, NULL, 0, 0,
- msg, (const char *) str1, (const char *) str2);
- }
- }
- /**
- * xmlFatalErrMsgInt:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @val: an integer value
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsgInt(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, int val)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL,
- ctxt, NULL, XML_FROM_PARSER, error, XML_ERR_FATAL,
- NULL, 0, NULL, NULL, NULL, val, 0, msg, val);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErrMsgStrIntStr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @str1: an string info
- * @val: an integer value
- * @str2: an string info
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsgStrIntStr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar *str1, int val,
- const xmlChar *str2)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL,
- ctxt, NULL, XML_FROM_PARSER, error, XML_ERR_FATAL,
- NULL, 0, (const char *) str1, (const char *) str2,
- NULL, val, 0, msg, str1, val, str2);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlFatalErrMsgStr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @val: a string value
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlFatalErrMsgStr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar * val)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL,
- XML_FROM_PARSER, error, XML_ERR_FATAL,
- NULL, 0, (const char *) val, NULL, NULL, 0, 0, msg,
- val);
- if (ctxt != NULL) {
- ctxt->wellFormed = 0;
- if (ctxt->recovery == 0)
- ctxt->disableSAX = 1;
- }
- }
- /**
- * xmlErrMsgStr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the error message
- * @val: a string value
- *
- * Handle a non fatal parser error
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlErrMsgStr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg, const xmlChar * val)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL,
- XML_FROM_PARSER, error, XML_ERR_ERROR,
- NULL, 0, (const char *) val, NULL, NULL, 0, 0, msg,
- val);
- }
- /**
- * xmlNsErr:
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the message
- * @info1: extra information string
- * @info2: extra information string
- *
- * Handle a fatal parser error, i.e. violating Well-Formedness constraints
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlNsErr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg,
- const xmlChar * info1, const xmlChar * info2,
- const xmlChar * info3)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- if (ctxt != NULL)
- ctxt->errNo = error;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_NAMESPACE, error,
- XML_ERR_ERROR, NULL, 0, (const char *) info1,
- (const char *) info2, (const char *) info3, 0, 0, msg,
- info1, info2, info3);
- if (ctxt != NULL)
- ctxt->nsWellFormed = 0;
- }
- /**
- * xmlNsWarn
- * @ctxt: an XML parser context
- * @error: the error number
- * @msg: the message
- * @info1: extra information string
- * @info2: extra information string
- *
- * Handle a namespace warning error
- */
- static void LIBXML_ATTR_FORMAT(3,0)
- xmlNsWarn(xmlParserCtxtPtr ctxt, xmlParserErrors error,
- const char *msg,
- const xmlChar * info1, const xmlChar * info2,
- const xmlChar * info3)
- {
- if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
- (ctxt->instate == XML_PARSER_EOF))
- return;
- __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_NAMESPACE, error,
- XML_ERR_WARNING, NULL, 0, (const char *) info1,
- (const char *) info2, (const char *) info3, 0, 0, msg,
- info1, info2, info3);
- }
- /************************************************************************
- * *
- * Library wide options *
- * *
- ************************************************************************/
- /**
- * xmlHasFeature:
- * @feature: the feature to be examined
- *
- * Examines if the library has been compiled with a given feature.
- *
- * Returns a non-zero value if the feature exist, otherwise zero.
- * Returns zero (0) if the feature does not exist or an unknown
- * unknown feature is requested, non-zero otherwise.
- */
- int
- xmlHasFeature(xmlFeature feature)
- {
- switch (feature) {
- case XML_WITH_THREAD:
- #ifdef LIBXML_THREAD_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_TREE:
- #ifdef LIBXML_TREE_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_OUTPUT:
- #ifdef LIBXML_OUTPUT_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_PUSH:
- #ifdef LIBXML_PUSH_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_READER:
- #ifdef LIBXML_READER_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_PATTERN:
- #ifdef LIBXML_PATTERN_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_WRITER:
- #ifdef LIBXML_WRITER_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_SAX1:
- #ifdef LIBXML_SAX1_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_FTP:
- #ifdef LIBXML_FTP_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_HTTP:
- #ifdef LIBXML_HTTP_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_VALID:
- #ifdef LIBXML_VALID_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_HTML:
- #ifdef LIBXML_HTML_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_LEGACY:
- #ifdef LIBXML_LEGACY_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_C14N:
- #ifdef LIBXML_C14N_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_CATALOG:
- #ifdef LIBXML_CATALOG_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_XPATH:
- #ifdef LIBXML_XPATH_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_XPTR:
- #ifdef LIBXML_XPTR_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_XINCLUDE:
- #ifdef LIBXML_XINCLUDE_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ICONV:
- #ifdef LIBXML_ICONV_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ISO8859X:
- #ifdef LIBXML_ISO8859X_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_UNICODE:
- #ifdef LIBXML_UNICODE_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_REGEXP:
- #ifdef LIBXML_REGEXP_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_AUTOMATA:
- #ifdef LIBXML_AUTOMATA_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_EXPR:
- #ifdef LIBXML_EXPR_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_SCHEMAS:
- #ifdef LIBXML_SCHEMAS_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_SCHEMATRON:
- #ifdef LIBXML_SCHEMATRON_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_MODULES:
- #ifdef LIBXML_MODULES_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_DEBUG:
- #ifdef LIBXML_DEBUG_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_DEBUG_MEM:
- #ifdef DEBUG_MEMORY_LOCATION
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_DEBUG_RUN:
- #ifdef LIBXML_DEBUG_RUNTIME
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ZLIB:
- #ifdef LIBXML_ZLIB_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_LZMA:
- #ifdef LIBXML_LZMA_ENABLED
- return(1);
- #else
- return(0);
- #endif
- case XML_WITH_ICU:
- #ifdef LIBXML_ICU_ENABLED
- return(1);
- #else
- return(0);
- #endif
- default:
- break;
- }
- return(0);
- }
- /************************************************************************
- * *
- * SAX2 defaulted attributes handling *
- * *
- ************************************************************************/
- /**
- * xmlDetectSAX2:
- * @ctxt: an XML parser context
- *
- * Do the SAX2 detection and specific initialization
- */
- static void
- xmlDetectSAX2(xmlParserCtxtPtr ctxt) {
- xmlSAXHandlerPtr sax;
- if (ctxt == NULL) return;
- sax = ctxt->sax;
- #ifdef LIBXML_SAX1_ENABLED
- if ((sax) && (sax->initialized == XML_SAX2_MAGIC) &&
- ((sax->startElementNs != NULL) ||
- (sax->endElementNs != NULL) ||
- ((sax->startElement == NULL) && (sax->endElement == NULL))))
- ctxt->sax2 = 1;
- #else
- ctxt->sax2 = 1;
- #endif /* LIBXML_SAX1_ENABLED */
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- if ((ctxt->str_xml==NULL) || (ctxt->str_xmlns==NULL) ||
- (ctxt->str_xml_ns == NULL)) {
- xmlErrMemory(ctxt, NULL);
- }
- }
- typedef struct _xmlDefAttrs xmlDefAttrs;
- typedef xmlDefAttrs *xmlDefAttrsPtr;
- struct _xmlDefAttrs {
- int nbAttrs; /* number of defaulted attributes on that element */
- int maxAttrs; /* the size of the array */
- #if __STDC_VERSION__ >= 199901L
- /* Using a C99 flexible array member avoids UBSan errors. */
- const xmlChar *values[]; /* array of localname/prefix/values/external */
- #else
- const xmlChar *values[5];
- #endif
- };
- /**
- * xmlAttrNormalizeSpace:
- * @src: the source string
- * @dst: the target string
- *
- * Normalize the space in non CDATA attribute values:
- * If the attribute type is not CDATA, then the XML processor MUST further
- * process the normalized attribute value by discarding any leading and
- * trailing space (#x20) characters, and by replacing sequences of space
- * (#x20) characters by a single space (#x20) character.
- * Note that the size of dst need to be at least src, and if one doesn't need
- * to preserve dst (and it doesn't come from a dictionary or read-only) then
- * passing src as dst is just fine.
- *
- * Returns a pointer to the normalized value (dst) or NULL if no conversion
- * is needed.
- */
- static xmlChar *
- xmlAttrNormalizeSpace(const xmlChar *src, xmlChar *dst)
- {
- if ((src == NULL) || (dst == NULL))
- return(NULL);
- while (*src == 0x20) src++;
- while (*src != 0) {
- if (*src == 0x20) {
- while (*src == 0x20) src++;
- if (*src != 0)
- *dst++ = 0x20;
- } else {
- *dst++ = *src++;
- }
- }
- *dst = 0;
- if (dst == src)
- return(NULL);
- return(dst);
- }
- /**
- * xmlAttrNormalizeSpace2:
- * @src: the source string
- *
- * Normalize the space in non CDATA attribute values, a slightly more complex
- * front end to avoid allocation problems when running on attribute values
- * coming from the input.
- *
- * Returns a pointer to the normalized value (dst) or NULL if no conversion
- * is needed.
- */
- static const xmlChar *
- xmlAttrNormalizeSpace2(xmlParserCtxtPtr ctxt, xmlChar *src, int *len)
- {
- int i;
- int remove_head = 0;
- int need_realloc = 0;
- const xmlChar *cur;
- if ((ctxt == NULL) || (src == NULL) || (len == NULL))
- return(NULL);
- i = *len;
- if (i <= 0)
- return(NULL);
- cur = src;
- while (*cur == 0x20) {
- cur++;
- remove_head++;
- }
- while (*cur != 0) {
- if (*cur == 0x20) {
- cur++;
- if ((*cur == 0x20) || (*cur == 0)) {
- need_realloc = 1;
- break;
- }
- } else
- cur++;
- }
- if (need_realloc) {
- xmlChar *ret;
- ret = xmlStrndup(src + remove_head, i - remove_head + 1);
- if (ret == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- xmlAttrNormalizeSpace(ret, ret);
- *len = (int) strlen((const char *)ret);
- return(ret);
- } else if (remove_head) {
- *len -= remove_head;
- memmove(src, src + remove_head, 1 + *len);
- return(src);
- }
- return(NULL);
- }
- /**
- * xmlAddDefAttrs:
- * @ctxt: an XML parser context
- * @fullname: the element fullname
- * @fullattr: the attribute fullname
- * @value: the attribute value
- *
- * Add a defaulted attribute for an element
- */
- static void
- xmlAddDefAttrs(xmlParserCtxtPtr ctxt,
- const xmlChar *fullname,
- const xmlChar *fullattr,
- const xmlChar *value) {
- xmlDefAttrsPtr defaults;
- int len;
- const xmlChar *name;
- const xmlChar *prefix;
- /*
- * Allows to detect attribute redefinitions
- */
- if (ctxt->attsSpecial != NULL) {
- if (xmlHashLookup2(ctxt->attsSpecial, fullname, fullattr) != NULL)
- return;
- }
- if (ctxt->attsDefault == NULL) {
- ctxt->attsDefault = xmlHashCreateDict(10, ctxt->dict);
- if (ctxt->attsDefault == NULL)
- goto mem_error;
- }
- /*
- * split the element name into prefix:localname , the string found
- * are within the DTD and then not associated to namespace names.
- */
- name = xmlSplitQName3(fullname, &len);
- if (name == NULL) {
- name = xmlDictLookup(ctxt->dict, fullname, -1);
- prefix = NULL;
- } else {
- name = xmlDictLookup(ctxt->dict, name, -1);
- prefix = xmlDictLookup(ctxt->dict, fullname, len);
- }
- /*
- * make sure there is some storage
- */
- defaults = xmlHashLookup2(ctxt->attsDefault, name, prefix);
- if (defaults == NULL) {
- defaults = (xmlDefAttrsPtr) xmlMalloc(sizeof(xmlDefAttrs) +
- (4 * 5) * sizeof(const xmlChar *));
- if (defaults == NULL)
- goto mem_error;
- defaults->nbAttrs = 0;
- defaults->maxAttrs = 4;
- if (xmlHashUpdateEntry2(ctxt->attsDefault, name, prefix,
- defaults, NULL) < 0) {
- xmlFree(defaults);
- goto mem_error;
- }
- } else if (defaults->nbAttrs >= defaults->maxAttrs) {
- xmlDefAttrsPtr temp;
- temp = (xmlDefAttrsPtr) xmlRealloc(defaults, sizeof(xmlDefAttrs) +
- (2 * defaults->maxAttrs * 5) * sizeof(const xmlChar *));
- if (temp == NULL)
- goto mem_error;
- defaults = temp;
- defaults->maxAttrs *= 2;
- if (xmlHashUpdateEntry2(ctxt->attsDefault, name, prefix,
- defaults, NULL) < 0) {
- xmlFree(defaults);
- goto mem_error;
- }
- }
- /*
- * Split the element name into prefix:localname , the string found
- * are within the DTD and hen not associated to namespace names.
- */
- name = xmlSplitQName3(fullattr, &len);
- if (name == NULL) {
- name = xmlDictLookup(ctxt->dict, fullattr, -1);
- prefix = NULL;
- } else {
- name = xmlDictLookup(ctxt->dict, name, -1);
- prefix = xmlDictLookup(ctxt->dict, fullattr, len);
- }
- defaults->values[5 * defaults->nbAttrs] = name;
- defaults->values[5 * defaults->nbAttrs + 1] = prefix;
- /* intern the string and precompute the end */
- len = xmlStrlen(value);
- value = xmlDictLookup(ctxt->dict, value, len);
- defaults->values[5 * defaults->nbAttrs + 2] = value;
- defaults->values[5 * defaults->nbAttrs + 3] = value + len;
- if (ctxt->external)
- defaults->values[5 * defaults->nbAttrs + 4] = BAD_CAST "external";
- else
- defaults->values[5 * defaults->nbAttrs + 4] = NULL;
- defaults->nbAttrs++;
- return;
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return;
- }
- /**
- * xmlAddSpecialAttr:
- * @ctxt: an XML parser context
- * @fullname: the element fullname
- * @fullattr: the attribute fullname
- * @type: the attribute type
- *
- * Register this attribute type
- */
- static void
- xmlAddSpecialAttr(xmlParserCtxtPtr ctxt,
- const xmlChar *fullname,
- const xmlChar *fullattr,
- int type)
- {
- if (ctxt->attsSpecial == NULL) {
- ctxt->attsSpecial = xmlHashCreateDict(10, ctxt->dict);
- if (ctxt->attsSpecial == NULL)
- goto mem_error;
- }
- if (xmlHashLookup2(ctxt->attsSpecial, fullname, fullattr) != NULL)
- return;
- xmlHashAddEntry2(ctxt->attsSpecial, fullname, fullattr,
- (void *) (ptrdiff_t) type);
- return;
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return;
- }
- /**
- * xmlCleanSpecialAttrCallback:
- *
- * Removes CDATA attributes from the special attribute table
- */
- static void
- xmlCleanSpecialAttrCallback(void *payload, void *data,
- const xmlChar *fullname, const xmlChar *fullattr,
- const xmlChar *unused ATTRIBUTE_UNUSED) {
- xmlParserCtxtPtr ctxt = (xmlParserCtxtPtr) data;
- if (((ptrdiff_t) payload) == XML_ATTRIBUTE_CDATA) {
- xmlHashRemoveEntry2(ctxt->attsSpecial, fullname, fullattr, NULL);
- }
- }
- /**
- * xmlCleanSpecialAttr:
- * @ctxt: an XML parser context
- *
- * Trim the list of attributes defined to remove all those of type
- * CDATA as they are not special. This call should be done when finishing
- * to parse the DTD and before starting to parse the document root.
- */
- static void
- xmlCleanSpecialAttr(xmlParserCtxtPtr ctxt)
- {
- if (ctxt->attsSpecial == NULL)
- return;
- xmlHashScanFull(ctxt->attsSpecial, xmlCleanSpecialAttrCallback, ctxt);
- if (xmlHashSize(ctxt->attsSpecial) == 0) {
- xmlHashFree(ctxt->attsSpecial, NULL);
- ctxt->attsSpecial = NULL;
- }
- return;
- }
- /**
- * xmlCheckLanguageID:
- * @lang: pointer to the string value
- *
- * Checks that the value conforms to the LanguageID production:
- *
- * NOTE: this is somewhat deprecated, those productions were removed from
- * the XML Second edition.
- *
- * [33] LanguageID ::= Langcode ('-' Subcode)*
- * [34] Langcode ::= ISO639Code | IanaCode | UserCode
- * [35] ISO639Code ::= ([a-z] | [A-Z]) ([a-z] | [A-Z])
- * [36] IanaCode ::= ('i' | 'I') '-' ([a-z] | [A-Z])+
- * [37] UserCode ::= ('x' | 'X') '-' ([a-z] | [A-Z])+
- * [38] Subcode ::= ([a-z] | [A-Z])+
- *
- * The current REC reference the successors of RFC 1766, currently 5646
- *
- * http://www.rfc-editor.org/rfc/rfc5646.txt
- * langtag = language
- * ["-" script]
- * ["-" region]
- * *("-" variant)
- * *("-" extension)
- * ["-" privateuse]
- * language = 2*3ALPHA ; shortest ISO 639 code
- * ["-" extlang] ; sometimes followed by
- * ; extended language subtags
- * / 4ALPHA ; or reserved for future use
- * / 5*8ALPHA ; or registered language subtag
- *
- * extlang = 3ALPHA ; selected ISO 639 codes
- * *2("-" 3ALPHA) ; permanently reserved
- *
- * script = 4ALPHA ; ISO 15924 code
- *
- * region = 2ALPHA ; ISO 3166-1 code
- * / 3DIGIT ; UN M.49 code
- *
- * variant = 5*8alphanum ; registered variants
- * / (DIGIT 3alphanum)
- *
- * extension = singleton 1*("-" (2*8alphanum))
- *
- * ; Single alphanumerics
- * ; "x" reserved for private use
- * singleton = DIGIT ; 0 - 9
- * / %x41-57 ; A - W
- * / %x59-5A ; Y - Z
- * / %x61-77 ; a - w
- * / %x79-7A ; y - z
- *
- * it sounds right to still allow Irregular i-xxx IANA and user codes too
- * The parser below doesn't try to cope with extension or privateuse
- * that could be added but that's not interoperable anyway
- *
- * Returns 1 if correct 0 otherwise
- **/
- int
- xmlCheckLanguageID(const xmlChar * lang)
- {
- const xmlChar *cur = lang, *nxt;
- if (cur == NULL)
- return (0);
- if (((cur[0] == 'i') && (cur[1] == '-')) ||
- ((cur[0] == 'I') && (cur[1] == '-')) ||
- ((cur[0] == 'x') && (cur[1] == '-')) ||
- ((cur[0] == 'X') && (cur[1] == '-'))) {
- /*
- * Still allow IANA code and user code which were coming
- * from the previous version of the XML-1.0 specification
- * it's deprecated but we should not fail
- */
- cur += 2;
- while (((cur[0] >= 'A') && (cur[0] <= 'Z')) ||
- ((cur[0] >= 'a') && (cur[0] <= 'z')))
- cur++;
- return(cur[0] == 0);
- }
- nxt = cur;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if (nxt - cur >= 4) {
- /*
- * Reserved
- */
- if ((nxt - cur > 8) || (nxt[0] != 0))
- return(0);
- return(1);
- }
- if (nxt - cur < 2)
- return(0);
- /* we got an ISO 639 code */
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can have extlang or script or region or variant */
- if ((nxt[0] >= '0') && (nxt[0] <= '9'))
- goto region_m49;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if (nxt - cur == 4)
- goto script;
- if (nxt - cur == 2)
- goto region;
- if ((nxt - cur >= 5) && (nxt - cur <= 8))
- goto variant;
- if (nxt - cur != 3)
- return(0);
- /* we parsed an extlang */
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can have script or region or variant */
- if ((nxt[0] >= '0') && (nxt[0] <= '9'))
- goto region_m49;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if (nxt - cur == 2)
- goto region;
- if ((nxt - cur >= 5) && (nxt - cur <= 8))
- goto variant;
- if (nxt - cur != 4)
- return(0);
- /* we parsed a script */
- script:
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can have region or variant */
- if ((nxt[0] >= '0') && (nxt[0] <= '9'))
- goto region_m49;
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if ((nxt - cur >= 5) && (nxt - cur <= 8))
- goto variant;
- if (nxt - cur != 2)
- return(0);
- /* we parsed a region */
- region:
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- nxt++;
- cur = nxt;
- /* now we can just have a variant */
- while (((nxt[0] >= 'A') && (nxt[0] <= 'Z')) ||
- ((nxt[0] >= 'a') && (nxt[0] <= 'z')))
- nxt++;
- if ((nxt - cur < 5) || (nxt - cur > 8))
- return(0);
- /* we parsed a variant */
- variant:
- if (nxt[0] == 0)
- return(1);
- if (nxt[0] != '-')
- return(0);
- /* extensions and private use subtags not checked */
- return (1);
- region_m49:
- if (((nxt[1] >= '0') && (nxt[1] <= '9')) &&
- ((nxt[2] >= '0') && (nxt[2] <= '9'))) {
- nxt += 3;
- goto region;
- }
- return(0);
- }
- /************************************************************************
- * *
- * Parser stacks related functions and macros *
- * *
- ************************************************************************/
- static xmlEntityPtr xmlParseStringEntityRef(xmlParserCtxtPtr ctxt,
- const xmlChar ** str);
- #ifdef SAX2
- /**
- * nsPush:
- * @ctxt: an XML parser context
- * @prefix: the namespace prefix or NULL
- * @URL: the namespace name
- *
- * Pushes a new parser namespace on top of the ns stack
- *
- * Returns -1 in case of error, -2 if the namespace should be discarded
- * and the index in the stack otherwise.
- */
- static int
- nsPush(xmlParserCtxtPtr ctxt, const xmlChar *prefix, const xmlChar *URL)
- {
- if (ctxt->options & XML_PARSE_NSCLEAN) {
- int i;
- for (i = ctxt->nsNr - 2;i >= 0;i -= 2) {
- if (ctxt->nsTab[i] == prefix) {
- /* in scope */
- if (ctxt->nsTab[i + 1] == URL)
- return(-2);
- /* out of scope keep it */
- break;
- }
- }
- }
- if ((ctxt->nsMax == 0) || (ctxt->nsTab == NULL)) {
- ctxt->nsMax = 10;
- ctxt->nsNr = 0;
- ctxt->nsTab = (const xmlChar **)
- xmlMalloc(ctxt->nsMax * sizeof(xmlChar *));
- if (ctxt->nsTab == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->nsMax = 0;
- return (-1);
- }
- } else if (ctxt->nsNr >= ctxt->nsMax) {
- const xmlChar ** tmp;
- ctxt->nsMax *= 2;
- tmp = (const xmlChar **) xmlRealloc((char *) ctxt->nsTab,
- ctxt->nsMax * sizeof(ctxt->nsTab[0]));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->nsMax /= 2;
- return (-1);
- }
- ctxt->nsTab = tmp;
- }
- ctxt->nsTab[ctxt->nsNr++] = prefix;
- ctxt->nsTab[ctxt->nsNr++] = URL;
- return (ctxt->nsNr);
- }
- /**
- * nsPop:
- * @ctxt: an XML parser context
- * @nr: the number to pop
- *
- * Pops the top @nr parser prefix/namespace from the ns stack
- *
- * Returns the number of namespaces removed
- */
- static int
- nsPop(xmlParserCtxtPtr ctxt, int nr)
- {
- int i;
- if (ctxt->nsTab == NULL) return(0);
- if (ctxt->nsNr < nr) {
- xmlGenericError(xmlGenericErrorContext, "Pbm popping %d NS\n", nr);
- nr = ctxt->nsNr;
- }
- if (ctxt->nsNr <= 0)
- return (0);
- for (i = 0;i < nr;i++) {
- ctxt->nsNr--;
- ctxt->nsTab[ctxt->nsNr] = NULL;
- }
- return(nr);
- }
- #endif
- static int
- xmlCtxtGrowAttrs(xmlParserCtxtPtr ctxt, int nr) {
- const xmlChar **atts;
- int *attallocs;
- int maxatts;
- if (ctxt->atts == NULL) {
- maxatts = 55; /* allow for 10 attrs by default */
- atts = (const xmlChar **)
- xmlMalloc(maxatts * sizeof(xmlChar *));
- if (atts == NULL) goto mem_error;
- ctxt->atts = atts;
- attallocs = (int *) xmlMalloc((maxatts / 5) * sizeof(int));
- if (attallocs == NULL) goto mem_error;
- ctxt->attallocs = attallocs;
- ctxt->maxatts = maxatts;
- } else if (nr + 5 > ctxt->maxatts) {
- maxatts = (nr + 5) * 2;
- atts = (const xmlChar **) xmlRealloc((void *) ctxt->atts,
- maxatts * sizeof(const xmlChar *));
- if (atts == NULL) goto mem_error;
- ctxt->atts = atts;
- attallocs = (int *) xmlRealloc((void *) ctxt->attallocs,
- (maxatts / 5) * sizeof(int));
- if (attallocs == NULL) goto mem_error;
- ctxt->attallocs = attallocs;
- ctxt->maxatts = maxatts;
- }
- return(ctxt->maxatts);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return(-1);
- }
- /**
- * inputPush:
- * @ctxt: an XML parser context
- * @value: the parser input
- *
- * Pushes a new parser input on top of the input stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- int
- inputPush(xmlParserCtxtPtr ctxt, xmlParserInputPtr value)
- {
- if ((ctxt == NULL) || (value == NULL))
- return(-1);
- if (ctxt->inputNr >= ctxt->inputMax) {
- ctxt->inputMax *= 2;
- ctxt->inputTab =
- (xmlParserInputPtr *) xmlRealloc(ctxt->inputTab,
- ctxt->inputMax *
- sizeof(ctxt->inputTab[0]));
- if (ctxt->inputTab == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFreeInputStream(value);
- ctxt->inputMax /= 2;
- value = NULL;
- return (-1);
- }
- }
- ctxt->inputTab[ctxt->inputNr] = value;
- ctxt->input = value;
- return (ctxt->inputNr++);
- }
- /**
- * inputPop:
- * @ctxt: an XML parser context
- *
- * Pops the top parser input from the input stack
- *
- * Returns the input just removed
- */
- xmlParserInputPtr
- inputPop(xmlParserCtxtPtr ctxt)
- {
- xmlParserInputPtr ret;
- if (ctxt == NULL)
- return(NULL);
- if (ctxt->inputNr <= 0)
- return (NULL);
- ctxt->inputNr--;
- if (ctxt->inputNr > 0)
- ctxt->input = ctxt->inputTab[ctxt->inputNr - 1];
- else
- ctxt->input = NULL;
- ret = ctxt->inputTab[ctxt->inputNr];
- ctxt->inputTab[ctxt->inputNr] = NULL;
- return (ret);
- }
- /**
- * nodePush:
- * @ctxt: an XML parser context
- * @value: the element node
- *
- * Pushes a new element node on top of the node stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- int
- nodePush(xmlParserCtxtPtr ctxt, xmlNodePtr value)
- {
- if (ctxt == NULL) return(0);
- if (ctxt->nodeNr >= ctxt->nodeMax) {
- xmlNodePtr *tmp;
- tmp = (xmlNodePtr *) xmlRealloc(ctxt->nodeTab,
- ctxt->nodeMax * 2 *
- sizeof(ctxt->nodeTab[0]));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- return (-1);
- }
- ctxt->nodeTab = tmp;
- ctxt->nodeMax *= 2;
- }
- if ((((unsigned int) ctxt->nodeNr) > xmlParserMaxDepth) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INTERNAL_ERROR,
- "Excessive depth in document: %d use XML_PARSE_HUGE option\n",
- xmlParserMaxDepth);
- xmlHaltParser(ctxt);
- return(-1);
- }
- ctxt->nodeTab[ctxt->nodeNr] = value;
- ctxt->node = value;
- return (ctxt->nodeNr++);
- }
- /**
- * nodePop:
- * @ctxt: an XML parser context
- *
- * Pops the top element node from the node stack
- *
- * Returns the node just removed
- */
- xmlNodePtr
- nodePop(xmlParserCtxtPtr ctxt)
- {
- xmlNodePtr ret;
- if (ctxt == NULL) return(NULL);
- if (ctxt->nodeNr <= 0)
- return (NULL);
- ctxt->nodeNr--;
- if (ctxt->nodeNr > 0)
- ctxt->node = ctxt->nodeTab[ctxt->nodeNr - 1];
- else
- ctxt->node = NULL;
- ret = ctxt->nodeTab[ctxt->nodeNr];
- ctxt->nodeTab[ctxt->nodeNr] = NULL;
- return (ret);
- }
- /**
- * nameNsPush:
- * @ctxt: an XML parser context
- * @value: the element name
- * @prefix: the element prefix
- * @URI: the element namespace name
- * @line: the current line number for error messages
- * @nsNr: the number of namespaces pushed on the namespace table
- *
- * Pushes a new element name/prefix/URL on top of the name stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- static int
- nameNsPush(xmlParserCtxtPtr ctxt, const xmlChar * value,
- const xmlChar *prefix, const xmlChar *URI, int line, int nsNr)
- {
- xmlStartTag *tag;
- if (ctxt->nameNr >= ctxt->nameMax) {
- const xmlChar * *tmp;
- xmlStartTag *tmp2;
- ctxt->nameMax *= 2;
- tmp = (const xmlChar * *) xmlRealloc((xmlChar * *)ctxt->nameTab,
- ctxt->nameMax *
- sizeof(ctxt->nameTab[0]));
- if (tmp == NULL) {
- ctxt->nameMax /= 2;
- goto mem_error;
- }
- ctxt->nameTab = tmp;
- tmp2 = (xmlStartTag *) xmlRealloc((void * *)ctxt->pushTab,
- ctxt->nameMax *
- sizeof(ctxt->pushTab[0]));
- if (tmp2 == NULL) {
- ctxt->nameMax /= 2;
- goto mem_error;
- }
- ctxt->pushTab = tmp2;
- } else if (ctxt->pushTab == NULL) {
- ctxt->pushTab = (xmlStartTag *) xmlMalloc(ctxt->nameMax *
- sizeof(ctxt->pushTab[0]));
- if (ctxt->pushTab == NULL)
- goto mem_error;
- }
- ctxt->nameTab[ctxt->nameNr] = value;
- ctxt->name = value;
- tag = &ctxt->pushTab[ctxt->nameNr];
- tag->prefix = prefix;
- tag->URI = URI;
- tag->line = line;
- tag->nsNr = nsNr;
- return (ctxt->nameNr++);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return (-1);
- }
- #ifdef LIBXML_PUSH_ENABLED
- /**
- * nameNsPop:
- * @ctxt: an XML parser context
- *
- * Pops the top element/prefix/URI name from the name stack
- *
- * Returns the name just removed
- */
- static const xmlChar *
- nameNsPop(xmlParserCtxtPtr ctxt)
- {
- const xmlChar *ret;
- if (ctxt->nameNr <= 0)
- return (NULL);
- ctxt->nameNr--;
- if (ctxt->nameNr > 0)
- ctxt->name = ctxt->nameTab[ctxt->nameNr - 1];
- else
- ctxt->name = NULL;
- ret = ctxt->nameTab[ctxt->nameNr];
- ctxt->nameTab[ctxt->nameNr] = NULL;
- return (ret);
- }
- #endif /* LIBXML_PUSH_ENABLED */
- /**
- * namePush:
- * @ctxt: an XML parser context
- * @value: the element name
- *
- * Pushes a new element name on top of the name stack
- *
- * Returns -1 in case of error, the index in the stack otherwise
- */
- int
- namePush(xmlParserCtxtPtr ctxt, const xmlChar * value)
- {
- if (ctxt == NULL) return (-1);
- if (ctxt->nameNr >= ctxt->nameMax) {
- const xmlChar * *tmp;
- tmp = (const xmlChar * *) xmlRealloc((xmlChar * *)ctxt->nameTab,
- ctxt->nameMax * 2 *
- sizeof(ctxt->nameTab[0]));
- if (tmp == NULL) {
- goto mem_error;
- }
- ctxt->nameTab = tmp;
- ctxt->nameMax *= 2;
- }
- ctxt->nameTab[ctxt->nameNr] = value;
- ctxt->name = value;
- return (ctxt->nameNr++);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- return (-1);
- }
- /**
- * namePop:
- * @ctxt: an XML parser context
- *
- * Pops the top element name from the name stack
- *
- * Returns the name just removed
- */
- const xmlChar *
- namePop(xmlParserCtxtPtr ctxt)
- {
- const xmlChar *ret;
- if ((ctxt == NULL) || (ctxt->nameNr <= 0))
- return (NULL);
- ctxt->nameNr--;
- if (ctxt->nameNr > 0)
- ctxt->name = ctxt->nameTab[ctxt->nameNr - 1];
- else
- ctxt->name = NULL;
- ret = ctxt->nameTab[ctxt->nameNr];
- ctxt->nameTab[ctxt->nameNr] = NULL;
- return (ret);
- }
- static int spacePush(xmlParserCtxtPtr ctxt, int val) {
- if (ctxt->spaceNr >= ctxt->spaceMax) {
- int *tmp;
- ctxt->spaceMax *= 2;
- tmp = (int *) xmlRealloc(ctxt->spaceTab,
- ctxt->spaceMax * sizeof(ctxt->spaceTab[0]));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->spaceMax /=2;
- return(-1);
- }
- ctxt->spaceTab = tmp;
- }
- ctxt->spaceTab[ctxt->spaceNr] = val;
- ctxt->space = &ctxt->spaceTab[ctxt->spaceNr];
- return(ctxt->spaceNr++);
- }
- static int spacePop(xmlParserCtxtPtr ctxt) {
- int ret;
- if (ctxt->spaceNr <= 0) return(0);
- ctxt->spaceNr--;
- if (ctxt->spaceNr > 0)
- ctxt->space = &ctxt->spaceTab[ctxt->spaceNr - 1];
- else
- ctxt->space = &ctxt->spaceTab[0];
- ret = ctxt->spaceTab[ctxt->spaceNr];
- ctxt->spaceTab[ctxt->spaceNr] = -1;
- return(ret);
- }
- /*
- * Macros for accessing the content. Those should be used only by the parser,
- * and not exported.
- *
- * Dirty macros, i.e. one often need to make assumption on the context to
- * use them
- *
- * CUR_PTR return the current pointer to the xmlChar to be parsed.
- * To be used with extreme caution since operations consuming
- * characters may move the input buffer to a different location !
- * CUR returns the current xmlChar value, i.e. a 8 bit value if compiled
- * This should be used internally by the parser
- * only to compare to ASCII values otherwise it would break when
- * running with UTF-8 encoding.
- * RAW same as CUR but in the input buffer, bypass any token
- * extraction that may have been done
- * NXT(n) returns the n'th next xmlChar. Same as CUR is should be used only
- * to compare on ASCII based substring.
- * SKIP(n) Skip n xmlChar, and must also be used only to skip ASCII defined
- * strings without newlines within the parser.
- * NEXT1(l) Skip 1 xmlChar, and must also be used only to skip 1 non-newline ASCII
- * defined char within the parser.
- * Clean macros, not dependent of an ASCII context, expect UTF-8 encoding
- *
- * NEXT Skip to the next character, this does the proper decoding
- * in UTF-8 mode. It also pop-up unfinished entities on the fly.
- * NEXTL(l) Skip the current unicode character of l xmlChars long.
- * CUR_CHAR(l) returns the current unicode character (int), set l
- * to the number of xmlChars used for the encoding [0-5].
- * CUR_SCHAR same but operate on a string instead of the context
- * COPY_BUF copy the current unicode char to the target buffer, increment
- * the index
- * GROW, SHRINK handling of input buffers
- */
- #define RAW (*ctxt->input->cur)
- #define CUR (*ctxt->input->cur)
- #define NXT(val) ctxt->input->cur[(val)]
- #define CUR_PTR ctxt->input->cur
- #define BASE_PTR ctxt->input->base
- #define CMP4( s, c1, c2, c3, c4 ) \
- ( ((unsigned char *) s)[ 0 ] == c1 && ((unsigned char *) s)[ 1 ] == c2 && \
- ((unsigned char *) s)[ 2 ] == c3 && ((unsigned char *) s)[ 3 ] == c4 )
- #define CMP5( s, c1, c2, c3, c4, c5 ) \
- ( CMP4( s, c1, c2, c3, c4 ) && ((unsigned char *) s)[ 4 ] == c5 )
- #define CMP6( s, c1, c2, c3, c4, c5, c6 ) \
- ( CMP5( s, c1, c2, c3, c4, c5 ) && ((unsigned char *) s)[ 5 ] == c6 )
- #define CMP7( s, c1, c2, c3, c4, c5, c6, c7 ) \
- ( CMP6( s, c1, c2, c3, c4, c5, c6 ) && ((unsigned char *) s)[ 6 ] == c7 )
- #define CMP8( s, c1, c2, c3, c4, c5, c6, c7, c8 ) \
- ( CMP7( s, c1, c2, c3, c4, c5, c6, c7 ) && ((unsigned char *) s)[ 7 ] == c8 )
- #define CMP9( s, c1, c2, c3, c4, c5, c6, c7, c8, c9 ) \
- ( CMP8( s, c1, c2, c3, c4, c5, c6, c7, c8 ) && \
- ((unsigned char *) s)[ 8 ] == c9 )
- #define CMP10( s, c1, c2, c3, c4, c5, c6, c7, c8, c9, c10 ) \
- ( CMP9( s, c1, c2, c3, c4, c5, c6, c7, c8, c9 ) && \
- ((unsigned char *) s)[ 9 ] == c10 )
- #define SKIP(val) do { \
- ctxt->input->cur += (val),ctxt->input->col+=(val); \
- if (*ctxt->input->cur == 0) \
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK); \
- } while (0)
- #define SKIPL(val) do { \
- int skipl; \
- for(skipl=0; skipl<val; skipl++) { \
- if (*(ctxt->input->cur) == '\n') { \
- ctxt->input->line++; ctxt->input->col = 1; \
- } else ctxt->input->col++; \
- ctxt->input->cur++; \
- } \
- if (*ctxt->input->cur == 0) \
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK); \
- } while (0)
- #define SHRINK if ((ctxt->progressive == 0) && \
- (ctxt->input->cur - ctxt->input->base > 2 * INPUT_CHUNK) && \
- (ctxt->input->end - ctxt->input->cur < 2 * INPUT_CHUNK)) \
- xmlSHRINK (ctxt);
- static void xmlSHRINK (xmlParserCtxtPtr ctxt) {
- xmlParserInputShrink(ctxt->input);
- if (*ctxt->input->cur == 0)
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- }
- #define GROW if ((ctxt->progressive == 0) && \
- (ctxt->input->end - ctxt->input->cur < INPUT_CHUNK)) \
- xmlGROW (ctxt);
- static void xmlGROW (xmlParserCtxtPtr ctxt) {
- ptrdiff_t curEnd = ctxt->input->end - ctxt->input->cur;
- ptrdiff_t curBase = ctxt->input->cur - ctxt->input->base;
- if (((curEnd > XML_MAX_LOOKUP_LIMIT) ||
- (curBase > XML_MAX_LOOKUP_LIMIT)) &&
- ((ctxt->input->buf) &&
- (ctxt->input->buf->readcallback != xmlInputReadCallbackNop)) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR, "Huge input lookup");
- xmlHaltParser(ctxt);
- return;
- }
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- if ((ctxt->input->cur > ctxt->input->end) ||
- (ctxt->input->cur < ctxt->input->base)) {
- xmlHaltParser(ctxt);
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR, "cur index out of bound");
- return;
- }
- if ((ctxt->input->cur != NULL) && (*ctxt->input->cur == 0))
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- }
- #define SKIP_BLANKS xmlSkipBlankChars(ctxt)
- #define NEXT xmlNextChar(ctxt)
- #define NEXT1 { \
- ctxt->input->col++; \
- ctxt->input->cur++; \
- if (*ctxt->input->cur == 0) \
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK); \
- }
- #define NEXTL(l) do { \
- if (*(ctxt->input->cur) == '\n') { \
- ctxt->input->line++; ctxt->input->col = 1; \
- } else ctxt->input->col++; \
- ctxt->input->cur += l; \
- } while (0)
- #define CUR_CHAR(l) xmlCurrentChar(ctxt, &l)
- #define CUR_SCHAR(s, l) xmlStringCurrentChar(ctxt, s, &l)
- #define COPY_BUF(l,b,i,v) \
- if (l == 1) b[i++] = (xmlChar) v; \
- else i += xmlCopyCharMultiByte(&b[i],v)
- /**
- * xmlSkipBlankChars:
- * @ctxt: the XML parser context
- *
- * skip all blanks character found at that point in the input streams.
- * It pops up finished entities in the process if allowable at that point.
- *
- * Returns the number of space chars skipped
- */
- int
- xmlSkipBlankChars(xmlParserCtxtPtr ctxt) {
- int res = 0;
- /*
- * It's Okay to use CUR/NEXT here since all the blanks are on
- * the ASCII range.
- */
- if (ctxt->instate != XML_PARSER_DTD) {
- const xmlChar *cur;
- /*
- * if we are in the document content, go really fast
- */
- cur = ctxt->input->cur;
- while (IS_BLANK_CH(*cur)) {
- if (*cur == '\n') {
- ctxt->input->line++; ctxt->input->col = 1;
- } else {
- ctxt->input->col++;
- }
- cur++;
- res++;
- if (*cur == 0) {
- ctxt->input->cur = cur;
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- cur = ctxt->input->cur;
- }
- }
- ctxt->input->cur = cur;
- } else {
- int expandPE = ((ctxt->external != 0) || (ctxt->inputNr != 1));
- while (1) {
- if (IS_BLANK_CH(CUR)) { /* CHECKED tstblanks.xml */
- NEXT;
- } else if (CUR == '%') {
- /*
- * Need to handle support of entities branching here
- */
- if ((expandPE == 0) || (IS_BLANK_CH(NXT(1))) || (NXT(1) == 0))
- break;
- xmlParsePEReference(ctxt);
- } else if (CUR == 0) {
- if (ctxt->inputNr <= 1)
- break;
- xmlPopInput(ctxt);
- } else {
- break;
- }
- /*
- * Also increase the counter when entering or exiting a PERef.
- * The spec says: "When a parameter-entity reference is recognized
- * in the DTD and included, its replacement text MUST be enlarged
- * by the attachment of one leading and one following space (#x20)
- * character."
- */
- res++;
- }
- }
- return(res);
- }
- /************************************************************************
- * *
- * Commodity functions to handle entities *
- * *
- ************************************************************************/
- /**
- * xmlPopInput:
- * @ctxt: an XML parser context
- *
- * xmlPopInput: the current input pointed by ctxt->input came to an end
- * pop it and return the next char.
- *
- * Returns the current xmlChar in the parser context
- */
- xmlChar
- xmlPopInput(xmlParserCtxtPtr ctxt) {
- if ((ctxt == NULL) || (ctxt->inputNr <= 1)) return(0);
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "Popping input %d\n", ctxt->inputNr);
- if ((ctxt->inputNr > 1) && (ctxt->inSubset == 0) &&
- (ctxt->instate != XML_PARSER_EOF))
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "Unfinished entity outside the DTD");
- xmlFreeInputStream(inputPop(ctxt));
- if (*ctxt->input->cur == 0)
- xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
- return(CUR);
- }
- /**
- * xmlPushInput:
- * @ctxt: an XML parser context
- * @input: an XML parser input fragment (entity, XML fragment ...).
- *
- * xmlPushInput: switch to a new input stream which is stacked on top
- * of the previous one(s).
- * Returns -1 in case of error or the index in the input stack
- */
- int
- xmlPushInput(xmlParserCtxtPtr ctxt, xmlParserInputPtr input) {
- int ret;
- if (input == NULL) return(-1);
- if (xmlParserDebugEntities) {
- if ((ctxt->input != NULL) && (ctxt->input->filename))
- xmlGenericError(xmlGenericErrorContext,
- "%s(%d): ", ctxt->input->filename,
- ctxt->input->line);
- xmlGenericError(xmlGenericErrorContext,
- "Pushing input %d : %.30s\n", ctxt->inputNr+1, input->cur);
- }
- if (((ctxt->inputNr > 40) && ((ctxt->options & XML_PARSE_HUGE) == 0)) ||
- (ctxt->inputNr > 1024)) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- while (ctxt->inputNr > 1)
- xmlFreeInputStream(inputPop(ctxt));
- return(-1);
- }
- ret = inputPush(ctxt, input);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- GROW;
- return(ret);
- }
- /**
- * xmlParseCharRef:
- * @ctxt: an XML parser context
- *
- * parse Reference declarations
- *
- * [66] CharRef ::= '&#' [0-9]+ ';' |
- * '&#x' [0-9a-fA-F]+ ';'
- *
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- *
- * Returns the value parsed (as an int), 0 in case of error
- */
- int
- xmlParseCharRef(xmlParserCtxtPtr ctxt) {
- int val = 0;
- int count = 0;
- /*
- * Using RAW/CUR/NEXT is okay since we are working on ASCII range here
- */
- if ((RAW == '&') && (NXT(1) == '#') &&
- (NXT(2) == 'x')) {
- SKIP(3);
- GROW;
- while (RAW != ';') { /* loop blocked by count */
- if (count++ > 20) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(0);
- }
- if ((RAW >= '0') && (RAW <= '9'))
- val = val * 16 + (CUR - '0');
- else if ((RAW >= 'a') && (RAW <= 'f') && (count < 20))
- val = val * 16 + (CUR - 'a') + 10;
- else if ((RAW >= 'A') && (RAW <= 'F') && (count < 20))
- val = val * 16 + (CUR - 'A') + 10;
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_HEX_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- NEXT;
- count++;
- }
- if (RAW == ';') {
- /* on purpose to avoid reentrancy problems with NEXT and SKIP */
- ctxt->input->col++;
- ctxt->input->cur++;
- }
- } else if ((RAW == '&') && (NXT(1) == '#')) {
- SKIP(2);
- GROW;
- while (RAW != ';') { /* loop blocked by count */
- if (count++ > 20) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(0);
- }
- if ((RAW >= '0') && (RAW <= '9'))
- val = val * 10 + (CUR - '0');
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_DEC_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- NEXT;
- count++;
- }
- if (RAW == ';') {
- /* on purpose to avoid reentrancy problems with NEXT and SKIP */
- ctxt->input->col++;
- ctxt->input->cur++;
- }
- } else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_CHARREF, NULL);
- }
- /*
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- */
- if (val >= 0x110000) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseCharRef: character reference out of bounds\n",
- val);
- } else if (IS_CHAR(val)) {
- return(val);
- } else {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseCharRef: invalid xmlChar value %d\n",
- val);
- }
- return(0);
- }
- /**
- * xmlParseStringCharRef:
- * @ctxt: an XML parser context
- * @str: a pointer to an index in the string
- *
- * parse Reference declarations, variant parsing from a string rather
- * than an an input flow.
- *
- * [66] CharRef ::= '&#' [0-9]+ ';' |
- * '&#x' [0-9a-fA-F]+ ';'
- *
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- *
- * Returns the value parsed (as an int), 0 in case of error, str will be
- * updated to the current value of the index
- */
- static int
- xmlParseStringCharRef(xmlParserCtxtPtr ctxt, const xmlChar **str) {
- const xmlChar *ptr;
- xmlChar cur;
- int val = 0;
- if ((str == NULL) || (*str == NULL)) return(0);
- ptr = *str;
- cur = *ptr;
- if ((cur == '&') && (ptr[1] == '#') && (ptr[2] == 'x')) {
- ptr += 3;
- cur = *ptr;
- while (cur != ';') { /* Non input consuming loop */
- if ((cur >= '0') && (cur <= '9'))
- val = val * 16 + (cur - '0');
- else if ((cur >= 'a') && (cur <= 'f'))
- val = val * 16 + (cur - 'a') + 10;
- else if ((cur >= 'A') && (cur <= 'F'))
- val = val * 16 + (cur - 'A') + 10;
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_HEX_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- ptr++;
- cur = *ptr;
- }
- if (cur == ';')
- ptr++;
- } else if ((cur == '&') && (ptr[1] == '#')){
- ptr += 2;
- cur = *ptr;
- while (cur != ';') { /* Non input consuming loops */
- if ((cur >= '0') && (cur <= '9'))
- val = val * 10 + (cur - '0');
- else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_DEC_CHARREF, NULL);
- val = 0;
- break;
- }
- if (val > 0x110000)
- val = 0x110000;
- ptr++;
- cur = *ptr;
- }
- if (cur == ';')
- ptr++;
- } else {
- xmlFatalErr(ctxt, XML_ERR_INVALID_CHARREF, NULL);
- return(0);
- }
- *str = ptr;
- /*
- * [ WFC: Legal Character ]
- * Characters referred to using character references must match the
- * production for Char.
- */
- if (val >= 0x110000) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseStringCharRef: character reference out of bounds\n",
- val);
- } else if (IS_CHAR(val)) {
- return(val);
- } else {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseStringCharRef: invalid xmlChar value %d\n",
- val);
- }
- return(0);
- }
- /**
- * xmlParserHandlePEReference:
- * @ctxt: the parser context
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive
- * reference to itself, either directly or indirectly.
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", ... ... The declaration of a parameter
- * entity must precede any reference to it...
- *
- * [ VC: Entity Declared ]
- * In a document with an external subset or external parameter entities
- * with "standalone='no'", ... ... The declaration of a parameter entity
- * must precede any reference to it...
- *
- * [ WFC: In DTD ]
- * Parameter-entity references may only appear in the DTD.
- * NOTE: misleading but this is handled.
- *
- * A PEReference may have been detected in the current input stream
- * the handling is done accordingly to
- * http://www.w3.org/TR/REC-xml#entproc
- * i.e.
- * - Included in literal in entity values
- * - Included as Parameter Entity reference within DTDs
- */
- void
- xmlParserHandlePEReference(xmlParserCtxtPtr ctxt) {
- switch(ctxt->instate) {
- case XML_PARSER_CDATA_SECTION:
- return;
- case XML_PARSER_COMMENT:
- return;
- case XML_PARSER_START_TAG:
- return;
- case XML_PARSER_END_TAG:
- return;
- case XML_PARSER_EOF:
- xmlFatalErr(ctxt, XML_ERR_PEREF_AT_EOF, NULL);
- return;
- case XML_PARSER_PROLOG:
- case XML_PARSER_START:
- case XML_PARSER_MISC:
- xmlFatalErr(ctxt, XML_ERR_PEREF_IN_PROLOG, NULL);
- return;
- case XML_PARSER_ENTITY_DECL:
- case XML_PARSER_CONTENT:
- case XML_PARSER_ATTRIBUTE_VALUE:
- case XML_PARSER_PI:
- case XML_PARSER_SYSTEM_LITERAL:
- case XML_PARSER_PUBLIC_LITERAL:
- /* we just ignore it there */
- return;
- case XML_PARSER_EPILOG:
- xmlFatalErr(ctxt, XML_ERR_PEREF_IN_EPILOG, NULL);
- return;
- case XML_PARSER_ENTITY_VALUE:
- /*
- * NOTE: in the case of entity values, we don't do the
- * substitution here since we need the literal
- * entity value to be able to save the internal
- * subset of the document.
- * This will be handled by xmlStringDecodeEntities
- */
- return;
- case XML_PARSER_DTD:
- /*
- * [WFC: Well-Formedness Constraint: PEs in Internal Subset]
- * In the internal DTD subset, parameter-entity references
- * can occur only where markup declarations can occur, not
- * within markup declarations.
- * In that case this is handled in xmlParseMarkupDecl
- */
- if ((ctxt->external == 0) && (ctxt->inputNr == 1))
- return;
- if (IS_BLANK_CH(NXT(1)) || NXT(1) == 0)
- return;
- break;
- case XML_PARSER_IGNORE:
- return;
- }
- xmlParsePEReference(ctxt);
- }
- /*
- * Macro used to grow the current buffer.
- * buffer##_size is expected to be a size_t
- * mem_error: is expected to handle memory allocation failures
- */
- #define growBuffer(buffer, n) { \
- xmlChar *tmp; \
- size_t new_size = buffer##_size * 2 + n; \
- if (new_size < buffer##_size) goto mem_error; \
- tmp = (xmlChar *) xmlRealloc(buffer, new_size); \
- if (tmp == NULL) goto mem_error; \
- buffer = tmp; \
- buffer##_size = new_size; \
- }
- /**
- * xmlStringLenDecodeEntities:
- * @ctxt: the parser context
- * @str: the input string
- * @len: the string length
- * @what: combination of XML_SUBSTITUTE_REF and XML_SUBSTITUTE_PEREF
- * @end: an end marker xmlChar, 0 if none
- * @end2: an end marker xmlChar, 0 if none
- * @end3: an end marker xmlChar, 0 if none
- *
- * Takes a entity string content and process to do the adequate substitutions.
- *
- * [67] Reference ::= EntityRef | CharRef
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * Returns A newly allocated string with the substitution done. The caller
- * must deallocate it !
- */
- xmlChar *
- xmlStringLenDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int len,
- int what, xmlChar end, xmlChar end2, xmlChar end3) {
- xmlChar *buffer = NULL;
- size_t buffer_size = 0;
- size_t nbchars = 0;
- xmlChar *current = NULL;
- xmlChar *rep = NULL;
- const xmlChar *last;
- xmlEntityPtr ent;
- int c,l;
- if ((ctxt == NULL) || (str == NULL) || (len < 0))
- return(NULL);
- last = str + len;
- if (((ctxt->depth > 40) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) ||
- (ctxt->depth > 1024)) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- return(NULL);
- }
- /*
- * allocate a translation buffer.
- */
- buffer_size = XML_PARSER_BIG_BUFFER_SIZE;
- buffer = (xmlChar *) xmlMallocAtomic(buffer_size);
- if (buffer == NULL) goto mem_error;
- /*
- * OK loop until we reach one of the ending char or a size limit.
- * we are operating on already parsed values.
- */
- if (str < last)
- c = CUR_SCHAR(str, l);
- else
- c = 0;
- while ((c != 0) && (c != end) && /* non input consuming loop */
- (c != end2) && (c != end3) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- if (c == 0) break;
- if ((c == '&') && (str[1] == '#')) {
- int val = xmlParseStringCharRef(ctxt, &str);
- if (val == 0)
- goto int_error;
- COPY_BUF(0,buffer,nbchars,val);
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- } else if ((c == '&') && (what & XML_SUBSTITUTE_REF)) {
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "String decoding Entity Reference: %.30s\n",
- str);
- ent = xmlParseStringEntityRef(ctxt, &str);
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- if (ent != NULL)
- ctxt->nbentities += ent->checked / 2;
- if ((ent != NULL) &&
- (ent->etype == XML_INTERNAL_PREDEFINED_ENTITY)) {
- if (ent->content != NULL) {
- COPY_BUF(0,buffer,nbchars,ent->content[0]);
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- } else {
- xmlFatalErrMsg(ctxt, XML_ERR_INTERNAL_ERROR,
- "predefined entity has no content\n");
- goto int_error;
- }
- } else if ((ent != NULL) && (ent->content != NULL)) {
- ctxt->depth++;
- rep = xmlStringDecodeEntities(ctxt, ent->content, what,
- 0, 0, 0);
- ctxt->depth--;
- if (rep == NULL) {
- ent->content[0] = 0;
- goto int_error;
- }
- current = rep;
- while (*current != 0) { /* non input consuming loop */
- buffer[nbchars++] = *current++;
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- if (xmlParserEntityCheck(ctxt, nbchars, ent, 0))
- goto int_error;
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- }
- xmlFree(rep);
- rep = NULL;
- } else if (ent != NULL) {
- int i = xmlStrlen(ent->name);
- const xmlChar *cur = ent->name;
- buffer[nbchars++] = '&';
- if (nbchars + i + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, i + XML_PARSER_BUFFER_SIZE);
- }
- for (;i > 0;i--)
- buffer[nbchars++] = *cur++;
- buffer[nbchars++] = ';';
- }
- } else if (c == '%' && (what & XML_SUBSTITUTE_PEREF)) {
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "String decoding PE Reference: %.30s\n", str);
- ent = xmlParseStringPEReference(ctxt, &str);
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- if (ent != NULL)
- ctxt->nbentities += ent->checked / 2;
- if (ent != NULL) {
- if (ent->content == NULL) {
- /*
- * Note: external parsed entities will not be loaded,
- * it is not required for a non-validating parser to
- * complete external PEReferences coming from the
- * internal subset
- */
- if (((ctxt->options & XML_PARSE_NOENT) != 0) ||
- ((ctxt->options & XML_PARSE_DTDVALID) != 0) ||
- (ctxt->validate != 0)) {
- xmlLoadEntityContent(ctxt, ent);
- } else {
- xmlWarningMsg(ctxt, XML_ERR_ENTITY_PROCESSING,
- "not validating will not read content for PE entity %s\n",
- ent->name, NULL);
- }
- }
- ctxt->depth++;
- rep = xmlStringDecodeEntities(ctxt, ent->content, what,
- 0, 0, 0);
- ctxt->depth--;
- if (rep == NULL) {
- if (ent->content != NULL)
- ent->content[0] = 0;
- goto int_error;
- }
- current = rep;
- while (*current != 0) { /* non input consuming loop */
- buffer[nbchars++] = *current++;
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- if (xmlParserEntityCheck(ctxt, nbchars, ent, 0))
- goto int_error;
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- }
- xmlFree(rep);
- rep = NULL;
- }
- } else {
- COPY_BUF(l,buffer,nbchars,c);
- str += l;
- if (nbchars + XML_PARSER_BUFFER_SIZE > buffer_size) {
- growBuffer(buffer, XML_PARSER_BUFFER_SIZE);
- }
- }
- if (str < last)
- c = CUR_SCHAR(str, l);
- else
- c = 0;
- }
- buffer[nbchars] = 0;
- return(buffer);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- int_error:
- if (rep != NULL)
- xmlFree(rep);
- if (buffer != NULL)
- xmlFree(buffer);
- return(NULL);
- }
- /**
- * xmlStringDecodeEntities:
- * @ctxt: the parser context
- * @str: the input string
- * @what: combination of XML_SUBSTITUTE_REF and XML_SUBSTITUTE_PEREF
- * @end: an end marker xmlChar, 0 if none
- * @end2: an end marker xmlChar, 0 if none
- * @end3: an end marker xmlChar, 0 if none
- *
- * Takes a entity string content and process to do the adequate substitutions.
- *
- * [67] Reference ::= EntityRef | CharRef
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * Returns A newly allocated string with the substitution done. The caller
- * must deallocate it !
- */
- xmlChar *
- xmlStringDecodeEntities(xmlParserCtxtPtr ctxt, const xmlChar *str, int what,
- xmlChar end, xmlChar end2, xmlChar end3) {
- if ((ctxt == NULL) || (str == NULL)) return(NULL);
- return(xmlStringLenDecodeEntities(ctxt, str, xmlStrlen(str), what,
- end, end2, end3));
- }
- /************************************************************************
- * *
- * Commodity functions, cleanup needed ? *
- * *
- ************************************************************************/
- /**
- * areBlanks:
- * @ctxt: an XML parser context
- * @str: a xmlChar *
- * @len: the size of @str
- * @blank_chars: we know the chars are blanks
- *
- * Is this a sequence of blank chars that one can ignore ?
- *
- * Returns 1 if ignorable 0 otherwise.
- */
- static int areBlanks(xmlParserCtxtPtr ctxt, const xmlChar *str, int len,
- int blank_chars) {
- int i, ret;
- xmlNodePtr lastChild;
- /*
- * Don't spend time trying to differentiate them, the same callback is
- * used !
- */
- if (ctxt->sax->ignorableWhitespace == ctxt->sax->characters)
- return(0);
- /*
- * Check for xml:space value.
- */
- if ((ctxt->space == NULL) || (*(ctxt->space) == 1) ||
- (*(ctxt->space) == -2))
- return(0);
- /*
- * Check that the string is made of blanks
- */
- if (blank_chars == 0) {
- for (i = 0;i < len;i++)
- if (!(IS_BLANK_CH(str[i]))) return(0);
- }
- /*
- * Look if the element is mixed content in the DTD if available
- */
- if (ctxt->node == NULL) return(0);
- if (ctxt->myDoc != NULL) {
- ret = xmlIsMixedElement(ctxt->myDoc, ctxt->node->name);
- if (ret == 0) return(1);
- if (ret == 1) return(0);
- }
- /*
- * Otherwise, heuristic :-\
- */
- if ((RAW != '<') && (RAW != 0xD)) return(0);
- if ((ctxt->node->children == NULL) &&
- (RAW == '<') && (NXT(1) == '/')) return(0);
- lastChild = xmlGetLastChild(ctxt->node);
- if (lastChild == NULL) {
- if ((ctxt->node->type != XML_ELEMENT_NODE) &&
- (ctxt->node->content != NULL)) return(0);
- } else if (xmlNodeIsText(lastChild))
- return(0);
- else if ((ctxt->node->children != NULL) &&
- (xmlNodeIsText(ctxt->node->children)))
- return(0);
- return(1);
- }
- /************************************************************************
- * *
- * Extra stuff for namespace support *
- * Relates to http://www.w3.org/TR/WD-xml-names *
- * *
- ************************************************************************/
- /**
- * xmlSplitQName:
- * @ctxt: an XML parser context
- * @name: an XML parser context
- * @prefix: a xmlChar **
- *
- * parse an UTF8 encoded XML qualified name string
- *
- * [NS 5] QName ::= (Prefix ':')? LocalPart
- *
- * [NS 6] Prefix ::= NCName
- *
- * [NS 7] LocalPart ::= NCName
- *
- * Returns the local part, and prefix is updated
- * to get the Prefix if any.
- */
- xmlChar *
- xmlSplitQName(xmlParserCtxtPtr ctxt, const xmlChar *name, xmlChar **prefix) {
- xmlChar buf[XML_MAX_NAMELEN + 5];
- xmlChar *buffer = NULL;
- int len = 0;
- int max = XML_MAX_NAMELEN;
- xmlChar *ret = NULL;
- const xmlChar *cur = name;
- int c;
- if (prefix == NULL) return(NULL);
- *prefix = NULL;
- if (cur == NULL) return(NULL);
- #ifndef XML_XML_NAMESPACE
- /* xml: prefix is not really a namespace */
- if ((cur[0] == 'x') && (cur[1] == 'm') &&
- (cur[2] == 'l') && (cur[3] == ':'))
- return(xmlStrdup(name));
- #endif
- /* nasty but well=formed */
- if (cur[0] == ':')
- return(xmlStrdup(name));
- c = *cur++;
- while ((c != 0) && (c != ':') && (len < max)) { /* tested bigname.xml */
- buf[len++] = c;
- c = *cur++;
- }
- if (len >= max) {
- /*
- * Okay someone managed to make a huge name, so he's ready to pay
- * for the processing speed.
- */
- max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while ((c != 0) && (c != ':')) { /* tested bigname.xml */
- if (len + 10 > max) {
- xmlChar *tmp;
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buffer);
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- buffer = tmp;
- }
- buffer[len++] = c;
- c = *cur++;
- }
- buffer[len] = 0;
- }
- if ((c == ':') && (*cur == 0)) {
- if (buffer != NULL)
- xmlFree(buffer);
- *prefix = NULL;
- return(xmlStrdup(name));
- }
- if (buffer == NULL)
- ret = xmlStrndup(buf, len);
- else {
- ret = buffer;
- buffer = NULL;
- max = XML_MAX_NAMELEN;
- }
- if (c == ':') {
- c = *cur;
- *prefix = ret;
- if (c == 0) {
- return(xmlStrndup(BAD_CAST "", 0));
- }
- len = 0;
- /*
- * Check that the first character is proper to start
- * a new name
- */
- if (!(((c >= 0x61) && (c <= 0x7A)) ||
- ((c >= 0x41) && (c <= 0x5A)) ||
- (c == '_') || (c == ':'))) {
- int l;
- int first = CUR_SCHAR(cur, l);
- if (!IS_LETTER(first) && (first != '_')) {
- xmlFatalErrMsgStr(ctxt, XML_NS_ERR_QNAME,
- "Name %s is not XML Namespace compliant\n",
- name);
- }
- }
- cur++;
- while ((c != 0) && (len < max)) { /* tested bigname2.xml */
- buf[len++] = c;
- c = *cur++;
- }
- if (len >= max) {
- /*
- * Okay someone managed to make a huge name, so he's ready to pay
- * for the processing speed.
- */
- max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while (c != 0) { /* tested bigname2.xml */
- if (len + 10 > max) {
- xmlChar *tmp;
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buffer);
- return(NULL);
- }
- buffer = tmp;
- }
- buffer[len++] = c;
- c = *cur++;
- }
- buffer[len] = 0;
- }
- if (buffer == NULL)
- ret = xmlStrndup(buf, len);
- else {
- ret = buffer;
- }
- }
- return(ret);
- }
- /************************************************************************
- * *
- * The parser itself *
- * Relates to http://www.w3.org/TR/REC-xml *
- * *
- ************************************************************************/
- /************************************************************************
- * *
- * Routines to parse Name, NCName and NmToken *
- * *
- ************************************************************************/
- #ifdef DEBUG
- static unsigned long nbParseName = 0;
- static unsigned long nbParseNmToken = 0;
- static unsigned long nbParseNCName = 0;
- static unsigned long nbParseNCNameComplex = 0;
- static unsigned long nbParseNameComplex = 0;
- static unsigned long nbParseStringName = 0;
- #endif
- /*
- * The two following functions are related to the change of accepted
- * characters for Name and NmToken in the Revision 5 of XML-1.0
- * They correspond to the modified production [4] and the new production [4a]
- * changes in that revision. Also note that the macros used for the
- * productions Letter, Digit, CombiningChar and Extender are not needed
- * anymore.
- * We still keep compatibility to pre-revision5 parsing semantic if the
- * new XML_PARSE_OLD10 option is given to the parser.
- */
- static int
- xmlIsNameStartChar(xmlParserCtxtPtr ctxt, int c) {
- if ((ctxt->options & XML_PARSE_OLD10) == 0) {
- /*
- * Use the new checks of production [4] [4a] amd [5] of the
- * Update 5 of XML-1.0
- */
- if ((c != ' ') && (c != '>') && (c != '/') && /* accelerators */
- (((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- (c == '_') || (c == ':') ||
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))))
- return(1);
- } else {
- if (IS_LETTER(c) || (c == '_') || (c == ':'))
- return(1);
- }
- return(0);
- }
- static int
- xmlIsNameChar(xmlParserCtxtPtr ctxt, int c) {
- if ((ctxt->options & XML_PARSE_OLD10) == 0) {
- /*
- * Use the new checks of production [4] [4a] amd [5] of the
- * Update 5 of XML-1.0
- */
- if ((c != ' ') && (c != '>') && (c != '/') && /* accelerators */
- (((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- ((c >= '0') && (c <= '9')) || /* !start */
- (c == '_') || (c == ':') ||
- (c == '-') || (c == '.') || (c == 0xB7) || /* !start */
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x300) && (c <= 0x36F)) || /* !start */
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x203F) && (c <= 0x2040)) || /* !start */
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))))
- return(1);
- } else {
- if ((IS_LETTER(c)) || (IS_DIGIT(c)) ||
- (c == '.') || (c == '-') ||
- (c == '_') || (c == ':') ||
- (IS_COMBINING(c)) ||
- (IS_EXTENDER(c)))
- return(1);
- }
- return(0);
- }
- static xmlChar * xmlParseAttValueInternal(xmlParserCtxtPtr ctxt,
- int *len, int *alloc, int normalize);
- static const xmlChar *
- xmlParseNameComplex(xmlParserCtxtPtr ctxt) {
- int len = 0, l;
- int c;
- int count = 0;
- #ifdef DEBUG
- nbParseNameComplex++;
- #endif
- /*
- * Handler for more complex cases
- */
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- c = CUR_CHAR(l);
- if ((ctxt->options & XML_PARSE_OLD10) == 0) {
- /*
- * Use the new checks of production [4] [4a] amd [5] of the
- * Update 5 of XML-1.0
- */
- if ((c == ' ') || (c == '>') || (c == '/') || /* accelerators */
- (!(((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- (c == '_') || (c == ':') ||
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))))) {
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- while ((c != ' ') && (c != '>') && (c != '/') && /* accelerators */
- (((c >= 'a') && (c <= 'z')) ||
- ((c >= 'A') && (c <= 'Z')) ||
- ((c >= '0') && (c <= '9')) || /* !start */
- (c == '_') || (c == ':') ||
- (c == '-') || (c == '.') || (c == 0xB7) || /* !start */
- ((c >= 0xC0) && (c <= 0xD6)) ||
- ((c >= 0xD8) && (c <= 0xF6)) ||
- ((c >= 0xF8) && (c <= 0x2FF)) ||
- ((c >= 0x300) && (c <= 0x36F)) || /* !start */
- ((c >= 0x370) && (c <= 0x37D)) ||
- ((c >= 0x37F) && (c <= 0x1FFF)) ||
- ((c >= 0x200C) && (c <= 0x200D)) ||
- ((c >= 0x203F) && (c <= 0x2040)) || /* !start */
- ((c >= 0x2070) && (c <= 0x218F)) ||
- ((c >= 0x2C00) && (c <= 0x2FEF)) ||
- ((c >= 0x3001) && (c <= 0xD7FF)) ||
- ((c >= 0xF900) && (c <= 0xFDCF)) ||
- ((c >= 0xFDF0) && (c <= 0xFFFD)) ||
- ((c >= 0x10000) && (c <= 0xEFFFF))
- )) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- }
- } else {
- if ((c == ' ') || (c == '>') || (c == '/') || /* accelerators */
- (!IS_LETTER(c) && (c != '_') &&
- (c != ':'))) {
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- while ((c != ' ') && (c != '>') && (c != '/') && /* test bigname.xml */
- ((IS_LETTER(c)) || (IS_DIGIT(c)) ||
- (c == '.') || (c == '-') ||
- (c == '_') || (c == ':') ||
- (IS_COMBINING(c)) ||
- (IS_EXTENDER(c)))) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- }
- }
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "Name");
- return(NULL);
- }
- if (ctxt->input->cur - ctxt->input->base < len) {
- /*
- * There were a couple of bugs where PERefs lead to to a change
- * of the buffer. Check the buffer size to avoid passing an invalid
- * pointer to xmlDictLookup.
- */
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "unexpected change of input buffer");
- return (NULL);
- }
- if ((*ctxt->input->cur == '\n') && (ctxt->input->cur[-1] == '\r'))
- return(xmlDictLookup(ctxt->dict, ctxt->input->cur - (len + 1), len));
- return(xmlDictLookup(ctxt->dict, ctxt->input->cur - len, len));
- }
- /**
- * xmlParseName:
- * @ctxt: an XML parser context
- *
- * parse an XML name.
- *
- * [4] NameChar ::= Letter | Digit | '.' | '-' | '_' | ':' |
- * CombiningChar | Extender
- *
- * [5] Name ::= (Letter | '_' | ':') (NameChar)*
- *
- * [6] Names ::= Name (#x20 Name)*
- *
- * Returns the Name parsed or NULL
- */
- const xmlChar *
- xmlParseName(xmlParserCtxtPtr ctxt) {
- const xmlChar *in;
- const xmlChar *ret;
- int count = 0;
- GROW;
- #ifdef DEBUG
- nbParseName++;
- #endif
- /*
- * Accelerator for simple ASCII names
- */
- in = ctxt->input->cur;
- if (((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- (*in == '_') || (*in == ':')) {
- in++;
- while (((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- ((*in >= 0x30) && (*in <= 0x39)) ||
- (*in == '_') || (*in == '-') ||
- (*in == ':') || (*in == '.'))
- in++;
- if ((*in > 0) && (*in < 0x80)) {
- count = in - ctxt->input->cur;
- if ((count > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "Name");
- return(NULL);
- }
- ret = xmlDictLookup(ctxt->dict, ctxt->input->cur, count);
- ctxt->input->cur = in;
- ctxt->input->col += count;
- if (ret == NULL)
- xmlErrMemory(ctxt, NULL);
- return(ret);
- }
- }
- /* accelerator for special cases */
- return(xmlParseNameComplex(ctxt));
- }
- static const xmlChar *
- xmlParseNCNameComplex(xmlParserCtxtPtr ctxt) {
- int len = 0, l;
- int c;
- int count = 0;
- size_t startPosition = 0;
- #ifdef DEBUG
- nbParseNCNameComplex++;
- #endif
- /*
- * Handler for more complex cases
- */
- GROW;
- startPosition = CUR_PTR - BASE_PTR;
- c = CUR_CHAR(l);
- if ((c == ' ') || (c == '>') || (c == '/') || /* accelerators */
- (!xmlIsNameStartChar(ctxt, c) || (c == ':'))) {
- return(NULL);
- }
- while ((c != ' ') && (c != '>') && (c != '/') && /* test bigname.xml */
- (xmlIsNameChar(ctxt, c) && (c != ':'))) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- }
- len += l;
- NEXTL(l);
- c = CUR_CHAR(l);
- if (c == 0) {
- count = 0;
- /*
- * when shrinking to extend the buffer we really need to preserve
- * the part of the name we already parsed. Hence rolling back
- * by current length.
- */
- ctxt->input->cur -= l;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- ctxt->input->cur += l;
- c = CUR_CHAR(l);
- }
- }
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- return(xmlDictLookup(ctxt->dict, (BASE_PTR + startPosition), len));
- }
- /**
- * xmlParseNCName:
- * @ctxt: an XML parser context
- * @len: length of the string parsed
- *
- * parse an XML name.
- *
- * [4NS] NCNameChar ::= Letter | Digit | '.' | '-' | '_' |
- * CombiningChar | Extender
- *
- * [5NS] NCName ::= (Letter | '_') (NCNameChar)*
- *
- * Returns the Name parsed or NULL
- */
- static const xmlChar *
- xmlParseNCName(xmlParserCtxtPtr ctxt) {
- const xmlChar *in, *e;
- const xmlChar *ret;
- int count = 0;
- #ifdef DEBUG
- nbParseNCName++;
- #endif
- /*
- * Accelerator for simple ASCII names
- */
- in = ctxt->input->cur;
- e = ctxt->input->end;
- if ((((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- (*in == '_')) && (in < e)) {
- in++;
- while ((((*in >= 0x61) && (*in <= 0x7A)) ||
- ((*in >= 0x41) && (*in <= 0x5A)) ||
- ((*in >= 0x30) && (*in <= 0x39)) ||
- (*in == '_') || (*in == '-') ||
- (*in == '.')) && (in < e))
- in++;
- if (in >= e)
- goto complex;
- if ((*in > 0) && (*in < 0x80)) {
- count = in - ctxt->input->cur;
- if ((count > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- ret = xmlDictLookup(ctxt->dict, ctxt->input->cur, count);
- ctxt->input->cur = in;
- ctxt->input->col += count;
- if (ret == NULL) {
- xmlErrMemory(ctxt, NULL);
- }
- return(ret);
- }
- }
- complex:
- return(xmlParseNCNameComplex(ctxt));
- }
- /**
- * xmlParseNameAndCompare:
- * @ctxt: an XML parser context
- *
- * parse an XML name and compares for match
- * (specialized for endtag parsing)
- *
- * Returns NULL for an illegal name, (xmlChar*) 1 for success
- * and the name for mismatch
- */
- static const xmlChar *
- xmlParseNameAndCompare(xmlParserCtxtPtr ctxt, xmlChar const *other) {
- register const xmlChar *cmp = other;
- register const xmlChar *in;
- const xmlChar *ret;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- in = ctxt->input->cur;
- while (*in != 0 && *in == *cmp) {
- ++in;
- ++cmp;
- }
- if (*cmp == 0 && (*in == '>' || IS_BLANK_CH (*in))) {
- /* success */
- ctxt->input->col += in - ctxt->input->cur;
- ctxt->input->cur = in;
- return (const xmlChar*) 1;
- }
- /* failure (or end of input buffer), check with full function */
- ret = xmlParseName (ctxt);
- /* strings coming from the dictionary direct compare possible */
- if (ret == other) {
- return (const xmlChar*) 1;
- }
- return ret;
- }
- /**
- * xmlParseStringName:
- * @ctxt: an XML parser context
- * @str: a pointer to the string pointer (IN/OUT)
- *
- * parse an XML name.
- *
- * [4] NameChar ::= Letter | Digit | '.' | '-' | '_' | ':' |
- * CombiningChar | Extender
- *
- * [5] Name ::= (Letter | '_' | ':') (NameChar)*
- *
- * [6] Names ::= Name (#x20 Name)*
- *
- * Returns the Name parsed or NULL. The @str pointer
- * is updated to the current location in the string.
- */
- static xmlChar *
- xmlParseStringName(xmlParserCtxtPtr ctxt, const xmlChar** str) {
- xmlChar buf[XML_MAX_NAMELEN + 5];
- const xmlChar *cur = *str;
- int len = 0, l;
- int c;
- #ifdef DEBUG
- nbParseStringName++;
- #endif
- c = CUR_SCHAR(cur, l);
- if (!xmlIsNameStartChar(ctxt, c)) {
- return(NULL);
- }
- COPY_BUF(l,buf,len,c);
- cur += l;
- c = CUR_SCHAR(cur, l);
- while (xmlIsNameChar(ctxt, c)) {
- COPY_BUF(l,buf,len,c);
- cur += l;
- c = CUR_SCHAR(cur, l);
- if (len >= XML_MAX_NAMELEN) { /* test bigentname.xml */
- /*
- * Okay someone managed to make a huge name, so he's ready to pay
- * for the processing speed.
- */
- xmlChar *buffer;
- int max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while (xmlIsNameChar(ctxt, c)) {
- if (len + 10 > max) {
- xmlChar *tmp;
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- xmlFree(buffer);
- return(NULL);
- }
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buffer);
- return(NULL);
- }
- buffer = tmp;
- }
- COPY_BUF(l,buffer,len,c);
- cur += l;
- c = CUR_SCHAR(cur, l);
- }
- buffer[len] = 0;
- *str = cur;
- return(buffer);
- }
- }
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NCName");
- return(NULL);
- }
- *str = cur;
- return(xmlStrndup(buf, len));
- }
- /**
- * xmlParseNmtoken:
- * @ctxt: an XML parser context
- *
- * parse an XML Nmtoken.
- *
- * [7] Nmtoken ::= (NameChar)+
- *
- * [8] Nmtokens ::= Nmtoken (#x20 Nmtoken)*
- *
- * Returns the Nmtoken parsed or NULL
- */
- xmlChar *
- xmlParseNmtoken(xmlParserCtxtPtr ctxt) {
- xmlChar buf[XML_MAX_NAMELEN + 5];
- int len = 0, l;
- int c;
- int count = 0;
- #ifdef DEBUG
- nbParseNmToken++;
- #endif
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- c = CUR_CHAR(l);
- while (xmlIsNameChar(ctxt, c)) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- }
- COPY_BUF(l,buf,len,c);
- NEXTL(l);
- c = CUR_CHAR(l);
- if (c == 0) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- c = CUR_CHAR(l);
- }
- if (len >= XML_MAX_NAMELEN) {
- /*
- * Okay someone managed to make a huge token, so he's ready to pay
- * for the processing speed.
- */
- xmlChar *buffer;
- int max = len * 2;
- buffer = (xmlChar *) xmlMallocAtomic(max * sizeof(xmlChar));
- if (buffer == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- memcpy(buffer, buf, len);
- while (xmlIsNameChar(ctxt, c)) {
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buffer);
- return(NULL);
- }
- }
- if (len + 10 > max) {
- xmlChar *tmp;
- if ((max > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NmToken");
- xmlFree(buffer);
- return(NULL);
- }
- max *= 2;
- tmp = (xmlChar *) xmlRealloc(buffer,
- max * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buffer);
- return(NULL);
- }
- buffer = tmp;
- }
- COPY_BUF(l,buffer,len,c);
- NEXTL(l);
- c = CUR_CHAR(l);
- }
- buffer[len] = 0;
- return(buffer);
- }
- }
- if (len == 0)
- return(NULL);
- if ((len > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "NmToken");
- return(NULL);
- }
- return(xmlStrndup(buf, len));
- }
- /**
- * xmlParseEntityValue:
- * @ctxt: an XML parser context
- * @orig: if non-NULL store a copy of the original entity value
- *
- * parse a value for ENTITY declarations
- *
- * [9] EntityValue ::= '"' ([^%&"] | PEReference | Reference)* '"' |
- * "'" ([^%&'] | PEReference | Reference)* "'"
- *
- * Returns the EntityValue parsed with reference substituted or NULL
- */
- xmlChar *
- xmlParseEntityValue(xmlParserCtxtPtr ctxt, xmlChar **orig) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- int c, l;
- xmlChar stop;
- xmlChar *ret = NULL;
- const xmlChar *cur = NULL;
- xmlParserInputPtr input;
- if (RAW == '"') stop = '"';
- else if (RAW == '\'') stop = '\'';
- else {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_NOT_STARTED, NULL);
- return(NULL);
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- /*
- * The content of the entity definition is copied in a buffer.
- */
- ctxt->instate = XML_PARSER_ENTITY_VALUE;
- input = ctxt->input;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- goto error;
- NEXT;
- c = CUR_CHAR(l);
- /*
- * NOTE: 4.4.5 Included in Literal
- * When a parameter entity reference appears in a literal entity
- * value, ... a single or double quote character in the replacement
- * text is always treated as a normal data character and will not
- * terminate the literal.
- * In practice it means we stop the loop only when back at parsing
- * the initial entity and the quote is found
- */
- while (((IS_CHAR(c)) && ((c != stop) || /* checked */
- (ctxt->input != input))) && (ctxt->instate != XML_PARSER_EOF)) {
- if (len + 5 >= size) {
- xmlChar *tmp;
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- goto error;
- }
- buf = tmp;
- }
- COPY_BUF(l,buf,len,c);
- NEXTL(l);
- GROW;
- c = CUR_CHAR(l);
- if (c == 0) {
- GROW;
- c = CUR_CHAR(l);
- }
- }
- buf[len] = 0;
- if (ctxt->instate == XML_PARSER_EOF)
- goto error;
- if (c != stop) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_NOT_FINISHED, NULL);
- goto error;
- }
- NEXT;
- /*
- * Raise problem w.r.t. '&' and '%' being used in non-entities
- * reference constructs. Note Charref will be handled in
- * xmlStringDecodeEntities()
- */
- cur = buf;
- while (*cur != 0) { /* non input consuming */
- if ((*cur == '%') || ((*cur == '&') && (cur[1] != '#'))) {
- xmlChar *name;
- xmlChar tmp = *cur;
- int nameOk = 0;
- cur++;
- name = xmlParseStringName(ctxt, &cur);
- if (name != NULL) {
- nameOk = 1;
- xmlFree(name);
- }
- if ((nameOk == 0) || (*cur != ';')) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_ENTITY_CHAR_ERROR,
- "EntityValue: '%c' forbidden except for entities references\n",
- tmp);
- goto error;
- }
- if ((tmp == '%') && (ctxt->inSubset == 1) &&
- (ctxt->inputNr == 1)) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_PE_INTERNAL, NULL);
- goto error;
- }
- if (*cur == 0)
- break;
- }
- cur++;
- }
- /*
- * Then PEReference entities are substituted.
- *
- * NOTE: 4.4.7 Bypassed
- * When a general entity reference appears in the EntityValue in
- * an entity declaration, it is bypassed and left as is.
- * so XML_SUBSTITUTE_REF is not set here.
- */
- ++ctxt->depth;
- ret = xmlStringDecodeEntities(ctxt, buf, XML_SUBSTITUTE_PEREF,
- 0, 0, 0);
- --ctxt->depth;
- if (orig != NULL) {
- *orig = buf;
- buf = NULL;
- }
- error:
- if (buf != NULL)
- xmlFree(buf);
- return(ret);
- }
- /**
- * xmlParseAttValueComplex:
- * @ctxt: an XML parser context
- * @len: the resulting attribute len
- * @normalize: whether to apply the inner normalization
- *
- * parse a value for an attribute, this is the fallback function
- * of xmlParseAttValue() when the attribute parsing requires handling
- * of non-ASCII characters, or normalization compaction.
- *
- * Returns the AttValue parsed or NULL. The value has to be freed by the caller.
- */
- static xmlChar *
- xmlParseAttValueComplex(xmlParserCtxtPtr ctxt, int *attlen, int normalize) {
- xmlChar limit = 0;
- xmlChar *buf = NULL;
- xmlChar *rep = NULL;
- size_t len = 0;
- size_t buf_size = 0;
- int c, l, in_space = 0;
- xmlChar *current = NULL;
- xmlEntityPtr ent;
- if (NXT(0) == '"') {
- ctxt->instate = XML_PARSER_ATTRIBUTE_VALUE;
- limit = '"';
- NEXT;
- } else if (NXT(0) == '\'') {
- limit = '\'';
- ctxt->instate = XML_PARSER_ATTRIBUTE_VALUE;
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_ATTRIBUTE_NOT_STARTED, NULL);
- return(NULL);
- }
- /*
- * allocate a translation buffer.
- */
- buf_size = XML_PARSER_BUFFER_SIZE;
- buf = (xmlChar *) xmlMallocAtomic(buf_size);
- if (buf == NULL) goto mem_error;
- /*
- * OK loop until we reach one of the ending char or a size limit.
- */
- c = CUR_CHAR(l);
- while (((NXT(0) != limit) && /* checked */
- (IS_CHAR(c)) && (c != '<')) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- /*
- * Impose a reasonable limit on attribute size, unless XML_PARSE_HUGE
- * special option is given
- */
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- goto mem_error;
- }
- if (c == '&') {
- in_space = 0;
- if (NXT(1) == '#') {
- int val = xmlParseCharRef(ctxt);
- if (val == '&') {
- if (ctxt->replaceEntities) {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- buf[len++] = '&';
- } else {
- /*
- * The reparsing will be done in xmlStringGetNodeList()
- * called by the attribute() function in SAX.c
- */
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- buf[len++] = '&';
- buf[len++] = '#';
- buf[len++] = '3';
- buf[len++] = '8';
- buf[len++] = ';';
- }
- } else if (val != 0) {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- len += xmlCopyChar(0, &buf[len], val);
- }
- } else {
- ent = xmlParseEntityRef(ctxt);
- ctxt->nbentities++;
- if (ent != NULL)
- ctxt->nbentities += ent->owner;
- if ((ent != NULL) &&
- (ent->etype == XML_INTERNAL_PREDEFINED_ENTITY)) {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- if ((ctxt->replaceEntities == 0) &&
- (ent->content[0] == '&')) {
- buf[len++] = '&';
- buf[len++] = '#';
- buf[len++] = '3';
- buf[len++] = '8';
- buf[len++] = ';';
- } else {
- buf[len++] = ent->content[0];
- }
- } else if ((ent != NULL) &&
- (ctxt->replaceEntities != 0)) {
- if (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) {
- ++ctxt->depth;
- rep = xmlStringDecodeEntities(ctxt, ent->content,
- XML_SUBSTITUTE_REF,
- 0, 0, 0);
- --ctxt->depth;
- if (rep != NULL) {
- current = rep;
- while (*current != 0) { /* non input consuming */
- if ((*current == 0xD) || (*current == 0xA) ||
- (*current == 0x9)) {
- buf[len++] = 0x20;
- current++;
- } else
- buf[len++] = *current++;
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- }
- xmlFree(rep);
- rep = NULL;
- }
- } else {
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- if (ent->content != NULL)
- buf[len++] = ent->content[0];
- }
- } else if (ent != NULL) {
- int i = xmlStrlen(ent->name);
- const xmlChar *cur = ent->name;
- /*
- * This may look absurd but is needed to detect
- * entities problems
- */
- if ((ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
- (ent->content != NULL) && (ent->checked == 0)) {
- unsigned long oldnbent = ctxt->nbentities, diff;
- ++ctxt->depth;
- rep = xmlStringDecodeEntities(ctxt, ent->content,
- XML_SUBSTITUTE_REF, 0, 0, 0);
- --ctxt->depth;
- diff = ctxt->nbentities - oldnbent + 1;
- if (diff > INT_MAX / 2)
- diff = INT_MAX / 2;
- ent->checked = diff * 2;
- if (rep != NULL) {
- if (xmlStrchr(rep, '<'))
- ent->checked |= 1;
- xmlFree(rep);
- rep = NULL;
- } else {
- ent->content[0] = 0;
- }
- }
- /*
- * Just output the reference
- */
- buf[len++] = '&';
- while (len + i + 10 > buf_size) {
- growBuffer(buf, i + 10);
- }
- for (;i > 0;i--)
- buf[len++] = *cur++;
- buf[len++] = ';';
- }
- }
- } else {
- if ((c == 0x20) || (c == 0xD) || (c == 0xA) || (c == 0x9)) {
- if ((len != 0) || (!normalize)) {
- if ((!normalize) || (!in_space)) {
- COPY_BUF(l,buf,len,0x20);
- while (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- }
- in_space = 1;
- }
- } else {
- in_space = 0;
- COPY_BUF(l,buf,len,c);
- if (len + 10 > buf_size) {
- growBuffer(buf, 10);
- }
- }
- NEXTL(l);
- }
- GROW;
- c = CUR_CHAR(l);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto error;
- if ((in_space) && (normalize)) {
- while ((len > 0) && (buf[len - 1] == 0x20)) len--;
- }
- buf[len] = 0;
- if (RAW == '<') {
- xmlFatalErr(ctxt, XML_ERR_LT_IN_ATTRIBUTE, NULL);
- } else if (RAW != limit) {
- if ((c != 0) && (!IS_CHAR(c))) {
- xmlFatalErrMsg(ctxt, XML_ERR_INVALID_CHAR,
- "invalid character in attribute value\n");
- } else {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue: ' expected\n");
- }
- } else
- NEXT;
- /*
- * There we potentially risk an overflow, don't allow attribute value of
- * length more than INT_MAX it is a very reasonable assumption !
- */
- if (len >= INT_MAX) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- goto mem_error;
- }
- if (attlen != NULL) *attlen = (int) len;
- return(buf);
- mem_error:
- xmlErrMemory(ctxt, NULL);
- error:
- if (buf != NULL)
- xmlFree(buf);
- if (rep != NULL)
- xmlFree(rep);
- return(NULL);
- }
- /**
- * xmlParseAttValue:
- * @ctxt: an XML parser context
- *
- * parse a value for an attribute
- * Note: the parser won't do substitution of entities here, this
- * will be handled later in xmlStringGetNodeList
- *
- * [10] AttValue ::= '"' ([^<&"] | Reference)* '"' |
- * "'" ([^<&'] | Reference)* "'"
- *
- * 3.3.3 Attribute-Value Normalization:
- * Before the value of an attribute is passed to the application or
- * checked for validity, the XML processor must normalize it as follows:
- * - a character reference is processed by appending the referenced
- * character to the attribute value
- * - an entity reference is processed by recursively processing the
- * replacement text of the entity
- * - a whitespace character (#x20, #xD, #xA, #x9) is processed by
- * appending #x20 to the normalized value, except that only a single
- * #x20 is appended for a "#xD#xA" sequence that is part of an external
- * parsed entity or the literal entity value of an internal parsed entity
- * - other characters are processed by appending them to the normalized value
- * If the declared value is not CDATA, then the XML processor must further
- * process the normalized attribute value by discarding any leading and
- * trailing space (#x20) characters, and by replacing sequences of space
- * (#x20) characters by a single space (#x20) character.
- * All attributes for which no declaration has been read should be treated
- * by a non-validating parser as if declared CDATA.
- *
- * Returns the AttValue parsed or NULL. The value has to be freed by the caller.
- */
- xmlChar *
- xmlParseAttValue(xmlParserCtxtPtr ctxt) {
- if ((ctxt == NULL) || (ctxt->input == NULL)) return(NULL);
- return(xmlParseAttValueInternal(ctxt, NULL, NULL, 0));
- }
- /**
- * xmlParseSystemLiteral:
- * @ctxt: an XML parser context
- *
- * parse an XML Literal
- *
- * [11] SystemLiteral ::= ('"' [^"]* '"') | ("'" [^']* "'")
- *
- * Returns the SystemLiteral parsed or NULL
- */
- xmlChar *
- xmlParseSystemLiteral(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- int cur, l;
- xmlChar stop;
- int state = ctxt->instate;
- int count = 0;
- SHRINK;
- if (RAW == '"') {
- NEXT;
- stop = '"';
- } else if (RAW == '\'') {
- NEXT;
- stop = '\'';
- } else {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_STARTED, NULL);
- return(NULL);
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- ctxt->instate = XML_PARSER_SYSTEM_LITERAL;
- cur = CUR_CHAR(l);
- while ((IS_CHAR(cur)) && (cur != stop)) { /* checked */
- if (len + 5 >= size) {
- xmlChar *tmp;
- if ((size > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "SystemLiteral");
- xmlFree(buf);
- ctxt->instate = (xmlParserInputState) state;
- return(NULL);
- }
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buf);
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = (xmlParserInputState) state;
- return(NULL);
- }
- buf = tmp;
- }
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return(NULL);
- }
- }
- COPY_BUF(l,buf,len,cur);
- NEXTL(l);
- cur = CUR_CHAR(l);
- if (cur == 0) {
- GROW;
- SHRINK;
- cur = CUR_CHAR(l);
- }
- }
- buf[len] = 0;
- ctxt->instate = (xmlParserInputState) state;
- if (!IS_CHAR(cur)) {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_FINISHED, NULL);
- } else {
- NEXT;
- }
- return(buf);
- }
- /**
- * xmlParsePubidLiteral:
- * @ctxt: an XML parser context
- *
- * parse an XML public literal
- *
- * [12] PubidLiteral ::= '"' PubidChar* '"' | "'" (PubidChar - "'")* "'"
- *
- * Returns the PubidLiteral parsed or NULL.
- */
- xmlChar *
- xmlParsePubidLiteral(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- xmlChar cur;
- xmlChar stop;
- int count = 0;
- xmlParserInputState oldstate = ctxt->instate;
- SHRINK;
- if (RAW == '"') {
- NEXT;
- stop = '"';
- } else if (RAW == '\'') {
- NEXT;
- stop = '\'';
- } else {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_STARTED, NULL);
- return(NULL);
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- ctxt->instate = XML_PARSER_PUBLIC_LITERAL;
- cur = CUR;
- while ((IS_PUBIDCHAR_CH(cur)) && (cur != stop)) { /* checked */
- if (len + 1 >= size) {
- xmlChar *tmp;
- if ((size > XML_MAX_NAME_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_NAME_TOO_LONG, "Public ID");
- xmlFree(buf);
- return(NULL);
- }
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buf);
- return(NULL);
- }
- buf = tmp;
- }
- buf[len++] = cur;
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return(NULL);
- }
- }
- NEXT;
- cur = CUR;
- if (cur == 0) {
- GROW;
- SHRINK;
- cur = CUR;
- }
- }
- buf[len] = 0;
- if (cur != stop) {
- xmlFatalErr(ctxt, XML_ERR_LITERAL_NOT_FINISHED, NULL);
- } else {
- NEXT;
- }
- ctxt->instate = oldstate;
- return(buf);
- }
- static void xmlParseCharDataComplex(xmlParserCtxtPtr ctxt, int cdata);
- /*
- * used for the test in the inner loop of the char data testing
- */
- static const unsigned char test_char_data[256] = {
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x9, CR/LF separated */
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x00, 0x27, /* & */
- 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E, 0x2F,
- 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
- 0x38, 0x39, 0x3A, 0x3B, 0x00, 0x3D, 0x3E, 0x3F, /* < */
- 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47,
- 0x48, 0x49, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F,
- 0x50, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57,
- 0x58, 0x59, 0x5A, 0x5B, 0x5C, 0x00, 0x5E, 0x5F, /* ] */
- 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67,
- 0x68, 0x69, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F,
- 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77,
- 0x78, 0x79, 0x7A, 0x7B, 0x7C, 0x7D, 0x7E, 0x7F,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* non-ascii */
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
- 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
- };
- /**
- * xmlParseCharData:
- * @ctxt: an XML parser context
- * @cdata: int indicating whether we are within a CDATA section
- *
- * parse a CharData section.
- * if we are within a CDATA section ']]>' marks an end of section.
- *
- * The right angle bracket (>) may be represented using the string ">",
- * and must, for compatibility, be escaped using ">" or a character
- * reference when it appears in the string "]]>" in content, when that
- * string is not marking the end of a CDATA section.
- *
- * [14] CharData ::= [^<&]* - ([^<&]* ']]>' [^<&]*)
- */
- void
- xmlParseCharData(xmlParserCtxtPtr ctxt, int cdata) {
- const xmlChar *in;
- int nbchar = 0;
- int line = ctxt->input->line;
- int col = ctxt->input->col;
- int ccol;
- SHRINK;
- GROW;
- /*
- * Accelerated common case where input don't need to be
- * modified before passing it to the handler.
- */
- if (!cdata) {
- in = ctxt->input->cur;
- do {
- get_more_space:
- while (*in == 0x20) { in++; ctxt->input->col++; }
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- goto get_more_space;
- }
- if (*in == '<') {
- nbchar = in - ctxt->input->cur;
- if (nbchar > 0) {
- const xmlChar *tmp = ctxt->input->cur;
- ctxt->input->cur = in;
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->ignorableWhitespace !=
- ctxt->sax->characters)) {
- if (areBlanks(ctxt, tmp, nbchar, 1)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData,
- tmp, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- tmp, nbchar);
- if (*ctxt->space == -1)
- *ctxt->space = -2;
- }
- } else if ((ctxt->sax != NULL) &&
- (ctxt->sax->characters != NULL)) {
- ctxt->sax->characters(ctxt->userData,
- tmp, nbchar);
- }
- }
- return;
- }
- get_more:
- ccol = ctxt->input->col;
- while (test_char_data[*in]) {
- in++;
- ccol++;
- }
- ctxt->input->col = ccol;
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- goto get_more;
- }
- if (*in == ']') {
- if ((in[1] == ']') && (in[2] == '>')) {
- xmlFatalErr(ctxt, XML_ERR_MISPLACED_CDATA_END, NULL);
- ctxt->input->cur = in + 1;
- return;
- }
- in++;
- ctxt->input->col++;
- goto get_more;
- }
- nbchar = in - ctxt->input->cur;
- if (nbchar > 0) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->ignorableWhitespace !=
- ctxt->sax->characters) &&
- (IS_BLANK_CH(*ctxt->input->cur))) {
- const xmlChar *tmp = ctxt->input->cur;
- ctxt->input->cur = in;
- if (areBlanks(ctxt, tmp, nbchar, 0)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData,
- tmp, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- tmp, nbchar);
- if (*ctxt->space == -1)
- *ctxt->space = -2;
- }
- line = ctxt->input->line;
- col = ctxt->input->col;
- } else if (ctxt->sax != NULL) {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- ctxt->input->cur, nbchar);
- line = ctxt->input->line;
- col = ctxt->input->col;
- }
- /* something really bad happened in the SAX callback */
- if (ctxt->instate != XML_PARSER_CONTENT)
- return;
- }
- ctxt->input->cur = in;
- if (*in == 0xD) {
- in++;
- if (*in == 0xA) {
- ctxt->input->cur = in;
- in++;
- ctxt->input->line++; ctxt->input->col = 1;
- continue; /* while */
- }
- in--;
- }
- if (*in == '<') {
- return;
- }
- if (*in == '&') {
- return;
- }
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- in = ctxt->input->cur;
- } while (((*in >= 0x20) && (*in <= 0x7F)) || (*in == 0x09) || (*in == 0x0a));
- nbchar = 0;
- }
- ctxt->input->line = line;
- ctxt->input->col = col;
- xmlParseCharDataComplex(ctxt, cdata);
- }
- /**
- * xmlParseCharDataComplex:
- * @ctxt: an XML parser context
- * @cdata: int indicating whether we are within a CDATA section
- *
- * parse a CharData section.this is the fallback function
- * of xmlParseCharData() when the parsing requires handling
- * of non-ASCII characters.
- */
- static void
- xmlParseCharDataComplex(xmlParserCtxtPtr ctxt, int cdata) {
- xmlChar buf[XML_PARSER_BIG_BUFFER_SIZE + 5];
- int nbchar = 0;
- int cur, l;
- int count = 0;
- SHRINK;
- GROW;
- cur = CUR_CHAR(l);
- while ((cur != '<') && /* checked */
- (cur != '&') &&
- (IS_CHAR(cur))) /* test also done in xmlCurrentChar() */ {
- if ((cur == ']') && (NXT(1) == ']') &&
- (NXT(2) == '>')) {
- if (cdata) break;
- else {
- xmlFatalErr(ctxt, XML_ERR_MISPLACED_CDATA_END, NULL);
- }
- }
- COPY_BUF(l,buf,nbchar,cur);
- if (nbchar >= XML_PARSER_BIG_BUFFER_SIZE) {
- buf[nbchar] = 0;
- /*
- * OK the segment is to be consumed as chars.
- */
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (areBlanks(ctxt, buf, nbchar, 0)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData,
- buf, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData, buf, nbchar);
- if ((ctxt->sax->characters !=
- ctxt->sax->ignorableWhitespace) &&
- (*ctxt->space == -1))
- *ctxt->space = -2;
- }
- }
- nbchar = 0;
- /* something really bad happened in the SAX callback */
- if (ctxt->instate != XML_PARSER_CONTENT)
- return;
- }
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- }
- NEXTL(l);
- cur = CUR_CHAR(l);
- }
- if (nbchar != 0) {
- buf[nbchar] = 0;
- /*
- * OK the segment is to be consumed as chars.
- */
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (areBlanks(ctxt, buf, nbchar, 0)) {
- if (ctxt->sax->ignorableWhitespace != NULL)
- ctxt->sax->ignorableWhitespace(ctxt->userData, buf, nbchar);
- } else {
- if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData, buf, nbchar);
- if ((ctxt->sax->characters != ctxt->sax->ignorableWhitespace) &&
- (*ctxt->space == -1))
- *ctxt->space = -2;
- }
- }
- }
- if ((cur != 0) && (!IS_CHAR(cur))) {
- /* Generate the error and skip the offending character */
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "PCDATA invalid Char value %d\n",
- cur);
- NEXTL(l);
- }
- }
- /**
- * xmlParseExternalID:
- * @ctxt: an XML parser context
- * @publicID: a xmlChar** receiving PubidLiteral
- * @strict: indicate whether we should restrict parsing to only
- * production [75], see NOTE below
- *
- * Parse an External ID or a Public ID
- *
- * NOTE: Productions [75] and [83] interact badly since [75] can generate
- * 'PUBLIC' S PubidLiteral S SystemLiteral
- *
- * [75] ExternalID ::= 'SYSTEM' S SystemLiteral
- * | 'PUBLIC' S PubidLiteral S SystemLiteral
- *
- * [83] PublicID ::= 'PUBLIC' S PubidLiteral
- *
- * Returns the function returns SystemLiteral and in the second
- * case publicID receives PubidLiteral, is strict is off
- * it is possible to return NULL and have publicID set.
- */
- xmlChar *
- xmlParseExternalID(xmlParserCtxtPtr ctxt, xmlChar **publicID, int strict) {
- xmlChar *URI = NULL;
- SHRINK;
- *publicID = NULL;
- if (CMP6(CUR_PTR, 'S', 'Y', 'S', 'T', 'E', 'M')) {
- SKIP(6);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'SYSTEM'\n");
- }
- URI = xmlParseSystemLiteral(ctxt);
- if (URI == NULL) {
- xmlFatalErr(ctxt, XML_ERR_URI_REQUIRED, NULL);
- }
- } else if (CMP6(CUR_PTR, 'P', 'U', 'B', 'L', 'I', 'C')) {
- SKIP(6);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'PUBLIC'\n");
- }
- *publicID = xmlParsePubidLiteral(ctxt);
- if (*publicID == NULL) {
- xmlFatalErr(ctxt, XML_ERR_PUBID_REQUIRED, NULL);
- }
- if (strict) {
- /*
- * We don't handle [83] so "S SystemLiteral" is required.
- */
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the Public Identifier\n");
- }
- } else {
- /*
- * We handle [83] so we return immediately, if
- * "S SystemLiteral" is not detected. We skip blanks if no
- * system literal was found, but this is harmless since we must
- * be at the end of a NotationDecl.
- */
- if (SKIP_BLANKS == 0) return(NULL);
- if ((CUR != '\'') && (CUR != '"')) return(NULL);
- }
- URI = xmlParseSystemLiteral(ctxt);
- if (URI == NULL) {
- xmlFatalErr(ctxt, XML_ERR_URI_REQUIRED, NULL);
- }
- }
- return(URI);
- }
- /**
- * xmlParseCommentComplex:
- * @ctxt: an XML parser context
- * @buf: the already parsed part of the buffer
- * @len: number of bytes in the buffer
- * @size: allocated size of the buffer
- *
- * Skip an XML (SGML) comment <!-- .... -->
- * The spec says that "For compatibility, the string "--" (double-hyphen)
- * must not occur within comments. "
- * This is the slow routine in case the accelerator for ascii didn't work
- *
- * [15] Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
- */
- static void
- xmlParseCommentComplex(xmlParserCtxtPtr ctxt, xmlChar *buf,
- size_t len, size_t size) {
- int q, ql;
- int r, rl;
- int cur, l;
- size_t count = 0;
- int inputid;
- inputid = ctxt->input->id;
- if (buf == NULL) {
- len = 0;
- size = XML_PARSER_BUFFER_SIZE;
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return;
- }
- }
- GROW; /* Assure there's enough input data */
- q = CUR_CHAR(ql);
- if (q == 0)
- goto not_terminated;
- if (!IS_CHAR(q)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseComment: invalid xmlChar value %d\n",
- q);
- xmlFree (buf);
- return;
- }
- NEXTL(ql);
- r = CUR_CHAR(rl);
- if (r == 0)
- goto not_terminated;
- if (!IS_CHAR(r)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseComment: invalid xmlChar value %d\n",
- q);
- xmlFree (buf);
- return;
- }
- NEXTL(rl);
- cur = CUR_CHAR(l);
- if (cur == 0)
- goto not_terminated;
- while (IS_CHAR(cur) && /* checked */
- ((cur != '>') ||
- (r != '-') || (q != '-'))) {
- if ((r == '-') && (q == '-')) {
- xmlFatalErr(ctxt, XML_ERR_HYPHEN_IN_COMMENT, NULL);
- }
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment too big found", NULL);
- xmlFree (buf);
- return;
- }
- if (len + 5 >= size) {
- xmlChar *new_buf;
- size_t new_size;
- new_size = size * 2;
- new_buf = (xmlChar *) xmlRealloc(buf, new_size);
- if (new_buf == NULL) {
- xmlFree (buf);
- xmlErrMemory(ctxt, NULL);
- return;
- }
- buf = new_buf;
- size = new_size;
- }
- COPY_BUF(ql,buf,len,q);
- q = r;
- ql = rl;
- r = cur;
- rl = l;
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- count = 0;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- }
- NEXTL(l);
- cur = CUR_CHAR(l);
- if (cur == 0) {
- SHRINK;
- GROW;
- cur = CUR_CHAR(l);
- }
- }
- buf[len] = 0;
- if (cur == 0) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment not terminated \n<!--%.50s\n", buf);
- } else if (!IS_CHAR(cur)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlParseComment: invalid xmlChar value %d\n",
- cur);
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Comment doesn't start and stop in the same"
- " entity\n");
- }
- NEXT;
- if ((ctxt->sax != NULL) && (ctxt->sax->comment != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->comment(ctxt->userData, buf);
- }
- xmlFree(buf);
- return;
- not_terminated:
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment not terminated\n", NULL);
- xmlFree(buf);
- return;
- }
- /**
- * xmlParseComment:
- * @ctxt: an XML parser context
- *
- * Skip an XML (SGML) comment <!-- .... -->
- * The spec says that "For compatibility, the string "--" (double-hyphen)
- * must not occur within comments. "
- *
- * [15] Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
- */
- void
- xmlParseComment(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- size_t size = XML_PARSER_BUFFER_SIZE;
- size_t len = 0;
- xmlParserInputState state;
- const xmlChar *in;
- size_t nbchar = 0;
- int ccol;
- int inputid;
- /*
- * Check that there is a comment right here.
- */
- if ((RAW != '<') || (NXT(1) != '!') ||
- (NXT(2) != '-') || (NXT(3) != '-')) return;
- state = ctxt->instate;
- ctxt->instate = XML_PARSER_COMMENT;
- inputid = ctxt->input->id;
- SKIP(4);
- SHRINK;
- GROW;
- /*
- * Accelerated common case where input don't need to be
- * modified before passing it to the handler.
- */
- in = ctxt->input->cur;
- do {
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- }
- get_more:
- ccol = ctxt->input->col;
- while (((*in > '-') && (*in <= 0x7F)) ||
- ((*in >= 0x20) && (*in < '-')) ||
- (*in == 0x09)) {
- in++;
- ccol++;
- }
- ctxt->input->col = ccol;
- if (*in == 0xA) {
- do {
- ctxt->input->line++; ctxt->input->col = 1;
- in++;
- } while (*in == 0xA);
- goto get_more;
- }
- nbchar = in - ctxt->input->cur;
- /*
- * save current set of data
- */
- if (nbchar > 0) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->comment != NULL)) {
- if (buf == NULL) {
- if ((*in == '-') && (in[1] == '-'))
- size = nbchar + 1;
- else
- size = XML_PARSER_BUFFER_SIZE + nbchar;
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = state;
- return;
- }
- len = 0;
- } else if (len + nbchar + 1 >= size) {
- xmlChar *new_buf;
- size += len + nbchar + XML_PARSER_BUFFER_SIZE;
- new_buf = (xmlChar *) xmlRealloc(buf,
- size * sizeof(xmlChar));
- if (new_buf == NULL) {
- xmlFree (buf);
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = state;
- return;
- }
- buf = new_buf;
- }
- memcpy(&buf[len], ctxt->input->cur, nbchar);
- len += nbchar;
- buf[len] = 0;
- }
- }
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_COMMENT_NOT_FINISHED,
- "Comment too big found", NULL);
- xmlFree (buf);
- return;
- }
- ctxt->input->cur = in;
- if (*in == 0xA) {
- in++;
- ctxt->input->line++; ctxt->input->col = 1;
- }
- if (*in == 0xD) {
- in++;
- if (*in == 0xA) {
- ctxt->input->cur = in;
- in++;
- ctxt->input->line++; ctxt->input->col = 1;
- continue; /* while */
- }
- in--;
- }
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- in = ctxt->input->cur;
- if (*in == '-') {
- if (in[1] == '-') {
- if (in[2] == '>') {
- if (ctxt->input->id != inputid) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "comment doesn't start and stop in the"
- " same entity\n");
- }
- SKIP(3);
- if ((ctxt->sax != NULL) && (ctxt->sax->comment != NULL) &&
- (!ctxt->disableSAX)) {
- if (buf != NULL)
- ctxt->sax->comment(ctxt->userData, buf);
- else
- ctxt->sax->comment(ctxt->userData, BAD_CAST "");
- }
- if (buf != NULL)
- xmlFree(buf);
- if (ctxt->instate != XML_PARSER_EOF)
- ctxt->instate = state;
- return;
- }
- if (buf != NULL) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_HYPHEN_IN_COMMENT,
- "Double hyphen within comment: "
- "<!--%.50s\n",
- buf);
- } else
- xmlFatalErrMsgStr(ctxt, XML_ERR_HYPHEN_IN_COMMENT,
- "Double hyphen within comment\n", NULL);
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- in++;
- ctxt->input->col++;
- }
- in++;
- ctxt->input->col++;
- goto get_more;
- }
- } while (((*in >= 0x20) && (*in <= 0x7F)) || (*in == 0x09) || (*in == 0x0a));
- xmlParseCommentComplex(ctxt, buf, len, size);
- ctxt->instate = state;
- return;
- }
- /**
- * xmlParsePITarget:
- * @ctxt: an XML parser context
- *
- * parse the name of a PI
- *
- * [17] PITarget ::= Name - (('X' | 'x') ('M' | 'm') ('L' | 'l'))
- *
- * Returns the PITarget name or NULL
- */
- const xmlChar *
- xmlParsePITarget(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- name = xmlParseName(ctxt);
- if ((name != NULL) &&
- ((name[0] == 'x') || (name[0] == 'X')) &&
- ((name[1] == 'm') || (name[1] == 'M')) &&
- ((name[2] == 'l') || (name[2] == 'L'))) {
- int i;
- if ((name[0] == 'x') && (name[1] == 'm') &&
- (name[2] == 'l') && (name[3] == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_RESERVED_XML_NAME,
- "XML declaration allowed only at the start of the document\n");
- return(name);
- } else if (name[3] == 0) {
- xmlFatalErr(ctxt, XML_ERR_RESERVED_XML_NAME, NULL);
- return(name);
- }
- for (i = 0;;i++) {
- if (xmlW3CPIs[i] == NULL) break;
- if (xmlStrEqual(name, (const xmlChar *)xmlW3CPIs[i]))
- return(name);
- }
- xmlWarningMsg(ctxt, XML_ERR_RESERVED_XML_NAME,
- "xmlParsePITarget: invalid name prefix 'xml'\n",
- NULL, NULL);
- }
- if ((name != NULL) && (xmlStrchr(name, ':') != NULL)) {
- xmlNsErr(ctxt, XML_NS_ERR_COLON,
- "colons are forbidden from PI names '%s'\n", name, NULL, NULL);
- }
- return(name);
- }
- #ifdef LIBXML_CATALOG_ENABLED
- /**
- * xmlParseCatalogPI:
- * @ctxt: an XML parser context
- * @catalog: the PI value string
- *
- * parse an XML Catalog Processing Instruction.
- *
- * <?oasis-xml-catalog catalog="http://example.com/catalog.xml"?>
- *
- * Occurs only if allowed by the user and if happening in the Misc
- * part of the document before any doctype information
- * This will add the given catalog to the parsing context in order
- * to be used if there is a resolution need further down in the document
- */
- static void
- xmlParseCatalogPI(xmlParserCtxtPtr ctxt, const xmlChar *catalog) {
- xmlChar *URL = NULL;
- const xmlChar *tmp, *base;
- xmlChar marker;
- tmp = catalog;
- while (IS_BLANK_CH(*tmp)) tmp++;
- if (xmlStrncmp(tmp, BAD_CAST"catalog", 7))
- goto error;
- tmp += 7;
- while (IS_BLANK_CH(*tmp)) tmp++;
- if (*tmp != '=') {
- return;
- }
- tmp++;
- while (IS_BLANK_CH(*tmp)) tmp++;
- marker = *tmp;
- if ((marker != '\'') && (marker != '"'))
- goto error;
- tmp++;
- base = tmp;
- while ((*tmp != 0) && (*tmp != marker)) tmp++;
- if (*tmp == 0)
- goto error;
- URL = xmlStrndup(base, tmp - base);
- tmp++;
- while (IS_BLANK_CH(*tmp)) tmp++;
- if (*tmp != 0)
- goto error;
- if (URL != NULL) {
- ctxt->catalogs = xmlCatalogAddLocal(ctxt->catalogs, URL);
- xmlFree(URL);
- }
- return;
- error:
- xmlWarningMsg(ctxt, XML_WAR_CATALOG_PI,
- "Catalog PI syntax error: %s\n",
- catalog, NULL);
- if (URL != NULL)
- xmlFree(URL);
- }
- #endif
- /**
- * xmlParsePI:
- * @ctxt: an XML parser context
- *
- * parse an XML Processing Instruction.
- *
- * [16] PI ::= '<?' PITarget (S (Char* - (Char* '?>' Char*)))? '?>'
- *
- * The processing is transferred to SAX once parsed.
- */
- void
- xmlParsePI(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- size_t len = 0;
- size_t size = XML_PARSER_BUFFER_SIZE;
- int cur, l;
- const xmlChar *target;
- xmlParserInputState state;
- int count = 0;
- if ((RAW == '<') && (NXT(1) == '?')) {
- int inputid = ctxt->input->id;
- state = ctxt->instate;
- ctxt->instate = XML_PARSER_PI;
- /*
- * this is a Processing Instruction.
- */
- SKIP(2);
- SHRINK;
- /*
- * Parse the target name and check for special support like
- * namespace.
- */
- target = xmlParsePITarget(ctxt);
- if (target != NULL) {
- if ((RAW == '?') && (NXT(1) == '>')) {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "PI declaration doesn't start and stop in"
- " the same entity\n");
- }
- SKIP(2);
- /*
- * SAX: PI detected.
- */
- if ((ctxt->sax) && (!ctxt->disableSAX) &&
- (ctxt->sax->processingInstruction != NULL))
- ctxt->sax->processingInstruction(ctxt->userData,
- target, NULL);
- if (ctxt->instate != XML_PARSER_EOF)
- ctxt->instate = state;
- return;
- }
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- ctxt->instate = state;
- return;
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_SPACE_REQUIRED,
- "ParsePI: PI %s space expected\n", target);
- }
- cur = CUR_CHAR(l);
- while (IS_CHAR(cur) && /* checked */
- ((cur != '?') || (NXT(1) != '>'))) {
- if (len + 5 >= size) {
- xmlChar *tmp;
- size_t new_size = size * 2;
- tmp = (xmlChar *) xmlRealloc(buf, new_size);
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buf);
- ctxt->instate = state;
- return;
- }
- buf = tmp;
- size = new_size;
- }
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- count = 0;
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_PI_NOT_FINISHED,
- "PI %s too big found", target);
- xmlFree(buf);
- ctxt->instate = state;
- return;
- }
- }
- COPY_BUF(l,buf,len,cur);
- NEXTL(l);
- cur = CUR_CHAR(l);
- if (cur == 0) {
- SHRINK;
- GROW;
- cur = CUR_CHAR(l);
- }
- }
- if ((len > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_PI_NOT_FINISHED,
- "PI %s too big found", target);
- xmlFree(buf);
- ctxt->instate = state;
- return;
- }
- buf[len] = 0;
- if (cur != '?') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_PI_NOT_FINISHED,
- "ParsePI: PI %s never end ...\n", target);
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "PI declaration doesn't start and stop in"
- " the same entity\n");
- }
- SKIP(2);
- #ifdef LIBXML_CATALOG_ENABLED
- if (((state == XML_PARSER_MISC) ||
- (state == XML_PARSER_START)) &&
- (xmlStrEqual(target, XML_CATALOG_PI))) {
- xmlCatalogAllow allow = xmlCatalogGetDefaults();
- if ((allow == XML_CATA_ALLOW_DOCUMENT) ||
- (allow == XML_CATA_ALLOW_ALL))
- xmlParseCatalogPI(ctxt, buf);
- }
- #endif
- /*
- * SAX: PI detected.
- */
- if ((ctxt->sax) && (!ctxt->disableSAX) &&
- (ctxt->sax->processingInstruction != NULL))
- ctxt->sax->processingInstruction(ctxt->userData,
- target, buf);
- }
- xmlFree(buf);
- } else {
- xmlFatalErr(ctxt, XML_ERR_PI_NOT_STARTED, NULL);
- }
- if (ctxt->instate != XML_PARSER_EOF)
- ctxt->instate = state;
- }
- }
- /**
- * xmlParseNotationDecl:
- * @ctxt: an XML parser context
- *
- * parse a notation declaration
- *
- * [82] NotationDecl ::= '<!NOTATION' S Name S (ExternalID | PublicID) S? '>'
- *
- * Hence there is actually 3 choices:
- * 'PUBLIC' S PubidLiteral
- * 'PUBLIC' S PubidLiteral S SystemLiteral
- * and 'SYSTEM' S SystemLiteral
- *
- * See the NOTE on xmlParseExternalID().
- */
- void
- xmlParseNotationDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- xmlChar *Pubid;
- xmlChar *Systemid;
- if (CMP10(CUR_PTR, '<', '!', 'N', 'O', 'T', 'A', 'T', 'I', 'O', 'N')) {
- int inputid = ctxt->input->id;
- SHRINK;
- SKIP(10);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '<!NOTATION'\n");
- return;
- }
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_STARTED, NULL);
- return;
- }
- if (xmlStrchr(name, ':') != NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_COLON,
- "colons are forbidden from notation names '%s'\n",
- name, NULL, NULL);
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the NOTATION name'\n");
- return;
- }
- /*
- * Parse the IDs.
- */
- Systemid = xmlParseExternalID(ctxt, &Pubid, 0);
- SKIP_BLANKS;
- if (RAW == '>') {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Notation declaration doesn't start and stop"
- " in the same entity\n");
- }
- NEXT;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->notationDecl != NULL))
- ctxt->sax->notationDecl(ctxt->userData, name, Pubid, Systemid);
- } else {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_FINISHED, NULL);
- }
- if (Systemid != NULL) xmlFree(Systemid);
- if (Pubid != NULL) xmlFree(Pubid);
- }
- }
- /**
- * xmlParseEntityDecl:
- * @ctxt: an XML parser context
- *
- * parse <!ENTITY declarations
- *
- * [70] EntityDecl ::= GEDecl | PEDecl
- *
- * [71] GEDecl ::= '<!ENTITY' S Name S EntityDef S? '>'
- *
- * [72] PEDecl ::= '<!ENTITY' S '%' S Name S PEDef S? '>'
- *
- * [73] EntityDef ::= EntityValue | (ExternalID NDataDecl?)
- *
- * [74] PEDef ::= EntityValue | ExternalID
- *
- * [76] NDataDecl ::= S 'NDATA' S Name
- *
- * [ VC: Notation Declared ]
- * The Name must match the declared name of a notation.
- */
- void
- xmlParseEntityDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name = NULL;
- xmlChar *value = NULL;
- xmlChar *URI = NULL, *literal = NULL;
- const xmlChar *ndata = NULL;
- int isParameter = 0;
- xmlChar *orig = NULL;
- /* GROW; done in the caller */
- if (CMP8(CUR_PTR, '<', '!', 'E', 'N', 'T', 'I', 'T', 'Y')) {
- int inputid = ctxt->input->id;
- SHRINK;
- SKIP(8);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '<!ENTITY'\n");
- }
- if (RAW == '%') {
- NEXT;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '%%'\n");
- }
- isParameter = 1;
- }
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseEntityDecl: no name\n");
- return;
- }
- if (xmlStrchr(name, ':') != NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_COLON,
- "colons are forbidden from entities names '%s'\n",
- name, NULL, NULL);
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the entity name\n");
- }
- ctxt->instate = XML_PARSER_ENTITY_DECL;
- /*
- * handle the various case of definitions...
- */
- if (isParameter) {
- if ((RAW == '"') || (RAW == '\'')) {
- value = xmlParseEntityValue(ctxt, &orig);
- if (value) {
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) && (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_INTERNAL_PARAMETER_ENTITY,
- NULL, NULL, value);
- }
- } else {
- URI = xmlParseExternalID(ctxt, &literal, 1);
- if ((URI == NULL) && (literal == NULL)) {
- xmlFatalErr(ctxt, XML_ERR_VALUE_REQUIRED, NULL);
- }
- if (URI) {
- xmlURIPtr uri;
- uri = xmlParseURI((const char *) URI);
- if (uri == NULL) {
- xmlErrMsgStr(ctxt, XML_ERR_INVALID_URI,
- "Invalid URI: %s\n", URI);
- /*
- * This really ought to be a well formedness error
- * but the XML Core WG decided otherwise c.f. issue
- * E26 of the XML erratas.
- */
- } else {
- if (uri->fragment != NULL) {
- /*
- * Okay this is foolish to block those but not
- * invalid URIs.
- */
- xmlFatalErr(ctxt, XML_ERR_URI_FRAGMENT, NULL);
- } else {
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) &&
- (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_EXTERNAL_PARAMETER_ENTITY,
- literal, URI, NULL);
- }
- xmlFreeURI(uri);
- }
- }
- }
- } else {
- if ((RAW == '"') || (RAW == '\'')) {
- value = xmlParseEntityValue(ctxt, &orig);
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) && (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_INTERNAL_GENERAL_ENTITY,
- NULL, NULL, value);
- /*
- * For expat compatibility in SAX mode.
- */
- if ((ctxt->myDoc == NULL) ||
- (xmlStrEqual(ctxt->myDoc->version, SAX_COMPAT_MODE))) {
- if (ctxt->myDoc == NULL) {
- ctxt->myDoc = xmlNewDoc(SAX_COMPAT_MODE);
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return;
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- }
- if (ctxt->myDoc->intSubset == NULL)
- ctxt->myDoc->intSubset = xmlNewDtd(ctxt->myDoc,
- BAD_CAST "fake", NULL, NULL);
- xmlSAX2EntityDecl(ctxt, name, XML_INTERNAL_GENERAL_ENTITY,
- NULL, NULL, value);
- }
- } else {
- URI = xmlParseExternalID(ctxt, &literal, 1);
- if ((URI == NULL) && (literal == NULL)) {
- xmlFatalErr(ctxt, XML_ERR_VALUE_REQUIRED, NULL);
- }
- if (URI) {
- xmlURIPtr uri;
- uri = xmlParseURI((const char *)URI);
- if (uri == NULL) {
- xmlErrMsgStr(ctxt, XML_ERR_INVALID_URI,
- "Invalid URI: %s\n", URI);
- /*
- * This really ought to be a well formedness error
- * but the XML Core WG decided otherwise c.f. issue
- * E26 of the XML erratas.
- */
- } else {
- if (uri->fragment != NULL) {
- /*
- * Okay this is foolish to block those but not
- * invalid URIs.
- */
- xmlFatalErr(ctxt, XML_ERR_URI_FRAGMENT, NULL);
- }
- xmlFreeURI(uri);
- }
- }
- if ((RAW != '>') && (SKIP_BLANKS == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required before 'NDATA'\n");
- }
- if (CMP5(CUR_PTR, 'N', 'D', 'A', 'T', 'A')) {
- SKIP(5);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'NDATA'\n");
- }
- ndata = xmlParseName(ctxt);
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->unparsedEntityDecl != NULL))
- ctxt->sax->unparsedEntityDecl(ctxt->userData, name,
- literal, URI, ndata);
- } else {
- if ((ctxt->sax != NULL) &&
- (!ctxt->disableSAX) && (ctxt->sax->entityDecl != NULL))
- ctxt->sax->entityDecl(ctxt->userData, name,
- XML_EXTERNAL_GENERAL_PARSED_ENTITY,
- literal, URI, NULL);
- /*
- * For expat compatibility in SAX mode.
- * assuming the entity replacement was asked for
- */
- if ((ctxt->replaceEntities != 0) &&
- ((ctxt->myDoc == NULL) ||
- (xmlStrEqual(ctxt->myDoc->version, SAX_COMPAT_MODE)))) {
- if (ctxt->myDoc == NULL) {
- ctxt->myDoc = xmlNewDoc(SAX_COMPAT_MODE);
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return;
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- }
- if (ctxt->myDoc->intSubset == NULL)
- ctxt->myDoc->intSubset = xmlNewDtd(ctxt->myDoc,
- BAD_CAST "fake", NULL, NULL);
- xmlSAX2EntityDecl(ctxt, name,
- XML_EXTERNAL_GENERAL_PARSED_ENTITY,
- literal, URI, NULL);
- }
- }
- }
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- SKIP_BLANKS;
- if (RAW != '>') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_NOT_FINISHED,
- "xmlParseEntityDecl: entity %s not terminated\n", name);
- xmlHaltParser(ctxt);
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Entity declaration doesn't start and stop in"
- " the same entity\n");
- }
- NEXT;
- }
- if (orig != NULL) {
- /*
- * Ugly mechanism to save the raw entity value.
- */
- xmlEntityPtr cur = NULL;
- if (isParameter) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getParameterEntity != NULL))
- cur = ctxt->sax->getParameterEntity(ctxt->userData, name);
- } else {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getEntity != NULL))
- cur = ctxt->sax->getEntity(ctxt->userData, name);
- if ((cur == NULL) && (ctxt->userData==ctxt)) {
- cur = xmlSAX2GetEntity(ctxt, name);
- }
- }
- if ((cur != NULL) && (cur->orig == NULL)) {
- cur->orig = orig;
- orig = NULL;
- }
- }
- done:
- if (value != NULL) xmlFree(value);
- if (URI != NULL) xmlFree(URI);
- if (literal != NULL) xmlFree(literal);
- if (orig != NULL) xmlFree(orig);
- }
- }
- /**
- * xmlParseDefaultDecl:
- * @ctxt: an XML parser context
- * @value: Receive a possible fixed default value for the attribute
- *
- * Parse an attribute default declaration
- *
- * [60] DefaultDecl ::= '#REQUIRED' | '#IMPLIED' | (('#FIXED' S)? AttValue)
- *
- * [ VC: Required Attribute ]
- * if the default declaration is the keyword #REQUIRED, then the
- * attribute must be specified for all elements of the type in the
- * attribute-list declaration.
- *
- * [ VC: Attribute Default Legal ]
- * The declared default value must meet the lexical constraints of
- * the declared attribute type c.f. xmlValidateAttributeDecl()
- *
- * [ VC: Fixed Attribute Default ]
- * if an attribute has a default value declared with the #FIXED
- * keyword, instances of that attribute must match the default value.
- *
- * [ WFC: No < in Attribute Values ]
- * handled in xmlParseAttValue()
- *
- * returns: XML_ATTRIBUTE_NONE, XML_ATTRIBUTE_REQUIRED, XML_ATTRIBUTE_IMPLIED
- * or XML_ATTRIBUTE_FIXED.
- */
- int
- xmlParseDefaultDecl(xmlParserCtxtPtr ctxt, xmlChar **value) {
- int val;
- xmlChar *ret;
- *value = NULL;
- if (CMP9(CUR_PTR, '#', 'R', 'E', 'Q', 'U', 'I', 'R', 'E', 'D')) {
- SKIP(9);
- return(XML_ATTRIBUTE_REQUIRED);
- }
- if (CMP8(CUR_PTR, '#', 'I', 'M', 'P', 'L', 'I', 'E', 'D')) {
- SKIP(8);
- return(XML_ATTRIBUTE_IMPLIED);
- }
- val = XML_ATTRIBUTE_NONE;
- if (CMP6(CUR_PTR, '#', 'F', 'I', 'X', 'E', 'D')) {
- SKIP(6);
- val = XML_ATTRIBUTE_FIXED;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '#FIXED'\n");
- }
- }
- ret = xmlParseAttValue(ctxt);
- ctxt->instate = XML_PARSER_DTD;
- if (ret == NULL) {
- xmlFatalErrMsg(ctxt, (xmlParserErrors)ctxt->errNo,
- "Attribute default value declaration error\n");
- } else
- *value = ret;
- return(val);
- }
- /**
- * xmlParseNotationType:
- * @ctxt: an XML parser context
- *
- * parse an Notation attribute type.
- *
- * Note: the leading 'NOTATION' S part has already being parsed...
- *
- * [58] NotationType ::= 'NOTATION' S '(' S? Name (S? '|' S? Name)* S? ')'
- *
- * [ VC: Notation Attributes ]
- * Values of this type must match one of the notation names included
- * in the declaration; all notation names in the declaration must be declared.
- *
- * Returns: the notation attribute tree built while parsing
- */
- xmlEnumerationPtr
- xmlParseNotationType(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- xmlEnumerationPtr ret = NULL, last = NULL, cur, tmp;
- if (RAW != '(') {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_STARTED, NULL);
- return(NULL);
- }
- SHRINK;
- do {
- NEXT;
- SKIP_BLANKS;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "Name expected in NOTATION declaration\n");
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- tmp = ret;
- while (tmp != NULL) {
- if (xmlStrEqual(name, tmp->name)) {
- xmlValidityError(ctxt, XML_DTD_DUP_TOKEN,
- "standalone: attribute notation value token %s duplicated\n",
- name, NULL);
- if (!xmlDictOwns(ctxt->dict, name))
- xmlFree((xmlChar *) name);
- break;
- }
- tmp = tmp->next;
- }
- if (tmp == NULL) {
- cur = xmlCreateEnumeration(name);
- if (cur == NULL) {
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- if (last == NULL) ret = last = cur;
- else {
- last->next = cur;
- last = cur;
- }
- }
- SKIP_BLANKS;
- } while (RAW == '|');
- if (RAW != ')') {
- xmlFatalErr(ctxt, XML_ERR_NOTATION_NOT_FINISHED, NULL);
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- NEXT;
- return(ret);
- }
- /**
- * xmlParseEnumerationType:
- * @ctxt: an XML parser context
- *
- * parse an Enumeration attribute type.
- *
- * [59] Enumeration ::= '(' S? Nmtoken (S? '|' S? Nmtoken)* S? ')'
- *
- * [ VC: Enumeration ]
- * Values of this type must match one of the Nmtoken tokens in
- * the declaration
- *
- * Returns: the enumeration attribute tree built while parsing
- */
- xmlEnumerationPtr
- xmlParseEnumerationType(xmlParserCtxtPtr ctxt) {
- xmlChar *name;
- xmlEnumerationPtr ret = NULL, last = NULL, cur, tmp;
- if (RAW != '(') {
- xmlFatalErr(ctxt, XML_ERR_ATTLIST_NOT_STARTED, NULL);
- return(NULL);
- }
- SHRINK;
- do {
- NEXT;
- SKIP_BLANKS;
- name = xmlParseNmtoken(ctxt);
- if (name == NULL) {
- xmlFatalErr(ctxt, XML_ERR_NMTOKEN_REQUIRED, NULL);
- return(ret);
- }
- tmp = ret;
- while (tmp != NULL) {
- if (xmlStrEqual(name, tmp->name)) {
- xmlValidityError(ctxt, XML_DTD_DUP_TOKEN,
- "standalone: attribute enumeration value token %s duplicated\n",
- name, NULL);
- if (!xmlDictOwns(ctxt->dict, name))
- xmlFree(name);
- break;
- }
- tmp = tmp->next;
- }
- if (tmp == NULL) {
- cur = xmlCreateEnumeration(name);
- if (!xmlDictOwns(ctxt->dict, name))
- xmlFree(name);
- if (cur == NULL) {
- xmlFreeEnumeration(ret);
- return(NULL);
- }
- if (last == NULL) ret = last = cur;
- else {
- last->next = cur;
- last = cur;
- }
- }
- SKIP_BLANKS;
- } while (RAW == '|');
- if (RAW != ')') {
- xmlFatalErr(ctxt, XML_ERR_ATTLIST_NOT_FINISHED, NULL);
- return(ret);
- }
- NEXT;
- return(ret);
- }
- /**
- * xmlParseEnumeratedType:
- * @ctxt: an XML parser context
- * @tree: the enumeration tree built while parsing
- *
- * parse an Enumerated attribute type.
- *
- * [57] EnumeratedType ::= NotationType | Enumeration
- *
- * [58] NotationType ::= 'NOTATION' S '(' S? Name (S? '|' S? Name)* S? ')'
- *
- *
- * Returns: XML_ATTRIBUTE_ENUMERATION or XML_ATTRIBUTE_NOTATION
- */
- int
- xmlParseEnumeratedType(xmlParserCtxtPtr ctxt, xmlEnumerationPtr *tree) {
- if (CMP8(CUR_PTR, 'N', 'O', 'T', 'A', 'T', 'I', 'O', 'N')) {
- SKIP(8);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'NOTATION'\n");
- return(0);
- }
- *tree = xmlParseNotationType(ctxt);
- if (*tree == NULL) return(0);
- return(XML_ATTRIBUTE_NOTATION);
- }
- *tree = xmlParseEnumerationType(ctxt);
- if (*tree == NULL) return(0);
- return(XML_ATTRIBUTE_ENUMERATION);
- }
- /**
- * xmlParseAttributeType:
- * @ctxt: an XML parser context
- * @tree: the enumeration tree built while parsing
- *
- * parse the Attribute list def for an element
- *
- * [54] AttType ::= StringType | TokenizedType | EnumeratedType
- *
- * [55] StringType ::= 'CDATA'
- *
- * [56] TokenizedType ::= 'ID' | 'IDREF' | 'IDREFS' | 'ENTITY' |
- * 'ENTITIES' | 'NMTOKEN' | 'NMTOKENS'
- *
- * Validity constraints for attribute values syntax are checked in
- * xmlValidateAttributeValue()
- *
- * [ VC: ID ]
- * Values of type ID must match the Name production. A name must not
- * appear more than once in an XML document as a value of this type;
- * i.e., ID values must uniquely identify the elements which bear them.
- *
- * [ VC: One ID per Element Type ]
- * No element type may have more than one ID attribute specified.
- *
- * [ VC: ID Attribute Default ]
- * An ID attribute must have a declared default of #IMPLIED or #REQUIRED.
- *
- * [ VC: IDREF ]
- * Values of type IDREF must match the Name production, and values
- * of type IDREFS must match Names; each IDREF Name must match the value
- * of an ID attribute on some element in the XML document; i.e. IDREF
- * values must match the value of some ID attribute.
- *
- * [ VC: Entity Name ]
- * Values of type ENTITY must match the Name production, values
- * of type ENTITIES must match Names; each Entity Name must match the
- * name of an unparsed entity declared in the DTD.
- *
- * [ VC: Name Token ]
- * Values of type NMTOKEN must match the Nmtoken production; values
- * of type NMTOKENS must match Nmtokens.
- *
- * Returns the attribute type
- */
- int
- xmlParseAttributeType(xmlParserCtxtPtr ctxt, xmlEnumerationPtr *tree) {
- SHRINK;
- if (CMP5(CUR_PTR, 'C', 'D', 'A', 'T', 'A')) {
- SKIP(5);
- return(XML_ATTRIBUTE_CDATA);
- } else if (CMP6(CUR_PTR, 'I', 'D', 'R', 'E', 'F', 'S')) {
- SKIP(6);
- return(XML_ATTRIBUTE_IDREFS);
- } else if (CMP5(CUR_PTR, 'I', 'D', 'R', 'E', 'F')) {
- SKIP(5);
- return(XML_ATTRIBUTE_IDREF);
- } else if ((RAW == 'I') && (NXT(1) == 'D')) {
- SKIP(2);
- return(XML_ATTRIBUTE_ID);
- } else if (CMP6(CUR_PTR, 'E', 'N', 'T', 'I', 'T', 'Y')) {
- SKIP(6);
- return(XML_ATTRIBUTE_ENTITY);
- } else if (CMP8(CUR_PTR, 'E', 'N', 'T', 'I', 'T', 'I', 'E', 'S')) {
- SKIP(8);
- return(XML_ATTRIBUTE_ENTITIES);
- } else if (CMP8(CUR_PTR, 'N', 'M', 'T', 'O', 'K', 'E', 'N', 'S')) {
- SKIP(8);
- return(XML_ATTRIBUTE_NMTOKENS);
- } else if (CMP7(CUR_PTR, 'N', 'M', 'T', 'O', 'K', 'E', 'N')) {
- SKIP(7);
- return(XML_ATTRIBUTE_NMTOKEN);
- }
- return(xmlParseEnumeratedType(ctxt, tree));
- }
- /**
- * xmlParseAttributeListDecl:
- * @ctxt: an XML parser context
- *
- * : parse the Attribute list def for an element
- *
- * [52] AttlistDecl ::= '<!ATTLIST' S Name AttDef* S? '>'
- *
- * [53] AttDef ::= S Name S AttType S DefaultDecl
- *
- */
- void
- xmlParseAttributeListDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *elemName;
- const xmlChar *attrName;
- xmlEnumerationPtr tree;
- if (CMP9(CUR_PTR, '<', '!', 'A', 'T', 'T', 'L', 'I', 'S', 'T')) {
- int inputid = ctxt->input->id;
- SKIP(9);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after '<!ATTLIST'\n");
- }
- elemName = xmlParseName(ctxt);
- if (elemName == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "ATTLIST: no name for Element\n");
- return;
- }
- SKIP_BLANKS;
- GROW;
- while ((RAW != '>') && (ctxt->instate != XML_PARSER_EOF)) {
- int type;
- int def;
- xmlChar *defaultValue = NULL;
- GROW;
- tree = NULL;
- attrName = xmlParseName(ctxt);
- if (attrName == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "ATTLIST: no name for Attribute\n");
- break;
- }
- GROW;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the attribute name\n");
- break;
- }
- type = xmlParseAttributeType(ctxt, &tree);
- if (type <= 0) {
- break;
- }
- GROW;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the attribute type\n");
- if (tree != NULL)
- xmlFreeEnumeration(tree);
- break;
- }
- def = xmlParseDefaultDecl(ctxt, &defaultValue);
- if (def <= 0) {
- if (defaultValue != NULL)
- xmlFree(defaultValue);
- if (tree != NULL)
- xmlFreeEnumeration(tree);
- break;
- }
- if ((type != XML_ATTRIBUTE_CDATA) && (defaultValue != NULL))
- xmlAttrNormalizeSpace(defaultValue, defaultValue);
- GROW;
- if (RAW != '>') {
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the attribute default value\n");
- if (defaultValue != NULL)
- xmlFree(defaultValue);
- if (tree != NULL)
- xmlFreeEnumeration(tree);
- break;
- }
- }
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->attributeDecl != NULL))
- ctxt->sax->attributeDecl(ctxt->userData, elemName, attrName,
- type, def, defaultValue, tree);
- else if (tree != NULL)
- xmlFreeEnumeration(tree);
- if ((ctxt->sax2) && (defaultValue != NULL) &&
- (def != XML_ATTRIBUTE_IMPLIED) &&
- (def != XML_ATTRIBUTE_REQUIRED)) {
- xmlAddDefAttrs(ctxt, elemName, attrName, defaultValue);
- }
- if (ctxt->sax2) {
- xmlAddSpecialAttr(ctxt, elemName, attrName, type);
- }
- if (defaultValue != NULL)
- xmlFree(defaultValue);
- GROW;
- }
- if (RAW == '>') {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Attribute list declaration doesn't start and"
- " stop in the same entity\n");
- }
- NEXT;
- }
- }
- }
- /**
- * xmlParseElementMixedContentDecl:
- * @ctxt: an XML parser context
- * @inputchk: the input used for the current entity, needed for boundary checks
- *
- * parse the declaration for a Mixed Element content
- * The leading '(' and spaces have been skipped in xmlParseElementContentDecl
- *
- * [51] Mixed ::= '(' S? '#PCDATA' (S? '|' S? Name)* S? ')*' |
- * '(' S? '#PCDATA' S? ')'
- *
- * [ VC: Proper Group/PE Nesting ] applies to [51] too (see [49])
- *
- * [ VC: No Duplicate Types ]
- * The same name must not appear more than once in a single
- * mixed-content declaration.
- *
- * returns: the list of the xmlElementContentPtr describing the element choices
- */
- xmlElementContentPtr
- xmlParseElementMixedContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
- xmlElementContentPtr ret = NULL, cur = NULL, n;
- const xmlChar *elem = NULL;
- GROW;
- if (CMP7(CUR_PTR, '#', 'P', 'C', 'D', 'A', 'T', 'A')) {
- SKIP(7);
- SKIP_BLANKS;
- SHRINK;
- if (RAW == ')') {
- if (ctxt->input->id != inputchk) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element content declaration doesn't start and"
- " stop in the same entity\n");
- }
- NEXT;
- ret = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_PCDATA);
- if (ret == NULL)
- return(NULL);
- if (RAW == '*') {
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- NEXT;
- }
- return(ret);
- }
- if ((RAW == '(') || (RAW == '|')) {
- ret = cur = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_PCDATA);
- if (ret == NULL) return(NULL);
- }
- while ((RAW == '|') && (ctxt->instate != XML_PARSER_EOF)) {
- NEXT;
- if (elem == NULL) {
- ret = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
- if (ret == NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, cur);
- return(NULL);
- }
- ret->c1 = cur;
- if (cur != NULL)
- cur->parent = ret;
- cur = ret;
- } else {
- n = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
- if (n == NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- n->c1 = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
- if (n->c1 != NULL)
- n->c1->parent = n;
- cur->c2 = n;
- if (n != NULL)
- n->parent = cur;
- cur = n;
- }
- SKIP_BLANKS;
- elem = xmlParseName(ctxt);
- if (elem == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseElementMixedContentDecl : Name expected\n");
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- SKIP_BLANKS;
- GROW;
- }
- if ((RAW == ')') && (NXT(1) == '*')) {
- if (elem != NULL) {
- cur->c2 = xmlNewDocElementContent(ctxt->myDoc, elem,
- XML_ELEMENT_CONTENT_ELEMENT);
- if (cur->c2 != NULL)
- cur->c2->parent = cur;
- }
- if (ret != NULL)
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- if (ctxt->input->id != inputchk) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element content declaration doesn't start and"
- " stop in the same entity\n");
- }
- SKIP(2);
- } else {
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- xmlFatalErr(ctxt, XML_ERR_MIXED_NOT_STARTED, NULL);
- return(NULL);
- }
- } else {
- xmlFatalErr(ctxt, XML_ERR_PCDATA_REQUIRED, NULL);
- }
- return(ret);
- }
- /**
- * xmlParseElementChildrenContentDeclPriv:
- * @ctxt: an XML parser context
- * @inputchk: the input used for the current entity, needed for boundary checks
- * @depth: the level of recursion
- *
- * parse the declaration for a Mixed Element content
- * The leading '(' and spaces have been skipped in xmlParseElementContentDecl
- *
- *
- * [47] children ::= (choice | seq) ('?' | '*' | '+')?
- *
- * [48] cp ::= (Name | choice | seq) ('?' | '*' | '+')?
- *
- * [49] choice ::= '(' S? cp ( S? '|' S? cp )* S? ')'
- *
- * [50] seq ::= '(' S? cp ( S? ',' S? cp )* S? ')'
- *
- * [ VC: Proper Group/PE Nesting ] applies to [49] and [50]
- * TODO Parameter-entity replacement text must be properly nested
- * with parenthesized groups. That is to say, if either of the
- * opening or closing parentheses in a choice, seq, or Mixed
- * construct is contained in the replacement text for a parameter
- * entity, both must be contained in the same replacement text. For
- * interoperability, if a parameter-entity reference appears in a
- * choice, seq, or Mixed construct, its replacement text should not
- * be empty, and neither the first nor last non-blank character of
- * the replacement text should be a connector (| or ,).
- *
- * Returns the tree of xmlElementContentPtr describing the element
- * hierarchy.
- */
- static xmlElementContentPtr
- xmlParseElementChildrenContentDeclPriv(xmlParserCtxtPtr ctxt, int inputchk,
- int depth) {
- xmlElementContentPtr ret = NULL, cur = NULL, last = NULL, op = NULL;
- const xmlChar *elem;
- xmlChar type = 0;
- if (((depth > 128) && ((ctxt->options & XML_PARSE_HUGE) == 0)) ||
- (depth > 2048)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_ELEMCONTENT_NOT_FINISHED,
- "xmlParseElementChildrenContentDecl : depth %d too deep, use XML_PARSE_HUGE\n",
- depth);
- return(NULL);
- }
- SKIP_BLANKS;
- GROW;
- if (RAW == '(') {
- int inputid = ctxt->input->id;
- /* Recurse on first child */
- NEXT;
- SKIP_BLANKS;
- cur = ret = xmlParseElementChildrenContentDeclPriv(ctxt, inputid,
- depth + 1);
- if (cur == NULL)
- return(NULL);
- SKIP_BLANKS;
- GROW;
- } else {
- elem = xmlParseName(ctxt);
- if (elem == NULL) {
- xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED, NULL);
- return(NULL);
- }
- cur = ret = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
- if (cur == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- GROW;
- if (RAW == '?') {
- cur->ocur = XML_ELEMENT_CONTENT_OPT;
- NEXT;
- } else if (RAW == '*') {
- cur->ocur = XML_ELEMENT_CONTENT_MULT;
- NEXT;
- } else if (RAW == '+') {
- cur->ocur = XML_ELEMENT_CONTENT_PLUS;
- NEXT;
- } else {
- cur->ocur = XML_ELEMENT_CONTENT_ONCE;
- }
- GROW;
- }
- SKIP_BLANKS;
- SHRINK;
- while ((RAW != ')') && (ctxt->instate != XML_PARSER_EOF)) {
- /*
- * Each loop we parse one separator and one element.
- */
- if (RAW == ',') {
- if (type == 0) type = CUR;
- /*
- * Detect "Name | Name , Name" error
- */
- else if (type != CUR) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_SEPARATOR_REQUIRED,
- "xmlParseElementChildrenContentDecl : '%c' expected\n",
- type);
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- NEXT;
- op = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_SEQ);
- if (op == NULL) {
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- if (last == NULL) {
- op->c1 = ret;
- if (ret != NULL)
- ret->parent = op;
- ret = cur = op;
- } else {
- cur->c2 = op;
- if (op != NULL)
- op->parent = cur;
- op->c1 = last;
- if (last != NULL)
- last->parent = op;
- cur =op;
- last = NULL;
- }
- } else if (RAW == '|') {
- if (type == 0) type = CUR;
- /*
- * Detect "Name , Name | Name" error
- */
- else if (type != CUR) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_SEPARATOR_REQUIRED,
- "xmlParseElementChildrenContentDecl : '%c' expected\n",
- type);
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- NEXT;
- op = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
- if (op == NULL) {
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- if (last == NULL) {
- op->c1 = ret;
- if (ret != NULL)
- ret->parent = op;
- ret = cur = op;
- } else {
- cur->c2 = op;
- if (op != NULL)
- op->parent = cur;
- op->c1 = last;
- if (last != NULL)
- last->parent = op;
- cur =op;
- last = NULL;
- }
- } else {
- xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_FINISHED, NULL);
- if ((last != NULL) && (last != ret))
- xmlFreeDocElementContent(ctxt->myDoc, last);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- GROW;
- SKIP_BLANKS;
- GROW;
- if (RAW == '(') {
- int inputid = ctxt->input->id;
- /* Recurse on second child */
- NEXT;
- SKIP_BLANKS;
- last = xmlParseElementChildrenContentDeclPriv(ctxt, inputid,
- depth + 1);
- if (last == NULL) {
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- SKIP_BLANKS;
- } else {
- elem = xmlParseName(ctxt);
- if (elem == NULL) {
- xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED, NULL);
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- last = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
- if (last == NULL) {
- if (ret != NULL)
- xmlFreeDocElementContent(ctxt->myDoc, ret);
- return(NULL);
- }
- if (RAW == '?') {
- last->ocur = XML_ELEMENT_CONTENT_OPT;
- NEXT;
- } else if (RAW == '*') {
- last->ocur = XML_ELEMENT_CONTENT_MULT;
- NEXT;
- } else if (RAW == '+') {
- last->ocur = XML_ELEMENT_CONTENT_PLUS;
- NEXT;
- } else {
- last->ocur = XML_ELEMENT_CONTENT_ONCE;
- }
- }
- SKIP_BLANKS;
- GROW;
- }
- if ((cur != NULL) && (last != NULL)) {
- cur->c2 = last;
- if (last != NULL)
- last->parent = cur;
- }
- if (ctxt->input->id != inputchk) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element content declaration doesn't start and stop in"
- " the same entity\n");
- }
- NEXT;
- if (RAW == '?') {
- if (ret != NULL) {
- if ((ret->ocur == XML_ELEMENT_CONTENT_PLUS) ||
- (ret->ocur == XML_ELEMENT_CONTENT_MULT))
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- else
- ret->ocur = XML_ELEMENT_CONTENT_OPT;
- }
- NEXT;
- } else if (RAW == '*') {
- if (ret != NULL) {
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- cur = ret;
- /*
- * Some normalization:
- * (a | b* | c?)* == (a | b | c)*
- */
- while ((cur != NULL) && (cur->type == XML_ELEMENT_CONTENT_OR)) {
- if ((cur->c1 != NULL) &&
- ((cur->c1->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c1->ocur == XML_ELEMENT_CONTENT_MULT)))
- cur->c1->ocur = XML_ELEMENT_CONTENT_ONCE;
- if ((cur->c2 != NULL) &&
- ((cur->c2->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c2->ocur == XML_ELEMENT_CONTENT_MULT)))
- cur->c2->ocur = XML_ELEMENT_CONTENT_ONCE;
- cur = cur->c2;
- }
- }
- NEXT;
- } else if (RAW == '+') {
- if (ret != NULL) {
- int found = 0;
- if ((ret->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (ret->ocur == XML_ELEMENT_CONTENT_MULT))
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- else
- ret->ocur = XML_ELEMENT_CONTENT_PLUS;
- /*
- * Some normalization:
- * (a | b*)+ == (a | b)*
- * (a | b?)+ == (a | b)*
- */
- while ((cur != NULL) && (cur->type == XML_ELEMENT_CONTENT_OR)) {
- if ((cur->c1 != NULL) &&
- ((cur->c1->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c1->ocur == XML_ELEMENT_CONTENT_MULT))) {
- cur->c1->ocur = XML_ELEMENT_CONTENT_ONCE;
- found = 1;
- }
- if ((cur->c2 != NULL) &&
- ((cur->c2->ocur == XML_ELEMENT_CONTENT_OPT) ||
- (cur->c2->ocur == XML_ELEMENT_CONTENT_MULT))) {
- cur->c2->ocur = XML_ELEMENT_CONTENT_ONCE;
- found = 1;
- }
- cur = cur->c2;
- }
- if (found)
- ret->ocur = XML_ELEMENT_CONTENT_MULT;
- }
- NEXT;
- }
- return(ret);
- }
- /**
- * xmlParseElementChildrenContentDecl:
- * @ctxt: an XML parser context
- * @inputchk: the input used for the current entity, needed for boundary checks
- *
- * parse the declaration for a Mixed Element content
- * The leading '(' and spaces have been skipped in xmlParseElementContentDecl
- *
- * [47] children ::= (choice | seq) ('?' | '*' | '+')?
- *
- * [48] cp ::= (Name | choice | seq) ('?' | '*' | '+')?
- *
- * [49] choice ::= '(' S? cp ( S? '|' S? cp )* S? ')'
- *
- * [50] seq ::= '(' S? cp ( S? ',' S? cp )* S? ')'
- *
- * [ VC: Proper Group/PE Nesting ] applies to [49] and [50]
- * TODO Parameter-entity replacement text must be properly nested
- * with parenthesized groups. That is to say, if either of the
- * opening or closing parentheses in a choice, seq, or Mixed
- * construct is contained in the replacement text for a parameter
- * entity, both must be contained in the same replacement text. For
- * interoperability, if a parameter-entity reference appears in a
- * choice, seq, or Mixed construct, its replacement text should not
- * be empty, and neither the first nor last non-blank character of
- * the replacement text should be a connector (| or ,).
- *
- * Returns the tree of xmlElementContentPtr describing the element
- * hierarchy.
- */
- xmlElementContentPtr
- xmlParseElementChildrenContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
- /* stub left for API/ABI compat */
- return(xmlParseElementChildrenContentDeclPriv(ctxt, inputchk, 1));
- }
- /**
- * xmlParseElementContentDecl:
- * @ctxt: an XML parser context
- * @name: the name of the element being defined.
- * @result: the Element Content pointer will be stored here if any
- *
- * parse the declaration for an Element content either Mixed or Children,
- * the cases EMPTY and ANY are handled directly in xmlParseElementDecl
- *
- * [46] contentspec ::= 'EMPTY' | 'ANY' | Mixed | children
- *
- * returns: the type of element content XML_ELEMENT_TYPE_xxx
- */
- int
- xmlParseElementContentDecl(xmlParserCtxtPtr ctxt, const xmlChar *name,
- xmlElementContentPtr *result) {
- xmlElementContentPtr tree = NULL;
- int inputid = ctxt->input->id;
- int res;
- *result = NULL;
- if (RAW != '(') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED,
- "xmlParseElementContentDecl : %s '(' expected\n", name);
- return(-1);
- }
- NEXT;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- SKIP_BLANKS;
- if (CMP7(CUR_PTR, '#', 'P', 'C', 'D', 'A', 'T', 'A')) {
- tree = xmlParseElementMixedContentDecl(ctxt, inputid);
- res = XML_ELEMENT_TYPE_MIXED;
- } else {
- tree = xmlParseElementChildrenContentDeclPriv(ctxt, inputid, 1);
- res = XML_ELEMENT_TYPE_ELEMENT;
- }
- SKIP_BLANKS;
- *result = tree;
- return(res);
- }
- /**
- * xmlParseElementDecl:
- * @ctxt: an XML parser context
- *
- * parse an Element declaration.
- *
- * [45] elementdecl ::= '<!ELEMENT' S Name S contentspec S? '>'
- *
- * [ VC: Unique Element Type Declaration ]
- * No element type may be declared more than once
- *
- * Returns the type of the element, or -1 in case of error
- */
- int
- xmlParseElementDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- int ret = -1;
- xmlElementContentPtr content = NULL;
- /* GROW; done in the caller */
- if (CMP9(CUR_PTR, '<', '!', 'E', 'L', 'E', 'M', 'E', 'N', 'T')) {
- int inputid = ctxt->input->id;
- SKIP(9);
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after 'ELEMENT'\n");
- return(-1);
- }
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseElementDecl: no name for Element\n");
- return(-1);
- }
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space required after the element name\n");
- }
- if (CMP5(CUR_PTR, 'E', 'M', 'P', 'T', 'Y')) {
- SKIP(5);
- /*
- * Element must always be empty.
- */
- ret = XML_ELEMENT_TYPE_EMPTY;
- } else if ((RAW == 'A') && (NXT(1) == 'N') &&
- (NXT(2) == 'Y')) {
- SKIP(3);
- /*
- * Element is a generic container.
- */
- ret = XML_ELEMENT_TYPE_ANY;
- } else if (RAW == '(') {
- ret = xmlParseElementContentDecl(ctxt, name, &content);
- } else {
- /*
- * [ WFC: PEs in Internal Subset ] error handling.
- */
- if ((RAW == '%') && (ctxt->external == 0) &&
- (ctxt->inputNr == 1)) {
- xmlFatalErrMsg(ctxt, XML_ERR_PEREF_IN_INT_SUBSET,
- "PEReference: forbidden within markup decl in internal subset\n");
- } else {
- xmlFatalErrMsg(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED,
- "xmlParseElementDecl: 'EMPTY', 'ANY' or '(' expected\n");
- }
- return(-1);
- }
- SKIP_BLANKS;
- if (RAW != '>') {
- xmlFatalErr(ctxt, XML_ERR_GT_REQUIRED, NULL);
- if (content != NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, content);
- }
- } else {
- if (inputid != ctxt->input->id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "Element declaration doesn't start and stop in"
- " the same entity\n");
- }
- NEXT;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->elementDecl != NULL)) {
- if (content != NULL)
- content->parent = NULL;
- ctxt->sax->elementDecl(ctxt->userData, name, ret,
- content);
- if ((content != NULL) && (content->parent == NULL)) {
- /*
- * this is a trick: if xmlAddElementDecl is called,
- * instead of copying the full tree it is plugged directly
- * if called from the parser. Avoid duplicating the
- * interfaces or change the API/ABI
- */
- xmlFreeDocElementContent(ctxt->myDoc, content);
- }
- } else if (content != NULL) {
- xmlFreeDocElementContent(ctxt->myDoc, content);
- }
- }
- }
- return(ret);
- }
- /**
- * xmlParseConditionalSections
- * @ctxt: an XML parser context
- *
- * [61] conditionalSect ::= includeSect | ignoreSect
- * [62] includeSect ::= '<![' S? 'INCLUDE' S? '[' extSubsetDecl ']]>'
- * [63] ignoreSect ::= '<![' S? 'IGNORE' S? '[' ignoreSectContents* ']]>'
- * [64] ignoreSectContents ::= Ignore ('<![' ignoreSectContents ']]>' Ignore)*
- * [65] Ignore ::= Char* - (Char* ('<![' | ']]>') Char*)
- */
- static void
- xmlParseConditionalSections(xmlParserCtxtPtr ctxt) {
- int *inputIds = NULL;
- size_t inputIdsSize = 0;
- size_t depth = 0;
- while (ctxt->instate != XML_PARSER_EOF) {
- if ((RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- int id = ctxt->input->id;
- SKIP(3);
- SKIP_BLANKS;
- if (CMP7(CUR_PTR, 'I', 'N', 'C', 'L', 'U', 'D', 'E')) {
- SKIP(7);
- SKIP_BLANKS;
- if (RAW != '[') {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_INVALID, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- if (ctxt->input->id != id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is"
- " not in the same entity\n");
- }
- NEXT;
- if (inputIdsSize <= depth) {
- int *tmp;
- inputIdsSize = (inputIdsSize == 0 ? 4 : inputIdsSize * 2);
- tmp = (int *) xmlRealloc(inputIds,
- inputIdsSize * sizeof(int));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- goto error;
- }
- inputIds = tmp;
- }
- inputIds[depth] = id;
- depth++;
- } else if (CMP6(CUR_PTR, 'I', 'G', 'N', 'O', 'R', 'E')) {
- int state;
- xmlParserInputState instate;
- size_t ignoreDepth = 0;
- SKIP(6);
- SKIP_BLANKS;
- if (RAW != '[') {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_INVALID, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- if (ctxt->input->id != id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is"
- " not in the same entity\n");
- }
- NEXT;
- /*
- * Parse up to the end of the conditional section but disable
- * SAX event generating DTD building in the meantime
- */
- state = ctxt->disableSAX;
- instate = ctxt->instate;
- if (ctxt->recovery == 0) ctxt->disableSAX = 1;
- ctxt->instate = XML_PARSER_IGNORE;
- while (RAW != 0) {
- if ((RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- SKIP(3);
- ignoreDepth++;
- /* Check for integer overflow */
- if (ignoreDepth == 0) {
- xmlErrMemory(ctxt, NULL);
- goto error;
- }
- } else if ((RAW == ']') && (NXT(1) == ']') &&
- (NXT(2) == '>')) {
- if (ignoreDepth == 0)
- break;
- SKIP(3);
- ignoreDepth--;
- } else {
- NEXT;
- }
- }
- ctxt->disableSAX = state;
- ctxt->instate = instate;
- if (RAW == 0) {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_NOT_FINISHED, NULL);
- goto error;
- }
- if (ctxt->input->id != id) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is"
- " not in the same entity\n");
- }
- SKIP(3);
- } else {
- xmlFatalErr(ctxt, XML_ERR_CONDSEC_INVALID_KEYWORD, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- } else if ((depth > 0) &&
- (RAW == ']') && (NXT(1) == ']') && (NXT(2) == '>')) {
- depth--;
- if (ctxt->input->id != inputIds[depth]) {
- xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
- "All markup of the conditional section is not"
- " in the same entity\n");
- }
- SKIP(3);
- } else {
- const xmlChar *check = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- xmlParseMarkupDecl(ctxt);
- if ((CUR_PTR == check) && (cons == ctxt->input->consumed)) {
- xmlFatalErr(ctxt, XML_ERR_EXT_SUBSET_NOT_FINISHED, NULL);
- xmlHaltParser(ctxt);
- goto error;
- }
- }
- if (depth == 0)
- break;
- SKIP_BLANKS;
- GROW;
- }
- error:
- xmlFree(inputIds);
- }
- /**
- * xmlParseMarkupDecl:
- * @ctxt: an XML parser context
- *
- * parse Markup declarations
- *
- * [29] markupdecl ::= elementdecl | AttlistDecl | EntityDecl |
- * NotationDecl | PI | Comment
- *
- * [ VC: Proper Declaration/PE Nesting ]
- * Parameter-entity replacement text must be properly nested with
- * markup declarations. That is to say, if either the first character
- * or the last character of a markup declaration (markupdecl above) is
- * contained in the replacement text for a parameter-entity reference,
- * both must be contained in the same replacement text.
- *
- * [ WFC: PEs in Internal Subset ]
- * In the internal DTD subset, parameter-entity references can occur
- * only where markup declarations can occur, not within markup declarations.
- * (This does not apply to references that occur in external parameter
- * entities or to the external subset.)
- */
- void
- xmlParseMarkupDecl(xmlParserCtxtPtr ctxt) {
- GROW;
- if (CUR == '<') {
- if (NXT(1) == '!') {
- switch (NXT(2)) {
- case 'E':
- if (NXT(3) == 'L')
- xmlParseElementDecl(ctxt);
- else if (NXT(3) == 'N')
- xmlParseEntityDecl(ctxt);
- break;
- case 'A':
- xmlParseAttributeListDecl(ctxt);
- break;
- case 'N':
- xmlParseNotationDecl(ctxt);
- break;
- case '-':
- xmlParseComment(ctxt);
- break;
- default:
- /* there is an error but it will be detected later */
- break;
- }
- } else if (NXT(1) == '?') {
- xmlParsePI(ctxt);
- }
- }
- /*
- * detect requirement to exit there and act accordingly
- * and avoid having instate overridden later on
- */
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- ctxt->instate = XML_PARSER_DTD;
- }
- /**
- * xmlParseTextDecl:
- * @ctxt: an XML parser context
- *
- * parse an XML declaration header for external entities
- *
- * [77] TextDecl ::= '<?xml' VersionInfo? EncodingDecl S? '?>'
- */
- void
- xmlParseTextDecl(xmlParserCtxtPtr ctxt) {
- xmlChar *version;
- const xmlChar *encoding;
- int oldstate;
- /*
- * We know that '<?xml' is here.
- */
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- SKIP(5);
- } else {
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_STARTED, NULL);
- return;
- }
- /* Avoid expansion of parameter entities when skipping blanks. */
- oldstate = ctxt->instate;
- ctxt->instate = XML_PARSER_START;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space needed after '<?xml'\n");
- }
- /*
- * We may have the VersionInfo here.
- */
- version = xmlParseVersionInfo(ctxt);
- if (version == NULL)
- version = xmlCharStrdup(XML_DEFAULT_VERSION);
- else {
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Space needed here\n");
- }
- }
- ctxt->input->version = version;
- /*
- * We must have the encoding declaration
- */
- encoding = xmlParseEncodingDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- ctxt->instate = oldstate;
- return;
- }
- if ((encoding == NULL) && (ctxt->errNo == XML_ERR_OK)) {
- xmlFatalErrMsg(ctxt, XML_ERR_MISSING_ENCODING,
- "Missing encoding in text declaration\n");
- }
- SKIP_BLANKS;
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- } else if (RAW == '>') {
- /* Deprecated old WD ... */
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- MOVETO_ENDTAG(CUR_PTR);
- NEXT;
- }
- ctxt->instate = oldstate;
- }
- /**
- * xmlParseExternalSubset:
- * @ctxt: an XML parser context
- * @ExternalID: the external identifier
- * @SystemID: the system identifier (or URL)
- *
- * parse Markup declarations from an external subset
- *
- * [30] extSubset ::= textDecl? extSubsetDecl
- *
- * [31] extSubsetDecl ::= (markupdecl | conditionalSect | PEReference | S) *
- */
- void
- xmlParseExternalSubset(xmlParserCtxtPtr ctxt, const xmlChar *ExternalID,
- const xmlChar *SystemID) {
- xmlDetectSAX2(ctxt);
- GROW;
- if ((ctxt->encoding == NULL) &&
- (ctxt->input->end - ctxt->input->cur >= 4)) {
- xmlChar start[4];
- xmlCharEncoding enc;
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE)
- xmlSwitchEncoding(ctxt, enc);
- }
- if (CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) {
- xmlParseTextDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- xmlHaltParser(ctxt);
- return;
- }
- }
- if (ctxt->myDoc == NULL) {
- ctxt->myDoc = xmlNewDoc(BAD_CAST "1.0");
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return;
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- }
- if ((ctxt->myDoc != NULL) && (ctxt->myDoc->intSubset == NULL))
- xmlCreateIntSubset(ctxt->myDoc, NULL, ExternalID, SystemID);
- ctxt->instate = XML_PARSER_DTD;
- ctxt->external = 1;
- SKIP_BLANKS;
- while (((RAW == '<') && (NXT(1) == '?')) ||
- ((RAW == '<') && (NXT(1) == '!')) ||
- (RAW == '%')) {
- const xmlChar *check = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- GROW;
- if ((RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- xmlParseConditionalSections(ctxt);
- } else
- xmlParseMarkupDecl(ctxt);
- SKIP_BLANKS;
- if ((CUR_PTR == check) && (cons == ctxt->input->consumed)) {
- xmlFatalErr(ctxt, XML_ERR_EXT_SUBSET_NOT_FINISHED, NULL);
- break;
- }
- }
- if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXT_SUBSET_NOT_FINISHED, NULL);
- }
- }
- /**
- * xmlParseReference:
- * @ctxt: an XML parser context
- *
- * parse and handle entity references in content, depending on the SAX
- * interface, this may end-up in a call to character() if this is a
- * CharRef, a predefined entity, if there is no reference() callback.
- * or if the parser was asked to switch to that mode.
- *
- * [67] Reference ::= EntityRef | CharRef
- */
- void
- xmlParseReference(xmlParserCtxtPtr ctxt) {
- xmlEntityPtr ent;
- xmlChar *val;
- int was_checked;
- xmlNodePtr list = NULL;
- xmlParserErrors ret = XML_ERR_OK;
- if (RAW != '&')
- return;
- /*
- * Simple case of a CharRef
- */
- if (NXT(1) == '#') {
- int i = 0;
- xmlChar out[16];
- int hex = NXT(2);
- int value = xmlParseCharRef(ctxt);
- if (value == 0)
- return;
- if (ctxt->charset != XML_CHAR_ENCODING_UTF8) {
- /*
- * So we are using non-UTF-8 buffers
- * Check that the char fit on 8bits, if not
- * generate a CharRef.
- */
- if (value <= 0xFF) {
- out[0] = value;
- out[1] = 0;
- if ((ctxt->sax != NULL) && (ctxt->sax->characters != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->characters(ctxt->userData, out, 1);
- } else {
- if ((hex == 'x') || (hex == 'X'))
- snprintf((char *)out, sizeof(out), "#x%X", value);
- else
- snprintf((char *)out, sizeof(out), "#%d", value);
- if ((ctxt->sax != NULL) && (ctxt->sax->reference != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->reference(ctxt->userData, out);
- }
- } else {
- /*
- * Just encode the value in UTF-8
- */
- COPY_BUF(0 ,out, i, value);
- out[i] = 0;
- if ((ctxt->sax != NULL) && (ctxt->sax->characters != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->characters(ctxt->userData, out, i);
- }
- return;
- }
- /*
- * We are seeing an entity reference
- */
- ent = xmlParseEntityRef(ctxt);
- if (ent == NULL) return;
- if (!ctxt->wellFormed)
- return;
- was_checked = ent->checked;
- /* special case of predefined entities */
- if ((ent->name == NULL) ||
- (ent->etype == XML_INTERNAL_PREDEFINED_ENTITY)) {
- val = ent->content;
- if (val == NULL) return;
- /*
- * inline the entity.
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->characters != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->characters(ctxt->userData, val, xmlStrlen(val));
- return;
- }
- /*
- * The first reference to the entity trigger a parsing phase
- * where the ent->children is filled with the result from
- * the parsing.
- * Note: external parsed entities will not be loaded, it is not
- * required for a non-validating parser, unless the parsing option
- * of validating, or substituting entities were given. Doing so is
- * far more secure as the parser will only process data coming from
- * the document entity by default.
- */
- if (((ent->checked == 0) ||
- ((ent->children == NULL) && (ctxt->options & XML_PARSE_NOENT))) &&
- ((ent->etype != XML_EXTERNAL_GENERAL_PARSED_ENTITY) ||
- (ctxt->options & (XML_PARSE_NOENT | XML_PARSE_DTDVALID)))) {
- unsigned long oldnbent = ctxt->nbentities, diff;
- /*
- * This is a bit hackish but this seems the best
- * way to make sure both SAX and DOM entity support
- * behaves okay.
- */
- void *user_data;
- if (ctxt->userData == ctxt)
- user_data = NULL;
- else
- user_data = ctxt->userData;
- /*
- * Check that this entity is well formed
- * 4.3.2: An internal general parsed entity is well-formed
- * if its replacement text matches the production labeled
- * content.
- */
- if (ent->etype == XML_INTERNAL_GENERAL_ENTITY) {
- ctxt->depth++;
- ret = xmlParseBalancedChunkMemoryInternal(ctxt, ent->content,
- user_data, &list);
- ctxt->depth--;
- } else if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY) {
- ctxt->depth++;
- ret = xmlParseExternalEntityPrivate(ctxt->myDoc, ctxt, ctxt->sax,
- user_data, ctxt->depth, ent->URI,
- ent->ExternalID, &list);
- ctxt->depth--;
- } else {
- ret = XML_ERR_ENTITY_PE_INTERNAL;
- xmlErrMsgStr(ctxt, XML_ERR_INTERNAL_ERROR,
- "invalid entity type found\n", NULL);
- }
- /*
- * Store the number of entities needing parsing for this entity
- * content and do checkings
- */
- diff = ctxt->nbentities - oldnbent + 1;
- if (diff > INT_MAX / 2)
- diff = INT_MAX / 2;
- ent->checked = diff * 2;
- if ((ent->content != NULL) && (xmlStrchr(ent->content, '<')))
- ent->checked |= 1;
- if (ret == XML_ERR_ENTITY_LOOP) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- xmlHaltParser(ctxt);
- xmlFreeNodeList(list);
- return;
- }
- if (xmlParserEntityCheck(ctxt, 0, ent, 0)) {
- xmlFreeNodeList(list);
- return;
- }
- if ((ret == XML_ERR_OK) && (list != NULL)) {
- if (((ent->etype == XML_INTERNAL_GENERAL_ENTITY) ||
- (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY))&&
- (ent->children == NULL)) {
- ent->children = list;
- /*
- * Prune it directly in the generated document
- * except for single text nodes.
- */
- if ((ctxt->replaceEntities == 0) ||
- (ctxt->parseMode == XML_PARSE_READER) ||
- ((list->type == XML_TEXT_NODE) &&
- (list->next == NULL))) {
- ent->owner = 1;
- while (list != NULL) {
- list->parent = (xmlNodePtr) ent;
- xmlSetTreeDoc(list, ent->doc);
- if (list->next == NULL)
- ent->last = list;
- list = list->next;
- }
- list = NULL;
- } else {
- ent->owner = 0;
- while (list != NULL) {
- list->parent = (xmlNodePtr) ctxt->node;
- list->doc = ctxt->myDoc;
- if (list->next == NULL)
- ent->last = list;
- list = list->next;
- }
- list = ent->children;
- #ifdef LIBXML_LEGACY_ENABLED
- if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)
- xmlAddEntityReference(ent, list, NULL);
- #endif /* LIBXML_LEGACY_ENABLED */
- }
- } else {
- xmlFreeNodeList(list);
- list = NULL;
- }
- } else if ((ret != XML_ERR_OK) &&
- (ret != XML_WAR_UNDECLARED_ENTITY)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "Entity '%s' failed to parse\n", ent->name);
- if (ent->content != NULL)
- ent->content[0] = 0;
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- } else if (list != NULL) {
- xmlFreeNodeList(list);
- list = NULL;
- }
- if (ent->checked == 0)
- ent->checked = 2;
- /* Prevent entity from being parsed and expanded twice (Bug 760367). */
- was_checked = 0;
- } else if (ent->checked != 1) {
- ctxt->nbentities += ent->checked / 2;
- }
- /*
- * Now that the entity content has been gathered
- * provide it to the application, this can take different forms based
- * on the parsing modes.
- */
- if (ent->children == NULL) {
- /*
- * Probably running in SAX mode and the callbacks don't
- * build the entity content. So unless we already went
- * though parsing for first checking go though the entity
- * content to generate callbacks associated to the entity
- */
- if (was_checked != 0) {
- void *user_data;
- /*
- * This is a bit hackish but this seems the best
- * way to make sure both SAX and DOM entity support
- * behaves okay.
- */
- if (ctxt->userData == ctxt)
- user_data = NULL;
- else
- user_data = ctxt->userData;
- if (ent->etype == XML_INTERNAL_GENERAL_ENTITY) {
- ctxt->depth++;
- ret = xmlParseBalancedChunkMemoryInternal(ctxt,
- ent->content, user_data, NULL);
- ctxt->depth--;
- } else if (ent->etype ==
- XML_EXTERNAL_GENERAL_PARSED_ENTITY) {
- ctxt->depth++;
- ret = xmlParseExternalEntityPrivate(ctxt->myDoc, ctxt,
- ctxt->sax, user_data, ctxt->depth,
- ent->URI, ent->ExternalID, NULL);
- ctxt->depth--;
- } else {
- ret = XML_ERR_ENTITY_PE_INTERNAL;
- xmlErrMsgStr(ctxt, XML_ERR_INTERNAL_ERROR,
- "invalid entity type found\n", NULL);
- }
- if (ret == XML_ERR_ENTITY_LOOP) {
- xmlFatalErr(ctxt, XML_ERR_ENTITY_LOOP, NULL);
- return;
- }
- }
- if ((ctxt->sax != NULL) && (ctxt->sax->reference != NULL) &&
- (ctxt->replaceEntities == 0) && (!ctxt->disableSAX)) {
- /*
- * Entity reference callback comes second, it's somewhat
- * superfluous but a compatibility to historical behaviour
- */
- ctxt->sax->reference(ctxt->userData, ent->name);
- }
- return;
- }
- /*
- * If we didn't get any children for the entity being built
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->reference != NULL) &&
- (ctxt->replaceEntities == 0) && (!ctxt->disableSAX)) {
- /*
- * Create a node.
- */
- ctxt->sax->reference(ctxt->userData, ent->name);
- return;
- }
- if ((ctxt->replaceEntities) || (ent->children == NULL)) {
- /*
- * There is a problem on the handling of _private for entities
- * (bug 155816): Should we copy the content of the field from
- * the entity (possibly overwriting some value set by the user
- * when a copy is created), should we leave it alone, or should
- * we try to take care of different situations? The problem
- * is exacerbated by the usage of this field by the xmlReader.
- * To fix this bug, we look at _private on the created node
- * and, if it's NULL, we copy in whatever was in the entity.
- * If it's not NULL we leave it alone. This is somewhat of a
- * hack - maybe we should have further tests to determine
- * what to do.
- */
- if ((ctxt->node != NULL) && (ent->children != NULL)) {
- /*
- * Seems we are generating the DOM content, do
- * a simple tree copy for all references except the first
- * In the first occurrence list contains the replacement.
- */
- if (((list == NULL) && (ent->owner == 0)) ||
- (ctxt->parseMode == XML_PARSE_READER)) {
- xmlNodePtr nw = NULL, cur, firstChild = NULL;
- /*
- * We are copying here, make sure there is no abuse
- */
- ctxt->sizeentcopy += ent->length + 5;
- if (xmlParserEntityCheck(ctxt, 0, ent, ctxt->sizeentcopy))
- return;
- /*
- * when operating on a reader, the entities definitions
- * are always owning the entities subtree.
- if (ctxt->parseMode == XML_PARSE_READER)
- ent->owner = 1;
- */
- cur = ent->children;
- while (cur != NULL) {
- nw = xmlDocCopyNode(cur, ctxt->myDoc, 1);
- if (nw != NULL) {
- if (nw->_private == NULL)
- nw->_private = cur->_private;
- if (firstChild == NULL){
- firstChild = nw;
- }
- nw = xmlAddChild(ctxt->node, nw);
- }
- if (cur == ent->last) {
- /*
- * needed to detect some strange empty
- * node cases in the reader tests
- */
- if ((ctxt->parseMode == XML_PARSE_READER) &&
- (nw != NULL) &&
- (nw->type == XML_ELEMENT_NODE) &&
- (nw->children == NULL))
- nw->extra = 1;
- break;
- }
- cur = cur->next;
- }
- #ifdef LIBXML_LEGACY_ENABLED
- if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)
- xmlAddEntityReference(ent, firstChild, nw);
- #endif /* LIBXML_LEGACY_ENABLED */
- } else if ((list == NULL) || (ctxt->inputNr > 0)) {
- xmlNodePtr nw = NULL, cur, next, last,
- firstChild = NULL;
- /*
- * We are copying here, make sure there is no abuse
- */
- ctxt->sizeentcopy += ent->length + 5;
- if (xmlParserEntityCheck(ctxt, 0, ent, ctxt->sizeentcopy))
- return;
- /*
- * Copy the entity child list and make it the new
- * entity child list. The goal is to make sure any
- * ID or REF referenced will be the one from the
- * document content and not the entity copy.
- */
- cur = ent->children;
- ent->children = NULL;
- last = ent->last;
- ent->last = NULL;
- while (cur != NULL) {
- next = cur->next;
- cur->next = NULL;
- cur->parent = NULL;
- nw = xmlDocCopyNode(cur, ctxt->myDoc, 1);
- if (nw != NULL) {
- if (nw->_private == NULL)
- nw->_private = cur->_private;
- if (firstChild == NULL){
- firstChild = cur;
- }
- xmlAddChild((xmlNodePtr) ent, nw);
- xmlAddChild(ctxt->node, cur);
- }
- if (cur == last)
- break;
- cur = next;
- }
- if (ent->owner == 0)
- ent->owner = 1;
- #ifdef LIBXML_LEGACY_ENABLED
- if (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)
- xmlAddEntityReference(ent, firstChild, nw);
- #endif /* LIBXML_LEGACY_ENABLED */
- } else {
- const xmlChar *nbktext;
- /*
- * the name change is to avoid coalescing of the
- * node with a possible previous text one which
- * would make ent->children a dangling pointer
- */
- nbktext = xmlDictLookup(ctxt->dict, BAD_CAST "nbktext",
- -1);
- if (ent->children->type == XML_TEXT_NODE)
- ent->children->name = nbktext;
- if ((ent->last != ent->children) &&
- (ent->last->type == XML_TEXT_NODE))
- ent->last->name = nbktext;
- xmlAddChildList(ctxt->node, ent->children);
- }
- /*
- * This is to avoid a nasty side effect, see
- * characters() in SAX.c
- */
- ctxt->nodemem = 0;
- ctxt->nodelen = 0;
- return;
- }
- }
- }
- /**
- * xmlParseEntityRef:
- * @ctxt: an XML parser context
- *
- * parse ENTITY references declarations
- *
- * [68] EntityRef ::= '&' Name ';'
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", the Name given in the entity reference
- * must match that in an entity declaration, except that well-formed
- * documents need not declare any of the following entities: amp, lt,
- * gt, apos, quot. The declaration of a parameter entity must precede
- * any reference to it. Similarly, the declaration of a general entity
- * must precede any reference to it which appears in a default value in an
- * attribute-list declaration. Note that if entities are declared in the
- * external subset or in external parameter entities, a non-validating
- * processor is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be declared is a
- * well-formedness constraint only if standalone='yes'.
- *
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an unparsed entity
- *
- * Returns the xmlEntityPtr if found, or NULL otherwise.
- */
- xmlEntityPtr
- xmlParseEntityRef(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- xmlEntityPtr ent = NULL;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- if (RAW != '&')
- return(NULL);
- NEXT;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseEntityRef: no name\n");
- return(NULL);
- }
- if (RAW != ';') {
- xmlFatalErr(ctxt, XML_ERR_ENTITYREF_SEMICOL_MISSING, NULL);
- return(NULL);
- }
- NEXT;
- /*
- * Predefined entities override any extra definition
- */
- if ((ctxt->options & XML_PARSE_OLDSAX) == 0) {
- ent = xmlGetPredefinedEntity(name);
- if (ent != NULL)
- return(ent);
- }
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Ask first SAX for entity resolution, otherwise try the
- * entities which may have stored in the parser context.
- */
- if (ctxt->sax != NULL) {
- if (ctxt->sax->getEntity != NULL)
- ent = ctxt->sax->getEntity(ctxt->userData, name);
- if ((ctxt->wellFormed == 1 ) && (ent == NULL) &&
- (ctxt->options & XML_PARSE_OLDSAX))
- ent = xmlGetPredefinedEntity(name);
- if ((ctxt->wellFormed == 1 ) && (ent == NULL) &&
- (ctxt->userData==ctxt)) {
- ent = xmlSAX2GetEntity(ctxt, name);
- }
- }
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", the
- * Name given in the entity reference must match that in an
- * entity declaration, except that well-formed documents
- * need not declare any of the following entities: amp, lt,
- * gt, apos, quot.
- * The declaration of a parameter entity must precede any
- * reference to it.
- * Similarly, the declaration of a general entity must
- * precede any reference to it which appears in a default
- * value in an attribute-list declaration. Note that if
- * entities are declared in the external subset or in
- * external parameter entities, a non-validating processor
- * is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be
- * declared is a well-formedness constraint only if
- * standalone='yes'.
- */
- if (ent == NULL) {
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) &&
- (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n", name);
- } else {
- xmlErrMsgStr(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n", name);
- if ((ctxt->inSubset == 0) &&
- (ctxt->sax != NULL) &&
- (ctxt->sax->reference != NULL)) {
- ctxt->sax->reference(ctxt->userData, name);
- }
- }
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- ctxt->valid = 0;
- }
- /*
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an
- * unparsed entity
- */
- else if (ent->etype == XML_EXTERNAL_GENERAL_UNPARSED_ENTITY) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNPARSED_ENTITY,
- "Entity reference to unparsed entity %s\n", name);
- }
- /*
- * [ WFC: No External Entity References ]
- * Attribute values cannot contain direct or indirect
- * entity references to external entities.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_EXTERNAL,
- "Attribute references external entity '%s'\n", name);
- }
- /*
- * [ WFC: No < in Attribute Values ]
- * The replacement text of any entity referred to directly or
- * indirectly in an attribute value (other than "<") must
- * not contain a <.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent != NULL) &&
- (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY)) {
- if (((ent->checked & 1) || (ent->checked == 0)) &&
- (ent->content != NULL) && (xmlStrchr(ent->content, '<'))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_LT_IN_ATTRIBUTE,
- "'<' in entity '%s' is not allowed in attributes values\n", name);
- }
- }
- /*
- * Internal check, no parameter entities here ...
- */
- else {
- switch (ent->etype) {
- case XML_INTERNAL_PARAMETER_ENTITY:
- case XML_EXTERNAL_PARAMETER_ENTITY:
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_PARAMETER,
- "Attempt to reference the parameter entity '%s'\n",
- name);
- break;
- default:
- break;
- }
- }
- /*
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive reference
- * to itself, either directly or indirectly.
- * Done somewhere else
- */
- return(ent);
- }
- /**
- * xmlParseStringEntityRef:
- * @ctxt: an XML parser context
- * @str: a pointer to an index in the string
- *
- * parse ENTITY references declarations, but this version parses it from
- * a string value.
- *
- * [68] EntityRef ::= '&' Name ';'
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", the Name given in the entity reference
- * must match that in an entity declaration, except that well-formed
- * documents need not declare any of the following entities: amp, lt,
- * gt, apos, quot. The declaration of a parameter entity must precede
- * any reference to it. Similarly, the declaration of a general entity
- * must precede any reference to it which appears in a default value in an
- * attribute-list declaration. Note that if entities are declared in the
- * external subset or in external parameter entities, a non-validating
- * processor is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be declared is a
- * well-formedness constraint only if standalone='yes'.
- *
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an unparsed entity
- *
- * Returns the xmlEntityPtr if found, or NULL otherwise. The str pointer
- * is updated to the current location in the string.
- */
- static xmlEntityPtr
- xmlParseStringEntityRef(xmlParserCtxtPtr ctxt, const xmlChar ** str) {
- xmlChar *name;
- const xmlChar *ptr;
- xmlChar cur;
- xmlEntityPtr ent = NULL;
- if ((str == NULL) || (*str == NULL))
- return(NULL);
- ptr = *str;
- cur = *ptr;
- if (cur != '&')
- return(NULL);
- ptr++;
- name = xmlParseStringName(ctxt, &ptr);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseStringEntityRef: no name\n");
- *str = ptr;
- return(NULL);
- }
- if (*ptr != ';') {
- xmlFatalErr(ctxt, XML_ERR_ENTITYREF_SEMICOL_MISSING, NULL);
- xmlFree(name);
- *str = ptr;
- return(NULL);
- }
- ptr++;
- /*
- * Predefined entities override any extra definition
- */
- if ((ctxt->options & XML_PARSE_OLDSAX) == 0) {
- ent = xmlGetPredefinedEntity(name);
- if (ent != NULL) {
- xmlFree(name);
- *str = ptr;
- return(ent);
- }
- }
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Ask first SAX for entity resolution, otherwise try the
- * entities which may have stored in the parser context.
- */
- if (ctxt->sax != NULL) {
- if (ctxt->sax->getEntity != NULL)
- ent = ctxt->sax->getEntity(ctxt->userData, name);
- if ((ent == NULL) && (ctxt->options & XML_PARSE_OLDSAX))
- ent = xmlGetPredefinedEntity(name);
- if ((ent == NULL) && (ctxt->userData==ctxt)) {
- ent = xmlSAX2GetEntity(ctxt, name);
- }
- }
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(name);
- return(NULL);
- }
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", the
- * Name given in the entity reference must match that in an
- * entity declaration, except that well-formed documents
- * need not declare any of the following entities: amp, lt,
- * gt, apos, quot.
- * The declaration of a parameter entity must precede any
- * reference to it.
- * Similarly, the declaration of a general entity must
- * precede any reference to it which appears in a default
- * value in an attribute-list declaration. Note that if
- * entities are declared in the external subset or in
- * external parameter entities, a non-validating processor
- * is not obligated to read and process their declarations;
- * for such documents, the rule that an entity must be
- * declared is a well-formedness constraint only if
- * standalone='yes'.
- */
- if (ent == NULL) {
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) &&
- (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n", name);
- } else {
- xmlErrMsgStr(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "Entity '%s' not defined\n",
- name);
- }
- xmlParserEntityCheck(ctxt, 0, ent, 0);
- /* TODO ? check regressions ctxt->valid = 0; */
- }
- /*
- * [ WFC: Parsed Entity ]
- * An entity reference must not contain the name of an
- * unparsed entity
- */
- else if (ent->etype == XML_EXTERNAL_GENERAL_UNPARSED_ENTITY) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNPARSED_ENTITY,
- "Entity reference to unparsed entity %s\n", name);
- }
- /*
- * [ WFC: No External Entity References ]
- * Attribute values cannot contain direct or indirect
- * entity references to external entities.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent->etype == XML_EXTERNAL_GENERAL_PARSED_ENTITY)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_EXTERNAL,
- "Attribute references external entity '%s'\n", name);
- }
- /*
- * [ WFC: No < in Attribute Values ]
- * The replacement text of any entity referred to directly or
- * indirectly in an attribute value (other than "<") must
- * not contain a <.
- */
- else if ((ctxt->instate == XML_PARSER_ATTRIBUTE_VALUE) &&
- (ent != NULL) && (ent->content != NULL) &&
- (ent->etype != XML_INTERNAL_PREDEFINED_ENTITY) &&
- (xmlStrchr(ent->content, '<'))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_LT_IN_ATTRIBUTE,
- "'<' in entity '%s' is not allowed in attributes values\n",
- name);
- }
- /*
- * Internal check, no parameter entities here ...
- */
- else {
- switch (ent->etype) {
- case XML_INTERNAL_PARAMETER_ENTITY:
- case XML_EXTERNAL_PARAMETER_ENTITY:
- xmlFatalErrMsgStr(ctxt, XML_ERR_ENTITY_IS_PARAMETER,
- "Attempt to reference the parameter entity '%s'\n",
- name);
- break;
- default:
- break;
- }
- }
- /*
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive reference
- * to itself, either directly or indirectly.
- * Done somewhere else
- */
- xmlFree(name);
- *str = ptr;
- return(ent);
- }
- /**
- * xmlParsePEReference:
- * @ctxt: an XML parser context
- *
- * parse PEReference declarations
- * The entity content is handled directly by pushing it's content as
- * a new input stream.
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive
- * reference to itself, either directly or indirectly.
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", ... ... The declaration of a parameter
- * entity must precede any reference to it...
- *
- * [ VC: Entity Declared ]
- * In a document with an external subset or external parameter entities
- * with "standalone='no'", ... ... The declaration of a parameter entity
- * must precede any reference to it...
- *
- * [ WFC: In DTD ]
- * Parameter-entity references may only appear in the DTD.
- * NOTE: misleading but this is handled.
- */
- void
- xmlParsePEReference(xmlParserCtxtPtr ctxt)
- {
- const xmlChar *name;
- xmlEntityPtr entity = NULL;
- xmlParserInputPtr input;
- if (RAW != '%')
- return;
- NEXT;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_PEREF_NO_NAME, "PEReference: no name\n");
- return;
- }
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "PEReference: %s\n", name);
- if (RAW != ';') {
- xmlFatalErr(ctxt, XML_ERR_PEREF_SEMICOL_MISSING, NULL);
- return;
- }
- NEXT;
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Request the entity from SAX
- */
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getParameterEntity != NULL))
- entity = ctxt->sax->getParameterEntity(ctxt->userData, name);
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- if (entity == NULL) {
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", ...
- * ... The declaration of a parameter entity must precede
- * any reference to it...
- */
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) &&
- (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name);
- } else {
- /*
- * [ VC: Entity Declared ]
- * In a document with an external subset or external
- * parameter entities with "standalone='no'", ...
- * ... The declaration of a parameter entity must
- * precede any reference to it...
- */
- if ((ctxt->validate) && (ctxt->vctxt.error != NULL)) {
- xmlValidityError(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name, NULL);
- } else
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name, NULL);
- ctxt->valid = 0;
- }
- xmlParserEntityCheck(ctxt, 0, NULL, 0);
- } else {
- /*
- * Internal checking in case the entity quest barfed
- */
- if ((entity->etype != XML_INTERNAL_PARAMETER_ENTITY) &&
- (entity->etype != XML_EXTERNAL_PARAMETER_ENTITY)) {
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "Internal: %%%s; is not a parameter entity\n",
- name, NULL);
- } else {
- xmlChar start[4];
- xmlCharEncoding enc;
- if (xmlParserEntityCheck(ctxt, 0, entity, 0))
- return;
- if ((entity->etype == XML_EXTERNAL_PARAMETER_ENTITY) &&
- ((ctxt->options & XML_PARSE_NOENT) == 0) &&
- ((ctxt->options & XML_PARSE_DTDVALID) == 0) &&
- ((ctxt->options & XML_PARSE_DTDLOAD) == 0) &&
- ((ctxt->options & XML_PARSE_DTDATTR) == 0) &&
- (ctxt->replaceEntities == 0) &&
- (ctxt->validate == 0))
- return;
- input = xmlNewEntityInputStream(ctxt, entity);
- if (xmlPushInput(ctxt, input) < 0) {
- xmlFreeInputStream(input);
- return;
- }
- if (entity->etype == XML_EXTERNAL_PARAMETER_ENTITY) {
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- * Note that, since we may have some non-UTF8
- * encoding (like UTF16, bug 135229), the 'length'
- * is not known, but we can calculate based upon
- * the amount of data in the buffer.
- */
- GROW
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- if ((ctxt->input->end - ctxt->input->cur)>=4) {
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) &&
- (IS_BLANK_CH(NXT(5)))) {
- xmlParseTextDecl(ctxt);
- }
- }
- }
- }
- ctxt->hasPErefs = 1;
- }
- /**
- * xmlLoadEntityContent:
- * @ctxt: an XML parser context
- * @entity: an unloaded system entity
- *
- * Load the original content of the given system entity from the
- * ExternalID/SystemID given. This is to be used for Included in Literal
- * http://www.w3.org/TR/REC-xml/#inliteral processing of entities references
- *
- * Returns 0 in case of success and -1 in case of failure
- */
- static int
- xmlLoadEntityContent(xmlParserCtxtPtr ctxt, xmlEntityPtr entity) {
- xmlParserInputPtr input;
- xmlBufferPtr buf;
- int l, c;
- int count = 0;
- if ((ctxt == NULL) || (entity == NULL) ||
- ((entity->etype != XML_EXTERNAL_PARAMETER_ENTITY) &&
- (entity->etype != XML_EXTERNAL_GENERAL_PARSED_ENTITY)) ||
- (entity->content != NULL)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlLoadEntityContent parameter error");
- return(-1);
- }
- if (xmlParserDebugEntities)
- xmlGenericError(xmlGenericErrorContext,
- "Reading %s entity content input\n", entity->name);
- buf = xmlBufferCreate();
- if (buf == NULL) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlLoadEntityContent parameter error");
- return(-1);
- }
- input = xmlNewEntityInputStream(ctxt, entity);
- if (input == NULL) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlLoadEntityContent input error");
- xmlBufferFree(buf);
- return(-1);
- }
- /*
- * Push the entity as the current input, read char by char
- * saving to the buffer until the end of the entity or an error
- */
- if (xmlPushInput(ctxt, input) < 0) {
- xmlBufferFree(buf);
- return(-1);
- }
- GROW;
- c = CUR_CHAR(l);
- while ((ctxt->input == input) && (ctxt->input->cur < ctxt->input->end) &&
- (IS_CHAR(c))) {
- xmlBufferAdd(buf, ctxt->input->cur, l);
- if (count++ > XML_PARSER_CHUNK_SIZE) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlBufferFree(buf);
- return(-1);
- }
- }
- NEXTL(l);
- c = CUR_CHAR(l);
- if (c == 0) {
- count = 0;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlBufferFree(buf);
- return(-1);
- }
- c = CUR_CHAR(l);
- }
- }
- if ((ctxt->input == input) && (ctxt->input->cur >= ctxt->input->end)) {
- xmlPopInput(ctxt);
- } else if (!IS_CHAR(c)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INVALID_CHAR,
- "xmlLoadEntityContent: invalid char value %d\n",
- c);
- xmlBufferFree(buf);
- return(-1);
- }
- entity->content = buf->content;
- buf->content = NULL;
- xmlBufferFree(buf);
- return(0);
- }
- /**
- * xmlParseStringPEReference:
- * @ctxt: an XML parser context
- * @str: a pointer to an index in the string
- *
- * parse PEReference declarations
- *
- * [69] PEReference ::= '%' Name ';'
- *
- * [ WFC: No Recursion ]
- * A parsed entity must not contain a recursive
- * reference to itself, either directly or indirectly.
- *
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an internal DTD
- * subset which contains no parameter entity references, or a document
- * with "standalone='yes'", ... ... The declaration of a parameter
- * entity must precede any reference to it...
- *
- * [ VC: Entity Declared ]
- * In a document with an external subset or external parameter entities
- * with "standalone='no'", ... ... The declaration of a parameter entity
- * must precede any reference to it...
- *
- * [ WFC: In DTD ]
- * Parameter-entity references may only appear in the DTD.
- * NOTE: misleading but this is handled.
- *
- * Returns the string of the entity content.
- * str is updated to the current value of the index
- */
- static xmlEntityPtr
- xmlParseStringPEReference(xmlParserCtxtPtr ctxt, const xmlChar **str) {
- const xmlChar *ptr;
- xmlChar cur;
- xmlChar *name;
- xmlEntityPtr entity = NULL;
- if ((str == NULL) || (*str == NULL)) return(NULL);
- ptr = *str;
- cur = *ptr;
- if (cur != '%')
- return(NULL);
- ptr++;
- name = xmlParseStringName(ctxt, &ptr);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseStringPEReference: no name\n");
- *str = ptr;
- return(NULL);
- }
- cur = *ptr;
- if (cur != ';') {
- xmlFatalErr(ctxt, XML_ERR_ENTITYREF_SEMICOL_MISSING, NULL);
- xmlFree(name);
- *str = ptr;
- return(NULL);
- }
- ptr++;
- /*
- * Increase the number of entity references parsed
- */
- ctxt->nbentities++;
- /*
- * Request the entity from SAX
- */
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->getParameterEntity != NULL))
- entity = ctxt->sax->getParameterEntity(ctxt->userData, name);
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(name);
- *str = ptr;
- return(NULL);
- }
- if (entity == NULL) {
- /*
- * [ WFC: Entity Declared ]
- * In a document without any DTD, a document with only an
- * internal DTD subset which contains no parameter entity
- * references, or a document with "standalone='yes'", ...
- * ... The declaration of a parameter entity must precede
- * any reference to it...
- */
- if ((ctxt->standalone == 1) ||
- ((ctxt->hasExternalSubset == 0) && (ctxt->hasPErefs == 0))) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n", name);
- } else {
- /*
- * [ VC: Entity Declared ]
- * In a document with an external subset or external
- * parameter entities with "standalone='no'", ...
- * ... The declaration of a parameter entity must
- * precede any reference to it...
- */
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "PEReference: %%%s; not found\n",
- name, NULL);
- ctxt->valid = 0;
- }
- xmlParserEntityCheck(ctxt, 0, NULL, 0);
- } else {
- /*
- * Internal checking in case the entity quest barfed
- */
- if ((entity->etype != XML_INTERNAL_PARAMETER_ENTITY) &&
- (entity->etype != XML_EXTERNAL_PARAMETER_ENTITY)) {
- xmlWarningMsg(ctxt, XML_WAR_UNDECLARED_ENTITY,
- "%%%s; is not a parameter entity\n",
- name, NULL);
- }
- }
- ctxt->hasPErefs = 1;
- xmlFree(name);
- *str = ptr;
- return(entity);
- }
- /**
- * xmlParseDocTypeDecl:
- * @ctxt: an XML parser context
- *
- * parse a DOCTYPE declaration
- *
- * [28] doctypedecl ::= '<!DOCTYPE' S Name (S ExternalID)? S?
- * ('[' (markupdecl | PEReference | S)* ']' S?)? '>'
- *
- * [ VC: Root Element Type ]
- * The Name in the document type declaration must match the element
- * type of the root element.
- */
- void
- xmlParseDocTypeDecl(xmlParserCtxtPtr ctxt) {
- const xmlChar *name = NULL;
- xmlChar *ExternalID = NULL;
- xmlChar *URI = NULL;
- /*
- * We know that '<!DOCTYPE' has been detected.
- */
- SKIP(9);
- SKIP_BLANKS;
- /*
- * Parse the DOCTYPE name.
- */
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseDocTypeDecl : no DOCTYPE name !\n");
- }
- ctxt->intSubName = name;
- SKIP_BLANKS;
- /*
- * Check for SystemID and ExternalID
- */
- URI = xmlParseExternalID(ctxt, &ExternalID, 1);
- if ((URI != NULL) || (ExternalID != NULL)) {
- ctxt->hasExternalSubset = 1;
- }
- ctxt->extSubURI = URI;
- ctxt->extSubSystem = ExternalID;
- SKIP_BLANKS;
- /*
- * Create and update the internal subset.
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->internalSubset != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->internalSubset(ctxt->userData, name, ExternalID, URI);
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- /*
- * Is there any internal subset declarations ?
- * they are handled separately in xmlParseInternalSubset()
- */
- if (RAW == '[')
- return;
- /*
- * We should be at the end of the DOCTYPE declaration.
- */
- if (RAW != '>') {
- xmlFatalErr(ctxt, XML_ERR_DOCTYPE_NOT_FINISHED, NULL);
- }
- NEXT;
- }
- /**
- * xmlParseInternalSubset:
- * @ctxt: an XML parser context
- *
- * parse the internal subset declaration
- *
- * [28 end] ('[' (markupdecl | PEReference | S)* ']' S?)? '>'
- */
- static void
- xmlParseInternalSubset(xmlParserCtxtPtr ctxt) {
- /*
- * Is there any DTD definition ?
- */
- if (RAW == '[') {
- int baseInputNr = ctxt->inputNr;
- ctxt->instate = XML_PARSER_DTD;
- NEXT;
- /*
- * Parse the succession of Markup declarations and
- * PEReferences.
- * Subsequence (markupdecl | PEReference | S)*
- */
- while (((RAW != ']') || (ctxt->inputNr > baseInputNr)) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *check = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- SKIP_BLANKS;
- xmlParseMarkupDecl(ctxt);
- xmlParsePEReference(ctxt);
- /*
- * Conditional sections are allowed from external entities included
- * by PE References in the internal subset.
- */
- if ((ctxt->inputNr > 1) && (ctxt->input->filename != NULL) &&
- (RAW == '<') && (NXT(1) == '!') && (NXT(2) == '[')) {
- xmlParseConditionalSections(ctxt);
- }
- if ((CUR_PTR == check) && (cons == ctxt->input->consumed)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlParseInternalSubset: error detected in Markup declaration\n");
- if (ctxt->inputNr > baseInputNr)
- xmlPopInput(ctxt);
- else
- break;
- }
- }
- if (RAW == ']') {
- NEXT;
- SKIP_BLANKS;
- }
- }
- /*
- * We should be at the end of the DOCTYPE declaration.
- */
- if (RAW != '>') {
- xmlFatalErr(ctxt, XML_ERR_DOCTYPE_NOT_FINISHED, NULL);
- return;
- }
- NEXT;
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlParseAttribute:
- * @ctxt: an XML parser context
- * @value: a xmlChar ** used to store the value of the attribute
- *
- * parse an attribute
- *
- * [41] Attribute ::= Name Eq AttValue
- *
- * [ WFC: No External Entity References ]
- * Attribute values cannot contain direct or indirect entity references
- * to external entities.
- *
- * [ WFC: No < in Attribute Values ]
- * The replacement text of any entity referred to directly or indirectly in
- * an attribute value (other than "<") must not contain a <.
- *
- * [ VC: Attribute Value Type ]
- * The attribute must have been declared; the value must be of the type
- * declared for it.
- *
- * [25] Eq ::= S? '=' S?
- *
- * With namespace:
- *
- * [NS 11] Attribute ::= QName Eq AttValue
- *
- * Also the case QName == xmlns:??? is handled independently as a namespace
- * definition.
- *
- * Returns the attribute name, and the value in *value.
- */
- const xmlChar *
- xmlParseAttribute(xmlParserCtxtPtr ctxt, xmlChar **value) {
- const xmlChar *name;
- xmlChar *val;
- *value = NULL;
- GROW;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "error parsing attribute name\n");
- return(NULL);
- }
- /*
- * read the value
- */
- SKIP_BLANKS;
- if (RAW == '=') {
- NEXT;
- SKIP_BLANKS;
- val = xmlParseAttValue(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ATTRIBUTE_WITHOUT_VALUE,
- "Specification mandates value for attribute %s\n", name);
- return(NULL);
- }
- /*
- * Check that xml:lang conforms to the specification
- * No more registered as an error, just generate a warning now
- * since this was deprecated in XML second edition
- */
- if ((ctxt->pedantic) && (xmlStrEqual(name, BAD_CAST "xml:lang"))) {
- if (!xmlCheckLanguageID(val)) {
- xmlWarningMsg(ctxt, XML_WAR_LANG_VALUE,
- "Malformed value for xml:lang : %s\n",
- val, NULL);
- }
- }
- /*
- * Check that xml:space conforms to the specification
- */
- if (xmlStrEqual(name, BAD_CAST "xml:space")) {
- if (xmlStrEqual(val, BAD_CAST "default"))
- *(ctxt->space) = 0;
- else if (xmlStrEqual(val, BAD_CAST "preserve"))
- *(ctxt->space) = 1;
- else {
- xmlWarningMsg(ctxt, XML_WAR_SPACE_VALUE,
- "Invalid value \"%s\" for xml:space : \"default\" or \"preserve\" expected\n",
- val, NULL);
- }
- }
- *value = val;
- return(name);
- }
- /**
- * xmlParseStartTag:
- * @ctxt: an XML parser context
- *
- * parse a start of tag either for rule element or
- * EmptyElement. In both case we don't parse the tag closing chars.
- *
- * [40] STag ::= '<' Name (S Attribute)* S? '>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * [44] EmptyElemTag ::= '<' Name (S Attribute)* S? '/>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * With namespace:
- *
- * [NS 8] STag ::= '<' QName (S Attribute)* S? '>'
- *
- * [NS 10] EmptyElement ::= '<' QName (S Attribute)* S? '/>'
- *
- * Returns the element name parsed
- */
- const xmlChar *
- xmlParseStartTag(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- const xmlChar *attname;
- xmlChar *attvalue;
- const xmlChar **atts = ctxt->atts;
- int nbatts = 0;
- int maxatts = ctxt->maxatts;
- int i;
- if (RAW != '<') return(NULL);
- NEXT1;
- name = xmlParseName(ctxt);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "xmlParseStartTag: invalid element name\n");
- return(NULL);
- }
- /*
- * Now parse the attributes, it ends up with the ending
- *
- * (S Attribute)* S?
- */
- SKIP_BLANKS;
- GROW;
- while (((RAW != '>') &&
- ((RAW != '/') || (NXT(1) != '>')) &&
- (IS_BYTE_CHAR(RAW))) && (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *q = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- attname = xmlParseAttribute(ctxt, &attvalue);
- if ((attname != NULL) && (attvalue != NULL)) {
- /*
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same
- * start-tag or empty-element tag.
- */
- for (i = 0; i < nbatts;i += 2) {
- if (xmlStrEqual(atts[i], attname)) {
- xmlErrAttributeDup(ctxt, NULL, attname);
- xmlFree(attvalue);
- goto failed;
- }
- }
- /*
- * Add the pair to atts
- */
- if (atts == NULL) {
- maxatts = 22; /* allow for 10 attrs by default */
- atts = (const xmlChar **)
- xmlMalloc(maxatts * sizeof(xmlChar *));
- if (atts == NULL) {
- xmlErrMemory(ctxt, NULL);
- if (attvalue != NULL)
- xmlFree(attvalue);
- goto failed;
- }
- ctxt->atts = atts;
- ctxt->maxatts = maxatts;
- } else if (nbatts + 4 > maxatts) {
- const xmlChar **n;
- maxatts *= 2;
- n = (const xmlChar **) xmlRealloc((void *) atts,
- maxatts * sizeof(const xmlChar *));
- if (n == NULL) {
- xmlErrMemory(ctxt, NULL);
- if (attvalue != NULL)
- xmlFree(attvalue);
- goto failed;
- }
- atts = n;
- ctxt->atts = atts;
- ctxt->maxatts = maxatts;
- }
- atts[nbatts++] = attname;
- atts[nbatts++] = attvalue;
- atts[nbatts] = NULL;
- atts[nbatts + 1] = NULL;
- } else {
- if (attvalue != NULL)
- xmlFree(attvalue);
- }
- failed:
- GROW
- if ((RAW == '>') || (((RAW == '/') && (NXT(1) == '>'))))
- break;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "attributes construct error\n");
- }
- if ((cons == ctxt->input->consumed) && (q == CUR_PTR) &&
- (attname == NULL) && (attvalue == NULL)) {
- xmlFatalErrMsg(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlParseStartTag: problem parsing attributes\n");
- break;
- }
- SHRINK;
- GROW;
- }
- /*
- * SAX: Start of Element !
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->startElement != NULL) &&
- (!ctxt->disableSAX)) {
- if (nbatts > 0)
- ctxt->sax->startElement(ctxt->userData, name, atts);
- else
- ctxt->sax->startElement(ctxt->userData, name, NULL);
- }
- if (atts != NULL) {
- /* Free only the content strings */
- for (i = 1;i < nbatts;i+=2)
- if (atts[i] != NULL)
- xmlFree((xmlChar *) atts[i]);
- }
- return(name);
- }
- /**
- * xmlParseEndTag1:
- * @ctxt: an XML parser context
- * @line: line of the start tag
- * @nsNr: number of namespaces on the start tag
- *
- * parse an end of tag
- *
- * [42] ETag ::= '</' Name S? '>'
- *
- * With namespace
- *
- * [NS 9] ETag ::= '</' QName S? '>'
- */
- static void
- xmlParseEndTag1(xmlParserCtxtPtr ctxt, int line) {
- const xmlChar *name;
- GROW;
- if ((RAW != '<') || (NXT(1) != '/')) {
- xmlFatalErrMsg(ctxt, XML_ERR_LTSLASH_REQUIRED,
- "xmlParseEndTag: '</' not found\n");
- return;
- }
- SKIP(2);
- name = xmlParseNameAndCompare(ctxt,ctxt->name);
- /*
- * We should definitely be at the ending "S? '>'" part
- */
- GROW;
- SKIP_BLANKS;
- if ((!IS_BYTE_CHAR(RAW)) || (RAW != '>')) {
- xmlFatalErr(ctxt, XML_ERR_GT_REQUIRED, NULL);
- } else
- NEXT1;
- /*
- * [ WFC: Element Type Match ]
- * The Name in an element's end-tag must match the element type in the
- * start-tag.
- *
- */
- if (name != (xmlChar*)1) {
- if (name == NULL) name = BAD_CAST "unparsable";
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NAME_MISMATCH,
- "Opening and ending tag mismatch: %s line %d and %s\n",
- ctxt->name, line, name);
- }
- /*
- * SAX: End of Tag
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->endElement != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElement(ctxt->userData, ctxt->name);
- namePop(ctxt);
- spacePop(ctxt);
- return;
- }
- /**
- * xmlParseEndTag:
- * @ctxt: an XML parser context
- *
- * parse an end of tag
- *
- * [42] ETag ::= '</' Name S? '>'
- *
- * With namespace
- *
- * [NS 9] ETag ::= '</' QName S? '>'
- */
- void
- xmlParseEndTag(xmlParserCtxtPtr ctxt) {
- xmlParseEndTag1(ctxt, 0);
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /************************************************************************
- * *
- * SAX 2 specific operations *
- * *
- ************************************************************************/
- /*
- * xmlGetNamespace:
- * @ctxt: an XML parser context
- * @prefix: the prefix to lookup
- *
- * Lookup the namespace name for the @prefix (which ca be NULL)
- * The prefix must come from the @ctxt->dict dictionary
- *
- * Returns the namespace name or NULL if not bound
- */
- static const xmlChar *
- xmlGetNamespace(xmlParserCtxtPtr ctxt, const xmlChar *prefix) {
- int i;
- if (prefix == ctxt->str_xml) return(ctxt->str_xml_ns);
- for (i = ctxt->nsNr - 2;i >= 0;i-=2)
- if (ctxt->nsTab[i] == prefix) {
- if ((prefix == NULL) && (*ctxt->nsTab[i + 1] == 0))
- return(NULL);
- return(ctxt->nsTab[i + 1]);
- }
- return(NULL);
- }
- /**
- * xmlParseQName:
- * @ctxt: an XML parser context
- * @prefix: pointer to store the prefix part
- *
- * parse an XML Namespace QName
- *
- * [6] QName ::= (Prefix ':')? LocalPart
- * [7] Prefix ::= NCName
- * [8] LocalPart ::= NCName
- *
- * Returns the Name parsed or NULL
- */
- static const xmlChar *
- xmlParseQName(xmlParserCtxtPtr ctxt, const xmlChar **prefix) {
- const xmlChar *l, *p;
- GROW;
- l = xmlParseNCName(ctxt);
- if (l == NULL) {
- if (CUR == ':') {
- l = xmlParseName(ctxt);
- if (l != NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_QNAME,
- "Failed to parse QName '%s'\n", l, NULL, NULL);
- *prefix = NULL;
- return(l);
- }
- }
- return(NULL);
- }
- if (CUR == ':') {
- NEXT;
- p = l;
- l = xmlParseNCName(ctxt);
- if (l == NULL) {
- xmlChar *tmp;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- xmlNsErr(ctxt, XML_NS_ERR_QNAME,
- "Failed to parse QName '%s:'\n", p, NULL, NULL);
- l = xmlParseNmtoken(ctxt);
- if (l == NULL) {
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- tmp = xmlBuildQName(BAD_CAST "", p, NULL, 0);
- } else {
- tmp = xmlBuildQName(l, p, NULL, 0);
- xmlFree((char *)l);
- }
- p = xmlDictLookup(ctxt->dict, tmp, -1);
- if (tmp != NULL) xmlFree(tmp);
- *prefix = NULL;
- return(p);
- }
- if (CUR == ':') {
- xmlChar *tmp;
- xmlNsErr(ctxt, XML_NS_ERR_QNAME,
- "Failed to parse QName '%s:%s:'\n", p, l, NULL);
- NEXT;
- tmp = (xmlChar *) xmlParseName(ctxt);
- if (tmp != NULL) {
- tmp = xmlBuildQName(tmp, l, NULL, 0);
- l = xmlDictLookup(ctxt->dict, tmp, -1);
- if (tmp != NULL) xmlFree(tmp);
- *prefix = p;
- return(l);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- tmp = xmlBuildQName(BAD_CAST "", l, NULL, 0);
- l = xmlDictLookup(ctxt->dict, tmp, -1);
- if (tmp != NULL) xmlFree(tmp);
- *prefix = p;
- return(l);
- }
- *prefix = p;
- } else
- *prefix = NULL;
- return(l);
- }
- /**
- * xmlParseQNameAndCompare:
- * @ctxt: an XML parser context
- * @name: the localname
- * @prefix: the prefix, if any.
- *
- * parse an XML name and compares for match
- * (specialized for endtag parsing)
- *
- * Returns NULL for an illegal name, (xmlChar*) 1 for success
- * and the name for mismatch
- */
- static const xmlChar *
- xmlParseQNameAndCompare(xmlParserCtxtPtr ctxt, xmlChar const *name,
- xmlChar const *prefix) {
- const xmlChar *cmp;
- const xmlChar *in;
- const xmlChar *ret;
- const xmlChar *prefix2;
- if (prefix == NULL) return(xmlParseNameAndCompare(ctxt, name));
- GROW;
- in = ctxt->input->cur;
- cmp = prefix;
- while (*in != 0 && *in == *cmp) {
- ++in;
- ++cmp;
- }
- if ((*cmp == 0) && (*in == ':')) {
- in++;
- cmp = name;
- while (*in != 0 && *in == *cmp) {
- ++in;
- ++cmp;
- }
- if (*cmp == 0 && (*in == '>' || IS_BLANK_CH (*in))) {
- /* success */
- ctxt->input->col += in - ctxt->input->cur;
- ctxt->input->cur = in;
- return((const xmlChar*) 1);
- }
- }
- /*
- * all strings coms from the dictionary, equality can be done directly
- */
- ret = xmlParseQName (ctxt, &prefix2);
- if ((ret == name) && (prefix == prefix2))
- return((const xmlChar*) 1);
- return ret;
- }
- /**
- * xmlParseAttValueInternal:
- * @ctxt: an XML parser context
- * @len: attribute len result
- * @alloc: whether the attribute was reallocated as a new string
- * @normalize: if 1 then further non-CDATA normalization must be done
- *
- * parse a value for an attribute.
- * NOTE: if no normalization is needed, the routine will return pointers
- * directly from the data buffer.
- *
- * 3.3.3 Attribute-Value Normalization:
- * Before the value of an attribute is passed to the application or
- * checked for validity, the XML processor must normalize it as follows:
- * - a character reference is processed by appending the referenced
- * character to the attribute value
- * - an entity reference is processed by recursively processing the
- * replacement text of the entity
- * - a whitespace character (#x20, #xD, #xA, #x9) is processed by
- * appending #x20 to the normalized value, except that only a single
- * #x20 is appended for a "#xD#xA" sequence that is part of an external
- * parsed entity or the literal entity value of an internal parsed entity
- * - other characters are processed by appending them to the normalized value
- * If the declared value is not CDATA, then the XML processor must further
- * process the normalized attribute value by discarding any leading and
- * trailing space (#x20) characters, and by replacing sequences of space
- * (#x20) characters by a single space (#x20) character.
- * All attributes for which no declaration has been read should be treated
- * by a non-validating parser as if declared CDATA.
- *
- * Returns the AttValue parsed or NULL. The value has to be freed by the
- * caller if it was copied, this can be detected by val[*len] == 0.
- */
- #define GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end) \
- const xmlChar *oldbase = ctxt->input->base;\
- GROW;\
- if (ctxt->instate == XML_PARSER_EOF)\
- return(NULL);\
- if (oldbase != ctxt->input->base) {\
- ptrdiff_t delta = ctxt->input->base - oldbase;\
- start = start + delta;\
- in = in + delta;\
- }\
- end = ctxt->input->end;
- static xmlChar *
- xmlParseAttValueInternal(xmlParserCtxtPtr ctxt, int *len, int *alloc,
- int normalize)
- {
- xmlChar limit = 0;
- const xmlChar *in = NULL, *start, *end, *last;
- xmlChar *ret = NULL;
- int line, col;
- GROW;
- in = (xmlChar *) CUR_PTR;
- line = ctxt->input->line;
- col = ctxt->input->col;
- if (*in != '"' && *in != '\'') {
- xmlFatalErr(ctxt, XML_ERR_ATTRIBUTE_NOT_STARTED, NULL);
- return (NULL);
- }
- ctxt->instate = XML_PARSER_ATTRIBUTE_VALUE;
- /*
- * try to handle in this routine the most common case where no
- * allocation of a new string is required and where content is
- * pure ASCII.
- */
- limit = *in++;
- col++;
- end = ctxt->input->end;
- start = in;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- }
- if (normalize) {
- /*
- * Skip any leading spaces
- */
- while ((in < end) && (*in != limit) &&
- ((*in == 0x20) || (*in == 0x9) ||
- (*in == 0xA) || (*in == 0xD))) {
- if (*in == 0xA) {
- line++; col = 1;
- } else {
- col++;
- }
- in++;
- start = in;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- while ((in < end) && (*in != limit) && (*in >= 0x20) &&
- (*in <= 0x7f) && (*in != '&') && (*in != '<')) {
- col++;
- if ((*in++ == 0x20) && (*in == 0x20)) break;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- last = in;
- /*
- * skip the trailing blanks
- */
- while ((last[-1] == 0x20) && (last > start)) last--;
- while ((in < end) && (*in != limit) &&
- ((*in == 0x20) || (*in == 0x9) ||
- (*in == 0xA) || (*in == 0xD))) {
- if (*in == 0xA) {
- line++, col = 1;
- } else {
- col++;
- }
- in++;
- if (in >= end) {
- const xmlChar *oldbase = ctxt->input->base;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return(NULL);
- if (oldbase != ctxt->input->base) {
- ptrdiff_t delta = ctxt->input->base - oldbase;
- start = start + delta;
- in = in + delta;
- last = last + delta;
- }
- end = ctxt->input->end;
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- if (*in != limit) goto need_complex;
- } else {
- while ((in < end) && (*in != limit) && (*in >= 0x20) &&
- (*in <= 0x7f) && (*in != '&') && (*in != '<')) {
- in++;
- col++;
- if (in >= end) {
- GROW_PARSE_ATT_VALUE_INTERNAL(ctxt, in, start, end)
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- }
- }
- last = in;
- if (((in - start) > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsg(ctxt, XML_ERR_ATTRIBUTE_NOT_FINISHED,
- "AttValue length too long\n");
- return(NULL);
- }
- if (*in != limit) goto need_complex;
- }
- in++;
- col++;
- if (len != NULL) {
- *len = last - start;
- ret = (xmlChar *) start;
- } else {
- if (alloc) *alloc = 1;
- ret = xmlStrndup(start, last - start);
- }
- CUR_PTR = in;
- ctxt->input->line = line;
- ctxt->input->col = col;
- if (alloc) *alloc = 0;
- return ret;
- need_complex:
- if (alloc) *alloc = 1;
- return xmlParseAttValueComplex(ctxt, len, normalize);
- }
- /**
- * xmlParseAttribute2:
- * @ctxt: an XML parser context
- * @pref: the element prefix
- * @elem: the element name
- * @prefix: a xmlChar ** used to store the value of the attribute prefix
- * @value: a xmlChar ** used to store the value of the attribute
- * @len: an int * to save the length of the attribute
- * @alloc: an int * to indicate if the attribute was allocated
- *
- * parse an attribute in the new SAX2 framework.
- *
- * Returns the attribute name, and the value in *value, .
- */
- static const xmlChar *
- xmlParseAttribute2(xmlParserCtxtPtr ctxt,
- const xmlChar * pref, const xmlChar * elem,
- const xmlChar ** prefix, xmlChar ** value,
- int *len, int *alloc)
- {
- const xmlChar *name;
- xmlChar *val, *internal_val = NULL;
- int normalize = 0;
- *value = NULL;
- GROW;
- name = xmlParseQName(ctxt, prefix);
- if (name == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "error parsing attribute name\n");
- return (NULL);
- }
- /*
- * get the type if needed
- */
- if (ctxt->attsSpecial != NULL) {
- int type;
- type = (int) (ptrdiff_t) xmlHashQLookup2(ctxt->attsSpecial,
- pref, elem, *prefix, name);
- if (type != 0)
- normalize = 1;
- }
- /*
- * read the value
- */
- SKIP_BLANKS;
- if (RAW == '=') {
- NEXT;
- SKIP_BLANKS;
- val = xmlParseAttValueInternal(ctxt, len, alloc, normalize);
- if (normalize) {
- /*
- * Sometimes a second normalisation pass for spaces is needed
- * but that only happens if charrefs or entities references
- * have been used in the attribute value, i.e. the attribute
- * value have been extracted in an allocated string already.
- */
- if (*alloc) {
- const xmlChar *val2;
- val2 = xmlAttrNormalizeSpace2(ctxt, val, len);
- if ((val2 != NULL) && (val2 != val)) {
- xmlFree(val);
- val = (xmlChar *) val2;
- }
- }
- }
- ctxt->instate = XML_PARSER_CONTENT;
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_ATTRIBUTE_WITHOUT_VALUE,
- "Specification mandates value for attribute %s\n",
- name);
- return (NULL);
- }
- if (*prefix == ctxt->str_xml) {
- /*
- * Check that xml:lang conforms to the specification
- * No more registered as an error, just generate a warning now
- * since this was deprecated in XML second edition
- */
- if ((ctxt->pedantic) && (xmlStrEqual(name, BAD_CAST "lang"))) {
- internal_val = xmlStrndup(val, *len);
- if (!xmlCheckLanguageID(internal_val)) {
- xmlWarningMsg(ctxt, XML_WAR_LANG_VALUE,
- "Malformed value for xml:lang : %s\n",
- internal_val, NULL);
- }
- }
- /*
- * Check that xml:space conforms to the specification
- */
- if (xmlStrEqual(name, BAD_CAST "space")) {
- internal_val = xmlStrndup(val, *len);
- if (xmlStrEqual(internal_val, BAD_CAST "default"))
- *(ctxt->space) = 0;
- else if (xmlStrEqual(internal_val, BAD_CAST "preserve"))
- *(ctxt->space) = 1;
- else {
- xmlWarningMsg(ctxt, XML_WAR_SPACE_VALUE,
- "Invalid value \"%s\" for xml:space : \"default\" or \"preserve\" expected\n",
- internal_val, NULL);
- }
- }
- if (internal_val) {
- xmlFree(internal_val);
- }
- }
- *value = val;
- return (name);
- }
- /**
- * xmlParseStartTag2:
- * @ctxt: an XML parser context
- *
- * parse a start of tag either for rule element or
- * EmptyElement. In both case we don't parse the tag closing chars.
- * This routine is called when running SAX2 parsing
- *
- * [40] STag ::= '<' Name (S Attribute)* S? '>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * [44] EmptyElemTag ::= '<' Name (S Attribute)* S? '/>'
- *
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same start-tag or
- * empty-element tag.
- *
- * With namespace:
- *
- * [NS 8] STag ::= '<' QName (S Attribute)* S? '>'
- *
- * [NS 10] EmptyElement ::= '<' QName (S Attribute)* S? '/>'
- *
- * Returns the element name parsed
- */
- static const xmlChar *
- xmlParseStartTag2(xmlParserCtxtPtr ctxt, const xmlChar **pref,
- const xmlChar **URI, int *tlen) {
- const xmlChar *localname;
- const xmlChar *prefix;
- const xmlChar *attname;
- const xmlChar *aprefix;
- const xmlChar *nsname;
- xmlChar *attvalue;
- const xmlChar **atts = ctxt->atts;
- int maxatts = ctxt->maxatts;
- int nratts, nbatts, nbdef, inputid;
- int i, j, nbNs, attval;
- unsigned long cur;
- int nsNr = ctxt->nsNr;
- if (RAW != '<') return(NULL);
- NEXT1;
- /*
- * NOTE: it is crucial with the SAX2 API to never call SHRINK beyond that
- * point since the attribute values may be stored as pointers to
- * the buffer and calling SHRINK would destroy them !
- * The Shrinking is only possible once the full set of attribute
- * callbacks have been done.
- */
- SHRINK;
- cur = ctxt->input->cur - ctxt->input->base;
- inputid = ctxt->input->id;
- nbatts = 0;
- nratts = 0;
- nbdef = 0;
- nbNs = 0;
- attval = 0;
- /* Forget any namespaces added during an earlier parse of this element. */
- ctxt->nsNr = nsNr;
- localname = xmlParseQName(ctxt, &prefix);
- if (localname == NULL) {
- xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
- "StartTag: invalid element name\n");
- return(NULL);
- }
- *tlen = ctxt->input->cur - ctxt->input->base - cur;
- /*
- * Now parse the attributes, it ends up with the ending
- *
- * (S Attribute)* S?
- */
- SKIP_BLANKS;
- GROW;
- while (((RAW != '>') &&
- ((RAW != '/') || (NXT(1) != '>')) &&
- (IS_BYTE_CHAR(RAW))) && (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *q = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- int len = -1, alloc = 0;
- attname = xmlParseAttribute2(ctxt, prefix, localname,
- &aprefix, &attvalue, &len, &alloc);
- if ((attname == NULL) || (attvalue == NULL))
- goto next_attr;
- if (len < 0) len = xmlStrlen(attvalue);
- if ((attname == ctxt->str_xmlns) && (aprefix == NULL)) {
- const xmlChar *URL = xmlDictLookup(ctxt->dict, attvalue, len);
- xmlURIPtr uri;
- if (URL == NULL) {
- xmlErrMemory(ctxt, "dictionary allocation failure");
- if ((attvalue != NULL) && (alloc != 0))
- xmlFree(attvalue);
- localname = NULL;
- goto done;
- }
- if (*URL != 0) {
- uri = xmlParseURI((const char *) URL);
- if (uri == NULL) {
- xmlNsErr(ctxt, XML_WAR_NS_URI,
- "xmlns: '%s' is not a valid URI\n",
- URL, NULL, NULL);
- } else {
- if (uri->scheme == NULL) {
- xmlNsWarn(ctxt, XML_WAR_NS_URI_RELATIVE,
- "xmlns: URI %s is not absolute\n",
- URL, NULL, NULL);
- }
- xmlFreeURI(uri);
- }
- if (URL == ctxt->str_xml_ns) {
- if (attname != ctxt->str_xml) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xml namespace URI cannot be the default namespace\n",
- NULL, NULL, NULL);
- }
- goto next_attr;
- }
- if ((len == 29) &&
- (xmlStrEqual(URL,
- BAD_CAST "http://www.w3.org/2000/xmlns/"))) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "reuse of the xmlns namespace name is forbidden\n",
- NULL, NULL, NULL);
- goto next_attr;
- }
- }
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == NULL)
- break;
- if (j <= nbNs)
- xmlErrAttributeDup(ctxt, NULL, attname);
- else
- if (nsPush(ctxt, NULL, URL) > 0) nbNs++;
- } else if (aprefix == ctxt->str_xmlns) {
- const xmlChar *URL = xmlDictLookup(ctxt->dict, attvalue, len);
- xmlURIPtr uri;
- if (attname == ctxt->str_xml) {
- if (URL != ctxt->str_xml_ns) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xml namespace prefix mapped to wrong URI\n",
- NULL, NULL, NULL);
- }
- /*
- * Do not keep a namespace definition node
- */
- goto next_attr;
- }
- if (URL == ctxt->str_xml_ns) {
- if (attname != ctxt->str_xml) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xml namespace URI mapped to wrong prefix\n",
- NULL, NULL, NULL);
- }
- goto next_attr;
- }
- if (attname == ctxt->str_xmlns) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "redefinition of the xmlns prefix is forbidden\n",
- NULL, NULL, NULL);
- goto next_attr;
- }
- if ((len == 29) &&
- (xmlStrEqual(URL,
- BAD_CAST "http://www.w3.org/2000/xmlns/"))) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "reuse of the xmlns namespace name is forbidden\n",
- NULL, NULL, NULL);
- goto next_attr;
- }
- if ((URL == NULL) || (URL[0] == 0)) {
- xmlNsErr(ctxt, XML_NS_ERR_XML_NAMESPACE,
- "xmlns:%s: Empty XML namespace is not allowed\n",
- attname, NULL, NULL);
- goto next_attr;
- } else {
- uri = xmlParseURI((const char *) URL);
- if (uri == NULL) {
- xmlNsErr(ctxt, XML_WAR_NS_URI,
- "xmlns:%s: '%s' is not a valid URI\n",
- attname, URL, NULL);
- } else {
- if ((ctxt->pedantic) && (uri->scheme == NULL)) {
- xmlNsWarn(ctxt, XML_WAR_NS_URI_RELATIVE,
- "xmlns:%s: URI %s is not absolute\n",
- attname, URL, NULL);
- }
- xmlFreeURI(uri);
- }
- }
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == attname)
- break;
- if (j <= nbNs)
- xmlErrAttributeDup(ctxt, aprefix, attname);
- else
- if (nsPush(ctxt, attname, URL) > 0) nbNs++;
- } else {
- /*
- * Add the pair to atts
- */
- if ((atts == NULL) || (nbatts + 5 > maxatts)) {
- if (xmlCtxtGrowAttrs(ctxt, nbatts + 5) < 0) {
- goto next_attr;
- }
- maxatts = ctxt->maxatts;
- atts = ctxt->atts;
- }
- ctxt->attallocs[nratts++] = alloc;
- atts[nbatts++] = attname;
- atts[nbatts++] = aprefix;
- /*
- * The namespace URI field is used temporarily to point at the
- * base of the current input buffer for non-alloced attributes.
- * When the input buffer is reallocated, all the pointers become
- * invalid, but they can be reconstructed later.
- */
- if (alloc)
- atts[nbatts++] = NULL;
- else
- atts[nbatts++] = ctxt->input->base;
- atts[nbatts++] = attvalue;
- attvalue += len;
- atts[nbatts++] = attvalue;
- /*
- * tag if some deallocation is needed
- */
- if (alloc != 0) attval = 1;
- attvalue = NULL; /* moved into atts */
- }
- next_attr:
- if ((attvalue != NULL) && (alloc != 0)) {
- xmlFree(attvalue);
- attvalue = NULL;
- }
- GROW
- if (ctxt->instate == XML_PARSER_EOF)
- break;
- if ((RAW == '>') || (((RAW == '/') && (NXT(1) == '>'))))
- break;
- if (SKIP_BLANKS == 0) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "attributes construct error\n");
- break;
- }
- if ((cons == ctxt->input->consumed) && (q == CUR_PTR) &&
- (attname == NULL) && (attvalue == NULL)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "xmlParseStartTag: problem parsing attributes\n");
- break;
- }
- GROW;
- }
- if (ctxt->input->id != inputid) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "Unexpected change of input\n");
- localname = NULL;
- goto done;
- }
- /* Reconstruct attribute value pointers. */
- for (i = 0, j = 0; j < nratts; i += 5, j++) {
- if (atts[i+2] != NULL) {
- /*
- * Arithmetic on dangling pointers is technically undefined
- * behavior, but well...
- */
- ptrdiff_t offset = ctxt->input->base - atts[i+2];
- atts[i+2] = NULL; /* Reset repurposed namespace URI */
- atts[i+3] += offset; /* value */
- atts[i+4] += offset; /* valuend */
- }
- }
- /*
- * The attributes defaulting
- */
- if (ctxt->attsDefault != NULL) {
- xmlDefAttrsPtr defaults;
- defaults = xmlHashLookup2(ctxt->attsDefault, localname, prefix);
- if (defaults != NULL) {
- for (i = 0;i < defaults->nbAttrs;i++) {
- attname = defaults->values[5 * i];
- aprefix = defaults->values[5 * i + 1];
- /*
- * special work for namespaces defaulted defs
- */
- if ((attname == ctxt->str_xmlns) && (aprefix == NULL)) {
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == NULL)
- break;
- if (j <= nbNs) continue;
- nsname = xmlGetNamespace(ctxt, NULL);
- if (nsname != defaults->values[5 * i + 2]) {
- if (nsPush(ctxt, NULL,
- defaults->values[5 * i + 2]) > 0)
- nbNs++;
- }
- } else if (aprefix == ctxt->str_xmlns) {
- /*
- * check that it's not a defined namespace
- */
- for (j = 1;j <= nbNs;j++)
- if (ctxt->nsTab[ctxt->nsNr - 2 * j] == attname)
- break;
- if (j <= nbNs) continue;
- nsname = xmlGetNamespace(ctxt, attname);
- if (nsname != defaults->values[2]) {
- if (nsPush(ctxt, attname,
- defaults->values[5 * i + 2]) > 0)
- nbNs++;
- }
- } else {
- /*
- * check that it's not a defined attribute
- */
- for (j = 0;j < nbatts;j+=5) {
- if ((attname == atts[j]) && (aprefix == atts[j+1]))
- break;
- }
- if (j < nbatts) continue;
- if ((atts == NULL) || (nbatts + 5 > maxatts)) {
- if (xmlCtxtGrowAttrs(ctxt, nbatts + 5) < 0) {
- localname = NULL;
- goto done;
- }
- maxatts = ctxt->maxatts;
- atts = ctxt->atts;
- }
- atts[nbatts++] = attname;
- atts[nbatts++] = aprefix;
- if (aprefix == NULL)
- atts[nbatts++] = NULL;
- else
- atts[nbatts++] = xmlGetNamespace(ctxt, aprefix);
- atts[nbatts++] = defaults->values[5 * i + 2];
- atts[nbatts++] = defaults->values[5 * i + 3];
- if ((ctxt->standalone == 1) &&
- (defaults->values[5 * i + 4] != NULL)) {
- xmlValidityError(ctxt, XML_DTD_STANDALONE_DEFAULTED,
- "standalone: attribute %s on %s defaulted from external subset\n",
- attname, localname);
- }
- nbdef++;
- }
- }
- }
- }
- /*
- * The attributes checkings
- */
- for (i = 0; i < nbatts;i += 5) {
- /*
- * The default namespace does not apply to attribute names.
- */
- if (atts[i + 1] != NULL) {
- nsname = xmlGetNamespace(ctxt, atts[i + 1]);
- if (nsname == NULL) {
- xmlNsErr(ctxt, XML_NS_ERR_UNDEFINED_NAMESPACE,
- "Namespace prefix %s for %s on %s is not defined\n",
- atts[i + 1], atts[i], localname);
- }
- atts[i + 2] = nsname;
- } else
- nsname = NULL;
- /*
- * [ WFC: Unique Att Spec ]
- * No attribute name may appear more than once in the same
- * start-tag or empty-element tag.
- * As extended by the Namespace in XML REC.
- */
- for (j = 0; j < i;j += 5) {
- if (atts[i] == atts[j]) {
- if (atts[i+1] == atts[j+1]) {
- xmlErrAttributeDup(ctxt, atts[i+1], atts[i]);
- break;
- }
- if ((nsname != NULL) && (atts[j + 2] == nsname)) {
- xmlNsErr(ctxt, XML_NS_ERR_ATTRIBUTE_REDEFINED,
- "Namespaced Attribute %s in '%s' redefined\n",
- atts[i], nsname, NULL);
- break;
- }
- }
- }
- }
- nsname = xmlGetNamespace(ctxt, prefix);
- if ((prefix != NULL) && (nsname == NULL)) {
- xmlNsErr(ctxt, XML_NS_ERR_UNDEFINED_NAMESPACE,
- "Namespace prefix %s on %s is not defined\n",
- prefix, localname, NULL);
- }
- *pref = prefix;
- *URI = nsname;
- /*
- * SAX: Start of Element !
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->startElementNs != NULL) &&
- (!ctxt->disableSAX)) {
- if (nbNs > 0)
- ctxt->sax->startElementNs(ctxt->userData, localname, prefix,
- nsname, nbNs, &ctxt->nsTab[ctxt->nsNr - 2 * nbNs],
- nbatts / 5, nbdef, atts);
- else
- ctxt->sax->startElementNs(ctxt->userData, localname, prefix,
- nsname, 0, NULL, nbatts / 5, nbdef, atts);
- }
- done:
- /*
- * Free up attribute allocated strings if needed
- */
- if (attval != 0) {
- for (i = 3,j = 0; j < nratts;i += 5,j++)
- if ((ctxt->attallocs[j] != 0) && (atts[i] != NULL))
- xmlFree((xmlChar *) atts[i]);
- }
- return(localname);
- }
- /**
- * xmlParseEndTag2:
- * @ctxt: an XML parser context
- * @line: line of the start tag
- * @nsNr: number of namespaces on the start tag
- *
- * parse an end of tag
- *
- * [42] ETag ::= '</' Name S? '>'
- *
- * With namespace
- *
- * [NS 9] ETag ::= '</' QName S? '>'
- */
- static void
- xmlParseEndTag2(xmlParserCtxtPtr ctxt, const xmlStartTag *tag) {
- const xmlChar *name;
- GROW;
- if ((RAW != '<') || (NXT(1) != '/')) {
- xmlFatalErr(ctxt, XML_ERR_LTSLASH_REQUIRED, NULL);
- return;
- }
- SKIP(2);
- if (tag->prefix == NULL)
- name = xmlParseNameAndCompare(ctxt, ctxt->name);
- else
- name = xmlParseQNameAndCompare(ctxt, ctxt->name, tag->prefix);
- /*
- * We should definitely be at the ending "S? '>'" part
- */
- GROW;
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- SKIP_BLANKS;
- if ((!IS_BYTE_CHAR(RAW)) || (RAW != '>')) {
- xmlFatalErr(ctxt, XML_ERR_GT_REQUIRED, NULL);
- } else
- NEXT1;
- /*
- * [ WFC: Element Type Match ]
- * The Name in an element's end-tag must match the element type in the
- * start-tag.
- *
- */
- if (name != (xmlChar*)1) {
- if (name == NULL) name = BAD_CAST "unparsable";
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NAME_MISMATCH,
- "Opening and ending tag mismatch: %s line %d and %s\n",
- ctxt->name, tag->line, name);
- }
- /*
- * SAX: End of Tag
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->endElementNs != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElementNs(ctxt->userData, ctxt->name, tag->prefix,
- tag->URI);
- spacePop(ctxt);
- if (tag->nsNr != 0)
- nsPop(ctxt, tag->nsNr);
- }
- /**
- * xmlParseCDSect:
- * @ctxt: an XML parser context
- *
- * Parse escaped pure raw content.
- *
- * [18] CDSect ::= CDStart CData CDEnd
- *
- * [19] CDStart ::= '<![CDATA['
- *
- * [20] Data ::= (Char* - (Char* ']]>' Char*))
- *
- * [21] CDEnd ::= ']]>'
- */
- void
- xmlParseCDSect(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = XML_PARSER_BUFFER_SIZE;
- int r, rl;
- int s, sl;
- int cur, l;
- int count = 0;
- /* Check 2.6.0 was NXT(0) not RAW */
- if (CMP9(CUR_PTR, '<', '!', '[', 'C', 'D', 'A', 'T', 'A', '[')) {
- SKIP(9);
- } else
- return;
- ctxt->instate = XML_PARSER_CDATA_SECTION;
- r = CUR_CHAR(rl);
- if (!IS_CHAR(r)) {
- xmlFatalErr(ctxt, XML_ERR_CDATA_NOT_FINISHED, NULL);
- ctxt->instate = XML_PARSER_CONTENT;
- return;
- }
- NEXTL(rl);
- s = CUR_CHAR(sl);
- if (!IS_CHAR(s)) {
- xmlFatalErr(ctxt, XML_ERR_CDATA_NOT_FINISHED, NULL);
- ctxt->instate = XML_PARSER_CONTENT;
- return;
- }
- NEXTL(sl);
- cur = CUR_CHAR(l);
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return;
- }
- while (IS_CHAR(cur) &&
- ((r != ']') || (s != ']') || (cur != '>'))) {
- if (len + 5 >= size) {
- xmlChar *tmp;
- if ((size > XML_MAX_TEXT_LENGTH) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_CDATA_NOT_FINISHED,
- "CData section too big found", NULL);
- xmlFree (buf);
- return;
- }
- tmp = (xmlChar *) xmlRealloc(buf, size * 2 * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buf);
- xmlErrMemory(ctxt, NULL);
- return;
- }
- buf = tmp;
- size *= 2;
- }
- COPY_BUF(rl,buf,len,r);
- r = s;
- rl = sl;
- s = cur;
- sl = l;
- count++;
- if (count > 50) {
- SHRINK;
- GROW;
- if (ctxt->instate == XML_PARSER_EOF) {
- xmlFree(buf);
- return;
- }
- count = 0;
- }
- NEXTL(l);
- cur = CUR_CHAR(l);
- }
- buf[len] = 0;
- ctxt->instate = XML_PARSER_CONTENT;
- if (cur != '>') {
- xmlFatalErrMsgStr(ctxt, XML_ERR_CDATA_NOT_FINISHED,
- "CData section not finished\n%.50s\n", buf);
- xmlFree(buf);
- return;
- }
- NEXTL(l);
- /*
- * OK the buffer is to be consumed as cdata.
- */
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (ctxt->sax->cdataBlock != NULL)
- ctxt->sax->cdataBlock(ctxt->userData, buf, len);
- else if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData, buf, len);
- }
- xmlFree(buf);
- }
- /**
- * xmlParseContentInternal:
- * @ctxt: an XML parser context
- *
- * Parse a content sequence. Stops at EOF or '</'. Leaves checking of
- * unexpected EOF to the caller.
- */
- static void
- xmlParseContentInternal(xmlParserCtxtPtr ctxt) {
- int nameNr = ctxt->nameNr;
- GROW;
- while ((RAW != 0) &&
- (ctxt->instate != XML_PARSER_EOF)) {
- const xmlChar *test = CUR_PTR;
- unsigned int cons = ctxt->input->consumed;
- const xmlChar *cur = ctxt->input->cur;
- /*
- * First case : a Processing Instruction.
- */
- if ((*cur == '<') && (cur[1] == '?')) {
- xmlParsePI(ctxt);
- }
- /*
- * Second case : a CDSection
- */
- /* 2.6.0 test was *cur not RAW */
- else if (CMP9(CUR_PTR, '<', '!', '[', 'C', 'D', 'A', 'T', 'A', '[')) {
- xmlParseCDSect(ctxt);
- }
- /*
- * Third case : a comment
- */
- else if ((*cur == '<') && (NXT(1) == '!') &&
- (NXT(2) == '-') && (NXT(3) == '-')) {
- xmlParseComment(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- }
- /*
- * Fourth case : a sub-element.
- */
- else if (*cur == '<') {
- if (NXT(1) == '/') {
- if (ctxt->nameNr <= nameNr)
- break;
- xmlParseElementEnd(ctxt);
- } else {
- xmlParseElementStart(ctxt);
- }
- }
- /*
- * Fifth case : a reference. If if has not been resolved,
- * parsing returns it's Name, create the node
- */
- else if (*cur == '&') {
- xmlParseReference(ctxt);
- }
- /*
- * Last case, text. Note that References are handled directly.
- */
- else {
- xmlParseCharData(ctxt, 0);
- }
- GROW;
- SHRINK;
- if ((cons == ctxt->input->consumed) && (test == CUR_PTR)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "detected an error in element content\n");
- xmlHaltParser(ctxt);
- break;
- }
- }
- }
- /**
- * xmlParseContent:
- * @ctxt: an XML parser context
- *
- * Parse a content sequence. Stops at EOF or '</'.
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- */
- void
- xmlParseContent(xmlParserCtxtPtr ctxt) {
- int nameNr = ctxt->nameNr;
- xmlParseContentInternal(ctxt);
- if ((ctxt->instate != XML_PARSER_EOF) && (ctxt->nameNr > nameNr)) {
- const xmlChar *name = ctxt->nameTab[ctxt->nameNr - 1];
- int line = ctxt->pushTab[ctxt->nameNr - 1].line;
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NOT_FINISHED,
- "Premature end of data in tag %s line %d\n",
- name, line, NULL);
- }
- }
- /**
- * xmlParseElement:
- * @ctxt: an XML parser context
- *
- * parse an XML element
- *
- * [39] element ::= EmptyElemTag | STag content ETag
- *
- * [ WFC: Element Type Match ]
- * The Name in an element's end-tag must match the element type in the
- * start-tag.
- *
- */
- void
- xmlParseElement(xmlParserCtxtPtr ctxt) {
- if (xmlParseElementStart(ctxt) != 0)
- return;
- xmlParseContentInternal(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- return;
- if (CUR == 0) {
- const xmlChar *name = ctxt->nameTab[ctxt->nameNr - 1];
- int line = ctxt->pushTab[ctxt->nameNr - 1].line;
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_TAG_NOT_FINISHED,
- "Premature end of data in tag %s line %d\n",
- name, line, NULL);
- return;
- }
- xmlParseElementEnd(ctxt);
- }
- /**
- * xmlParseElementStart:
- * @ctxt: an XML parser context
- *
- * Parse the start of an XML element. Returns -1 in case of error, 0 if an
- * opening tag was parsed, 1 if an empty element was parsed.
- */
- static int
- xmlParseElementStart(xmlParserCtxtPtr ctxt) {
- const xmlChar *name;
- const xmlChar *prefix = NULL;
- const xmlChar *URI = NULL;
- xmlParserNodeInfo node_info;
- int line, tlen = 0;
- xmlNodePtr ret;
- int nsNr = ctxt->nsNr;
- if (((unsigned int) ctxt->nameNr > xmlParserMaxDepth) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErrMsgInt(ctxt, XML_ERR_INTERNAL_ERROR,
- "Excessive depth in document: %d use XML_PARSE_HUGE option\n",
- xmlParserMaxDepth);
- xmlHaltParser(ctxt);
- return(-1);
- }
- /* Capture start position */
- if (ctxt->record_info) {
- node_info.begin_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.begin_line = ctxt->input->line;
- }
- if (ctxt->spaceNr == 0)
- spacePush(ctxt, -1);
- else if (*ctxt->space == -2)
- spacePush(ctxt, -1);
- else
- spacePush(ctxt, *ctxt->space);
- line = ctxt->input->line;
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax2)
- #endif /* LIBXML_SAX1_ENABLED */
- name = xmlParseStartTag2(ctxt, &prefix, &URI, &tlen);
- #ifdef LIBXML_SAX1_ENABLED
- else
- name = xmlParseStartTag(ctxt);
- #endif /* LIBXML_SAX1_ENABLED */
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if (name == NULL) {
- spacePop(ctxt);
- return(-1);
- }
- nameNsPush(ctxt, name, prefix, URI, line, ctxt->nsNr - nsNr);
- ret = ctxt->node;
- #ifdef LIBXML_VALID_ENABLED
- /*
- * [ VC: Root Element Type ]
- * The Name in the document type declaration must match the element
- * type of the root element.
- */
- if (ctxt->validate && ctxt->wellFormed && ctxt->myDoc &&
- ctxt->node && (ctxt->node == ctxt->myDoc->children))
- ctxt->valid &= xmlValidateRoot(&ctxt->vctxt, ctxt->myDoc);
- #endif /* LIBXML_VALID_ENABLED */
- /*
- * Check for an Empty Element.
- */
- if ((RAW == '/') && (NXT(1) == '>')) {
- SKIP(2);
- if (ctxt->sax2) {
- if ((ctxt->sax != NULL) && (ctxt->sax->endElementNs != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElementNs(ctxt->userData, name, prefix, URI);
- #ifdef LIBXML_SAX1_ENABLED
- } else {
- if ((ctxt->sax != NULL) && (ctxt->sax->endElement != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElement(ctxt->userData, name);
- #endif /* LIBXML_SAX1_ENABLED */
- }
- namePop(ctxt);
- spacePop(ctxt);
- if (nsNr != ctxt->nsNr)
- nsPop(ctxt, ctxt->nsNr - nsNr);
- if ( ret != NULL && ctxt->record_info ) {
- node_info.end_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.end_line = ctxt->input->line;
- node_info.node = ret;
- xmlParserAddNodeInfo(ctxt, &node_info);
- }
- return(1);
- }
- if (RAW == '>') {
- NEXT1;
- } else {
- xmlFatalErrMsgStrIntStr(ctxt, XML_ERR_GT_REQUIRED,
- "Couldn't find end of Start Tag %s line %d\n",
- name, line, NULL);
- /*
- * end of parsing of this node.
- */
- nodePop(ctxt);
- namePop(ctxt);
- spacePop(ctxt);
- if (nsNr != ctxt->nsNr)
- nsPop(ctxt, ctxt->nsNr - nsNr);
- /*
- * Capture end position and add node
- */
- if ( ret != NULL && ctxt->record_info ) {
- node_info.end_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.end_line = ctxt->input->line;
- node_info.node = ret;
- xmlParserAddNodeInfo(ctxt, &node_info);
- }
- return(-1);
- }
- return(0);
- }
- /**
- * xmlParseElementEnd:
- * @ctxt: an XML parser context
- *
- * Parse the end of an XML element.
- */
- static void
- xmlParseElementEnd(xmlParserCtxtPtr ctxt) {
- xmlParserNodeInfo node_info;
- xmlNodePtr ret = ctxt->node;
- if (ctxt->nameNr <= 0)
- return;
- /*
- * parse the end of tag: '</' should be here.
- */
- if (ctxt->sax2) {
- xmlParseEndTag2(ctxt, &ctxt->pushTab[ctxt->nameNr - 1]);
- namePop(ctxt);
- }
- #ifdef LIBXML_SAX1_ENABLED
- else
- xmlParseEndTag1(ctxt, 0);
- #endif /* LIBXML_SAX1_ENABLED */
- /*
- * Capture end position and add node
- */
- if ( ret != NULL && ctxt->record_info ) {
- node_info.end_pos = ctxt->input->consumed +
- (CUR_PTR - ctxt->input->base);
- node_info.end_line = ctxt->input->line;
- node_info.node = ret;
- xmlParserAddNodeInfo(ctxt, &node_info);
- }
- }
- /**
- * xmlParseVersionNum:
- * @ctxt: an XML parser context
- *
- * parse the XML version value.
- *
- * [26] VersionNum ::= '1.' [0-9]+
- *
- * In practice allow [0-9].[0-9]+ at that level
- *
- * Returns the string giving the XML version number, or NULL
- */
- xmlChar *
- xmlParseVersionNum(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = 10;
- xmlChar cur;
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- cur = CUR;
- if (!((cur >= '0') && (cur <= '9'))) {
- xmlFree(buf);
- return(NULL);
- }
- buf[len++] = cur;
- NEXT;
- cur=CUR;
- if (cur != '.') {
- xmlFree(buf);
- return(NULL);
- }
- buf[len++] = cur;
- NEXT;
- cur=CUR;
- while ((cur >= '0') && (cur <= '9')) {
- if (len + 1 >= size) {
- xmlChar *tmp;
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlFree(buf);
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- buf = tmp;
- }
- buf[len++] = cur;
- NEXT;
- cur=CUR;
- }
- buf[len] = 0;
- return(buf);
- }
- /**
- * xmlParseVersionInfo:
- * @ctxt: an XML parser context
- *
- * parse the XML version.
- *
- * [24] VersionInfo ::= S 'version' Eq (' VersionNum ' | " VersionNum ")
- *
- * [25] Eq ::= S? '=' S?
- *
- * Returns the version string, e.g. "1.0"
- */
- xmlChar *
- xmlParseVersionInfo(xmlParserCtxtPtr ctxt) {
- xmlChar *version = NULL;
- if (CMP7(CUR_PTR, 'v', 'e', 'r', 's', 'i', 'o', 'n')) {
- SKIP(7);
- SKIP_BLANKS;
- if (RAW != '=') {
- xmlFatalErr(ctxt, XML_ERR_EQUAL_REQUIRED, NULL);
- return(NULL);
- }
- NEXT;
- SKIP_BLANKS;
- if (RAW == '"') {
- NEXT;
- version = xmlParseVersionNum(ctxt);
- if (RAW != '"') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else if (RAW == '\''){
- NEXT;
- version = xmlParseVersionNum(ctxt);
- if (RAW != '\'') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_STARTED, NULL);
- }
- }
- return(version);
- }
- /**
- * xmlParseEncName:
- * @ctxt: an XML parser context
- *
- * parse the XML encoding name
- *
- * [81] EncName ::= [A-Za-z] ([A-Za-z0-9._] | '-')*
- *
- * Returns the encoding name value or NULL
- */
- xmlChar *
- xmlParseEncName(xmlParserCtxtPtr ctxt) {
- xmlChar *buf = NULL;
- int len = 0;
- int size = 10;
- xmlChar cur;
- cur = CUR;
- if (((cur >= 'a') && (cur <= 'z')) ||
- ((cur >= 'A') && (cur <= 'Z'))) {
- buf = (xmlChar *) xmlMallocAtomic(size * sizeof(xmlChar));
- if (buf == NULL) {
- xmlErrMemory(ctxt, NULL);
- return(NULL);
- }
- buf[len++] = cur;
- NEXT;
- cur = CUR;
- while (((cur >= 'a') && (cur <= 'z')) ||
- ((cur >= 'A') && (cur <= 'Z')) ||
- ((cur >= '0') && (cur <= '9')) ||
- (cur == '.') || (cur == '_') ||
- (cur == '-')) {
- if (len + 1 >= size) {
- xmlChar *tmp;
- size *= 2;
- tmp = (xmlChar *) xmlRealloc(buf, size * sizeof(xmlChar));
- if (tmp == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFree(buf);
- return(NULL);
- }
- buf = tmp;
- }
- buf[len++] = cur;
- NEXT;
- cur = CUR;
- if (cur == 0) {
- SHRINK;
- GROW;
- cur = CUR;
- }
- }
- buf[len] = 0;
- } else {
- xmlFatalErr(ctxt, XML_ERR_ENCODING_NAME, NULL);
- }
- return(buf);
- }
- /**
- * xmlParseEncodingDecl:
- * @ctxt: an XML parser context
- *
- * parse the XML encoding declaration
- *
- * [80] EncodingDecl ::= S 'encoding' Eq ('"' EncName '"' | "'" EncName "'")
- *
- * this setups the conversion filters.
- *
- * Returns the encoding value or NULL
- */
- const xmlChar *
- xmlParseEncodingDecl(xmlParserCtxtPtr ctxt) {
- xmlChar *encoding = NULL;
- SKIP_BLANKS;
- if (CMP8(CUR_PTR, 'e', 'n', 'c', 'o', 'd', 'i', 'n', 'g')) {
- SKIP(8);
- SKIP_BLANKS;
- if (RAW != '=') {
- xmlFatalErr(ctxt, XML_ERR_EQUAL_REQUIRED, NULL);
- return(NULL);
- }
- NEXT;
- SKIP_BLANKS;
- if (RAW == '"') {
- NEXT;
- encoding = xmlParseEncName(ctxt);
- if (RAW != '"') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- xmlFree((xmlChar *) encoding);
- return(NULL);
- } else
- NEXT;
- } else if (RAW == '\''){
- NEXT;
- encoding = xmlParseEncName(ctxt);
- if (RAW != '\'') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- xmlFree((xmlChar *) encoding);
- return(NULL);
- } else
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_STARTED, NULL);
- }
- /*
- * Non standard parsing, allowing the user to ignore encoding
- */
- if (ctxt->options & XML_PARSE_IGNORE_ENC) {
- xmlFree((xmlChar *) encoding);
- return(NULL);
- }
- /*
- * UTF-16 encoding switch has already taken place at this stage,
- * more over the little-endian/big-endian selection is already done
- */
- if ((encoding != NULL) &&
- ((!xmlStrcasecmp(encoding, BAD_CAST "UTF-16")) ||
- (!xmlStrcasecmp(encoding, BAD_CAST "UTF16")))) {
- /*
- * If no encoding was passed to the parser, that we are
- * using UTF-16 and no decoder is present i.e. the
- * document is apparently UTF-8 compatible, then raise an
- * encoding mismatch fatal error
- */
- if ((ctxt->encoding == NULL) &&
- (ctxt->input->buf != NULL) &&
- (ctxt->input->buf->encoder == NULL)) {
- xmlFatalErrMsg(ctxt, XML_ERR_INVALID_ENCODING,
- "Document labelled UTF-16 but has UTF-8 content\n");
- }
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = encoding;
- }
- /*
- * UTF-8 encoding is handled natively
- */
- else if ((encoding != NULL) &&
- ((!xmlStrcasecmp(encoding, BAD_CAST "UTF-8")) ||
- (!xmlStrcasecmp(encoding, BAD_CAST "UTF8")))) {
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = encoding;
- }
- else if (encoding != NULL) {
- xmlCharEncodingHandlerPtr handler;
- if (ctxt->input->encoding != NULL)
- xmlFree((xmlChar *) ctxt->input->encoding);
- ctxt->input->encoding = encoding;
- handler = xmlFindCharEncodingHandler((const char *) encoding);
- if (handler != NULL) {
- if (xmlSwitchToEncoding(ctxt, handler) < 0) {
- /* failed to convert */
- ctxt->errNo = XML_ERR_UNSUPPORTED_ENCODING;
- return(NULL);
- }
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNSUPPORTED_ENCODING,
- "Unsupported encoding %s\n", encoding);
- return(NULL);
- }
- }
- }
- return(encoding);
- }
- /**
- * xmlParseSDDecl:
- * @ctxt: an XML parser context
- *
- * parse the XML standalone declaration
- *
- * [32] SDDecl ::= S 'standalone' Eq
- * (("'" ('yes' | 'no') "'") | ('"' ('yes' | 'no')'"'))
- *
- * [ VC: Standalone Document Declaration ]
- * TODO The standalone document declaration must have the value "no"
- * if any external markup declarations contain declarations of:
- * - attributes with default values, if elements to which these
- * attributes apply appear in the document without specifications
- * of values for these attributes, or
- * - entities (other than amp, lt, gt, apos, quot), if references
- * to those entities appear in the document, or
- * - attributes with values subject to normalization, where the
- * attribute appears in the document with a value which will change
- * as a result of normalization, or
- * - element types with element content, if white space occurs directly
- * within any instance of those types.
- *
- * Returns:
- * 1 if standalone="yes"
- * 0 if standalone="no"
- * -2 if standalone attribute is missing or invalid
- * (A standalone value of -2 means that the XML declaration was found,
- * but no value was specified for the standalone attribute).
- */
- int
- xmlParseSDDecl(xmlParserCtxtPtr ctxt) {
- int standalone = -2;
- SKIP_BLANKS;
- if (CMP10(CUR_PTR, 's', 't', 'a', 'n', 'd', 'a', 'l', 'o', 'n', 'e')) {
- SKIP(10);
- SKIP_BLANKS;
- if (RAW != '=') {
- xmlFatalErr(ctxt, XML_ERR_EQUAL_REQUIRED, NULL);
- return(standalone);
- }
- NEXT;
- SKIP_BLANKS;
- if (RAW == '\''){
- NEXT;
- if ((RAW == 'n') && (NXT(1) == 'o')) {
- standalone = 0;
- SKIP(2);
- } else if ((RAW == 'y') && (NXT(1) == 'e') &&
- (NXT(2) == 's')) {
- standalone = 1;
- SKIP(3);
- } else {
- xmlFatalErr(ctxt, XML_ERR_STANDALONE_VALUE, NULL);
- }
- if (RAW != '\'') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else if (RAW == '"'){
- NEXT;
- if ((RAW == 'n') && (NXT(1) == 'o')) {
- standalone = 0;
- SKIP(2);
- } else if ((RAW == 'y') && (NXT(1) == 'e') &&
- (NXT(2) == 's')) {
- standalone = 1;
- SKIP(3);
- } else {
- xmlFatalErr(ctxt, XML_ERR_STANDALONE_VALUE, NULL);
- }
- if (RAW != '"') {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_CLOSED, NULL);
- } else
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_STRING_NOT_STARTED, NULL);
- }
- }
- return(standalone);
- }
- /**
- * xmlParseXMLDecl:
- * @ctxt: an XML parser context
- *
- * parse an XML declaration header
- *
- * [23] XMLDecl ::= '<?xml' VersionInfo EncodingDecl? SDDecl? S? '?>'
- */
- void
- xmlParseXMLDecl(xmlParserCtxtPtr ctxt) {
- xmlChar *version;
- /*
- * This value for standalone indicates that the document has an
- * XML declaration but it does not have a standalone attribute.
- * It will be overwritten later if a standalone attribute is found.
- */
- ctxt->input->standalone = -2;
- /*
- * We know that '<?xml' is here.
- */
- SKIP(5);
- if (!IS_BLANK_CH(RAW)) {
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED,
- "Blank needed after '<?xml'\n");
- }
- SKIP_BLANKS;
- /*
- * We must have the VersionInfo here.
- */
- version = xmlParseVersionInfo(ctxt);
- if (version == NULL) {
- xmlFatalErr(ctxt, XML_ERR_VERSION_MISSING, NULL);
- } else {
- if (!xmlStrEqual(version, (const xmlChar *) XML_DEFAULT_VERSION)) {
- /*
- * Changed here for XML-1.0 5th edition
- */
- if (ctxt->options & XML_PARSE_OLD10) {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNKNOWN_VERSION,
- "Unsupported version '%s'\n",
- version);
- } else {
- if ((version[0] == '1') && ((version[1] == '.'))) {
- xmlWarningMsg(ctxt, XML_WAR_UNKNOWN_VERSION,
- "Unsupported version '%s'\n",
- version, NULL);
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNKNOWN_VERSION,
- "Unsupported version '%s'\n",
- version);
- }
- }
- }
- if (ctxt->version != NULL)
- xmlFree((void *) ctxt->version);
- ctxt->version = version;
- }
- /*
- * We may have the encoding declaration
- */
- if (!IS_BLANK_CH(RAW)) {
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- return;
- }
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED, "Blank needed here\n");
- }
- xmlParseEncodingDecl(ctxt);
- if ((ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) ||
- (ctxt->instate == XML_PARSER_EOF)) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- return;
- }
- /*
- * We may have the standalone status.
- */
- if ((ctxt->input->encoding != NULL) && (!IS_BLANK_CH(RAW))) {
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- return;
- }
- xmlFatalErrMsg(ctxt, XML_ERR_SPACE_REQUIRED, "Blank needed here\n");
- }
- /*
- * We can grow the input buffer freely at that point
- */
- GROW;
- SKIP_BLANKS;
- ctxt->input->standalone = xmlParseSDDecl(ctxt);
- SKIP_BLANKS;
- if ((RAW == '?') && (NXT(1) == '>')) {
- SKIP(2);
- } else if (RAW == '>') {
- /* Deprecated old WD ... */
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- NEXT;
- } else {
- xmlFatalErr(ctxt, XML_ERR_XMLDECL_NOT_FINISHED, NULL);
- MOVETO_ENDTAG(CUR_PTR);
- NEXT;
- }
- }
- /**
- * xmlParseMisc:
- * @ctxt: an XML parser context
- *
- * parse an XML Misc* optional field.
- *
- * [27] Misc ::= Comment | PI | S
- */
- void
- xmlParseMisc(xmlParserCtxtPtr ctxt) {
- while ((ctxt->instate != XML_PARSER_EOF) &&
- (((RAW == '<') && (NXT(1) == '?')) ||
- (CMP4(CUR_PTR, '<', '!', '-', '-')) ||
- IS_BLANK_CH(CUR))) {
- if ((RAW == '<') && (NXT(1) == '?')) {
- xmlParsePI(ctxt);
- } else if (IS_BLANK_CH(CUR)) {
- NEXT;
- } else
- xmlParseComment(ctxt);
- }
- }
- /**
- * xmlParseDocument:
- * @ctxt: an XML parser context
- *
- * parse an XML document (and build a tree if using the standard SAX
- * interface).
- *
- * [1] document ::= prolog element Misc*
- *
- * [22] prolog ::= XMLDecl? Misc* (doctypedecl Misc*)?
- *
- * Returns 0, -1 in case of error. the parser context is augmented
- * as a result of the parsing.
- */
- int
- xmlParseDocument(xmlParserCtxtPtr ctxt) {
- xmlChar start[4];
- xmlCharEncoding enc;
- xmlInitParser();
- if ((ctxt == NULL) || (ctxt->input == NULL))
- return(-1);
- GROW;
- /*
- * SAX: detecting the level.
- */
- xmlDetectSAX2(ctxt);
- /*
- * SAX: beginning of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData, &xmlDefaultSAXLocator);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if ((ctxt->encoding == NULL) &&
- ((ctxt->input->end - ctxt->input->cur) >= 4)) {
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(&start[0], 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- if (CUR == 0) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- return(-1);
- }
- /*
- * Check for the XMLDecl in the Prolog.
- * do not GROW here to avoid the detected encoder to decode more
- * than just the first line, unless the amount of data is really
- * too small to hold "<?xml version="1.0" encoding="foo"
- */
- if ((ctxt->input->end - ctxt->input->cur) < 35) {
- GROW;
- }
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- /*
- * Note that we will switch encoding on the fly.
- */
- xmlParseXMLDecl(ctxt);
- if ((ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) ||
- (ctxt->instate == XML_PARSER_EOF)) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- return(-1);
- }
- ctxt->standalone = ctxt->input->standalone;
- SKIP_BLANKS;
- } else {
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- }
- if ((ctxt->sax) && (ctxt->sax->startDocument) && (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if ((ctxt->myDoc != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL) && (ctxt->input->buf->compressed >= 0)) {
- ctxt->myDoc->compression = ctxt->input->buf->compressed;
- }
- /*
- * The Misc part of the Prolog
- */
- GROW;
- xmlParseMisc(ctxt);
- /*
- * Then possibly doc type declaration(s) and more Misc
- * (doctypedecl Misc*)?
- */
- GROW;
- if (CMP9(CUR_PTR, '<', '!', 'D', 'O', 'C', 'T', 'Y', 'P', 'E')) {
- ctxt->inSubset = 1;
- xmlParseDocTypeDecl(ctxt);
- if (RAW == '[') {
- ctxt->instate = XML_PARSER_DTD;
- xmlParseInternalSubset(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- }
- /*
- * Create and update the external subset.
- */
- ctxt->inSubset = 2;
- if ((ctxt->sax != NULL) && (ctxt->sax->externalSubset != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->externalSubset(ctxt->userData, ctxt->intSubName,
- ctxt->extSubSystem, ctxt->extSubURI);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- ctxt->inSubset = 0;
- xmlCleanSpecialAttr(ctxt);
- ctxt->instate = XML_PARSER_PROLOG;
- xmlParseMisc(ctxt);
- }
- /*
- * Time to start parsing the tree itself
- */
- GROW;
- if (RAW != '<') {
- xmlFatalErrMsg(ctxt, XML_ERR_DOCUMENT_EMPTY,
- "Start tag expected, '<' not found\n");
- } else {
- ctxt->instate = XML_PARSER_CONTENT;
- xmlParseElement(ctxt);
- ctxt->instate = XML_PARSER_EPILOG;
- /*
- * The Misc part at the end
- */
- xmlParseMisc(ctxt);
- if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- }
- ctxt->instate = XML_PARSER_EOF;
- }
- /*
- * SAX: end of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- /*
- * Remove locally kept entity definitions if the tree was not built
- */
- if ((ctxt->myDoc != NULL) &&
- (xmlStrEqual(ctxt->myDoc->version, SAX_COMPAT_MODE))) {
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if ((ctxt->wellFormed) && (ctxt->myDoc != NULL)) {
- ctxt->myDoc->properties |= XML_DOC_WELLFORMED;
- if (ctxt->valid)
- ctxt->myDoc->properties |= XML_DOC_DTDVALID;
- if (ctxt->nsWellFormed)
- ctxt->myDoc->properties |= XML_DOC_NSVALID;
- if (ctxt->options & XML_PARSE_OLD10)
- ctxt->myDoc->properties |= XML_DOC_OLD10;
- }
- if (! ctxt->wellFormed) {
- ctxt->valid = 0;
- return(-1);
- }
- return(0);
- }
- /**
- * xmlParseExtParsedEnt:
- * @ctxt: an XML parser context
- *
- * parse a general parsed entity
- * An external general parsed entity is well-formed if it matches the
- * production labeled extParsedEnt.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * Returns 0, -1 in case of error. the parser context is augmented
- * as a result of the parsing.
- */
- int
- xmlParseExtParsedEnt(xmlParserCtxtPtr ctxt) {
- xmlChar start[4];
- xmlCharEncoding enc;
- if ((ctxt == NULL) || (ctxt->input == NULL))
- return(-1);
- xmlDefaultSAXHandlerInit();
- xmlDetectSAX2(ctxt);
- GROW;
- /*
- * SAX: beginning of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData, &xmlDefaultSAXLocator);
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- if ((ctxt->input->end - ctxt->input->cur) >= 4) {
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- if (CUR == 0) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- }
- /*
- * Check for the XMLDecl in the Prolog.
- */
- GROW;
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- /*
- * Note that we will switch encoding on the fly.
- */
- xmlParseXMLDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right here
- */
- return(-1);
- }
- SKIP_BLANKS;
- } else {
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- }
- if ((ctxt->sax) && (ctxt->sax->startDocument) && (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- /*
- * Doing validity checking on chunk doesn't make sense
- */
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->validate = 0;
- ctxt->loadsubset = 0;
- ctxt->depth = 0;
- xmlParseContent(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- /*
- * SAX: end of the document processing.
- */
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- if (! ctxt->wellFormed) return(-1);
- return(0);
- }
- #ifdef LIBXML_PUSH_ENABLED
- /************************************************************************
- * *
- * Progressive parsing interfaces *
- * *
- ************************************************************************/
- /**
- * xmlParseLookupSequence:
- * @ctxt: an XML parser context
- * @first: the first char to lookup
- * @next: the next char to lookup or zero
- * @third: the next char to lookup or zero
- *
- * Try to find if a sequence (first, next, third) or just (first next) or
- * (first) is available in the input stream.
- * This function has a side effect of (possibly) incrementing ctxt->checkIndex
- * to avoid rescanning sequences of bytes, it DOES change the state of the
- * parser, do not use liberally.
- *
- * Returns the index to the current parsing point if the full sequence
- * is available, -1 otherwise.
- */
- static int
- xmlParseLookupSequence(xmlParserCtxtPtr ctxt, xmlChar first,
- xmlChar next, xmlChar third) {
- int base, len;
- xmlParserInputPtr in;
- const xmlChar *buf;
- in = ctxt->input;
- if (in == NULL) return(-1);
- base = in->cur - in->base;
- if (base < 0) return(-1);
- if (ctxt->checkIndex > base)
- base = ctxt->checkIndex;
- if (in->buf == NULL) {
- buf = in->base;
- len = in->length;
- } else {
- buf = xmlBufContent(in->buf->buffer);
- len = xmlBufUse(in->buf->buffer);
- }
- /* take into account the sequence length */
- if (third) len -= 2;
- else if (next) len --;
- for (;base < len;base++) {
- if (buf[base] == first) {
- if (third != 0) {
- if ((buf[base + 1] != next) ||
- (buf[base + 2] != third)) continue;
- } else if (next != 0) {
- if (buf[base + 1] != next) continue;
- }
- ctxt->checkIndex = 0;
- #ifdef DEBUG_PUSH
- if (next == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c' found at %d\n",
- first, base);
- else if (third == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c' found at %d\n",
- first, next, base);
- else
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c%c' found at %d\n",
- first, next, third, base);
- #endif
- return(base - (in->cur - in->base));
- }
- }
- ctxt->checkIndex = base;
- #ifdef DEBUG_PUSH
- if (next == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c' failed\n", first);
- else if (third == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c' failed\n", first, next);
- else
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup '%c%c%c' failed\n", first, next, third);
- #endif
- return(-1);
- }
- /**
- * xmlParseGetLasts:
- * @ctxt: an XML parser context
- * @lastlt: pointer to store the last '<' from the input
- * @lastgt: pointer to store the last '>' from the input
- *
- * Lookup the last < and > in the current chunk
- */
- static void
- xmlParseGetLasts(xmlParserCtxtPtr ctxt, const xmlChar **lastlt,
- const xmlChar **lastgt) {
- const xmlChar *tmp;
- if ((ctxt == NULL) || (lastlt == NULL) || (lastgt == NULL)) {
- xmlGenericError(xmlGenericErrorContext,
- "Internal error: xmlParseGetLasts\n");
- return;
- }
- if ((ctxt->progressive != 0) && (ctxt->inputNr == 1)) {
- tmp = ctxt->input->end;
- tmp--;
- while ((tmp >= ctxt->input->base) && (*tmp != '<')) tmp--;
- if (tmp < ctxt->input->base) {
- *lastlt = NULL;
- *lastgt = NULL;
- } else {
- *lastlt = tmp;
- tmp++;
- while ((tmp < ctxt->input->end) && (*tmp != '>')) {
- if (*tmp == '\'') {
- tmp++;
- while ((tmp < ctxt->input->end) && (*tmp != '\'')) tmp++;
- if (tmp < ctxt->input->end) tmp++;
- } else if (*tmp == '"') {
- tmp++;
- while ((tmp < ctxt->input->end) && (*tmp != '"')) tmp++;
- if (tmp < ctxt->input->end) tmp++;
- } else
- tmp++;
- }
- if (tmp < ctxt->input->end)
- *lastgt = tmp;
- else {
- tmp = *lastlt;
- tmp--;
- while ((tmp >= ctxt->input->base) && (*tmp != '>')) tmp--;
- if (tmp >= ctxt->input->base)
- *lastgt = tmp;
- else
- *lastgt = NULL;
- }
- }
- } else {
- *lastlt = NULL;
- *lastgt = NULL;
- }
- }
- /**
- * xmlCheckCdataPush:
- * @cur: pointer to the block of characters
- * @len: length of the block in bytes
- * @complete: 1 if complete CDATA block is passed in, 0 if partial block
- *
- * Check that the block of characters is okay as SCdata content [20]
- *
- * Returns the number of bytes to pass if okay, a negative index where an
- * UTF-8 error occurred otherwise
- */
- static int
- xmlCheckCdataPush(const xmlChar *utf, int len, int complete) {
- int ix;
- unsigned char c;
- int codepoint;
- if ((utf == NULL) || (len <= 0))
- return(0);
- for (ix = 0; ix < len;) { /* string is 0-terminated */
- c = utf[ix];
- if ((c & 0x80) == 0x00) { /* 1-byte code, starts with 10 */
- if (c >= 0x20)
- ix++;
- else if ((c == 0xA) || (c == 0xD) || (c == 0x9))
- ix++;
- else
- return(-ix);
- } else if ((c & 0xe0) == 0xc0) {/* 2-byte code, starts with 110 */
- if (ix + 2 > len) return(complete ? -ix : ix);
- if ((utf[ix+1] & 0xc0 ) != 0x80)
- return(-ix);
- codepoint = (utf[ix] & 0x1f) << 6;
- codepoint |= utf[ix+1] & 0x3f;
- if (!xmlIsCharQ(codepoint))
- return(-ix);
- ix += 2;
- } else if ((c & 0xf0) == 0xe0) {/* 3-byte code, starts with 1110 */
- if (ix + 3 > len) return(complete ? -ix : ix);
- if (((utf[ix+1] & 0xc0) != 0x80) ||
- ((utf[ix+2] & 0xc0) != 0x80))
- return(-ix);
- codepoint = (utf[ix] & 0xf) << 12;
- codepoint |= (utf[ix+1] & 0x3f) << 6;
- codepoint |= utf[ix+2] & 0x3f;
- if (!xmlIsCharQ(codepoint))
- return(-ix);
- ix += 3;
- } else if ((c & 0xf8) == 0xf0) {/* 4-byte code, starts with 11110 */
- if (ix + 4 > len) return(complete ? -ix : ix);
- if (((utf[ix+1] & 0xc0) != 0x80) ||
- ((utf[ix+2] & 0xc0) != 0x80) ||
- ((utf[ix+3] & 0xc0) != 0x80))
- return(-ix);
- codepoint = (utf[ix] & 0x7) << 18;
- codepoint |= (utf[ix+1] & 0x3f) << 12;
- codepoint |= (utf[ix+2] & 0x3f) << 6;
- codepoint |= utf[ix+3] & 0x3f;
- if (!xmlIsCharQ(codepoint))
- return(-ix);
- ix += 4;
- } else /* unknown encoding */
- return(-ix);
- }
- return(ix);
- }
- /**
- * xmlParseTryOrFinish:
- * @ctxt: an XML parser context
- * @terminate: last chunk indicator
- *
- * Try to progress on parsing
- *
- * Returns zero if no parsing was possible
- */
- static int
- xmlParseTryOrFinish(xmlParserCtxtPtr ctxt, int terminate) {
- int ret = 0;
- int avail, tlen;
- xmlChar cur, next;
- const xmlChar *lastlt, *lastgt;
- if (ctxt->input == NULL)
- return(0);
- #ifdef DEBUG_PUSH
- switch (ctxt->instate) {
- case XML_PARSER_EOF:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try EOF\n"); break;
- case XML_PARSER_START:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try START\n"); break;
- case XML_PARSER_MISC:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try MISC\n");break;
- case XML_PARSER_COMMENT:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try COMMENT\n");break;
- case XML_PARSER_PROLOG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try PROLOG\n");break;
- case XML_PARSER_START_TAG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try START_TAG\n");break;
- case XML_PARSER_CONTENT:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try CONTENT\n");break;
- case XML_PARSER_CDATA_SECTION:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try CDATA_SECTION\n");break;
- case XML_PARSER_END_TAG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try END_TAG\n");break;
- case XML_PARSER_ENTITY_DECL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try ENTITY_DECL\n");break;
- case XML_PARSER_ENTITY_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try ENTITY_VALUE\n");break;
- case XML_PARSER_ATTRIBUTE_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try ATTRIBUTE_VALUE\n");break;
- case XML_PARSER_DTD:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try DTD\n");break;
- case XML_PARSER_EPILOG:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try EPILOG\n");break;
- case XML_PARSER_PI:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try PI\n");break;
- case XML_PARSER_IGNORE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: try IGNORE\n");break;
- }
- #endif
- if ((ctxt->input != NULL) &&
- (ctxt->input->cur - ctxt->input->base > 4096)) {
- xmlSHRINK(ctxt);
- ctxt->checkIndex = 0;
- }
- xmlParseGetLasts(ctxt, &lastlt, &lastgt);
- while (ctxt->instate != XML_PARSER_EOF) {
- if ((ctxt->errNo != XML_ERR_OK) && (ctxt->disableSAX == 1))
- return(0);
- if (ctxt->input == NULL) break;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length -
- (ctxt->input->cur - ctxt->input->base);
- else {
- /*
- * If we are operating on converted input, try to flush
- * remaining chars to avoid them stalling in the non-converted
- * buffer. But do not do this in document start where
- * encoding="..." may not have been read and we work on a
- * guessed encoding.
- */
- if ((ctxt->instate != XML_PARSER_START) &&
- (ctxt->input->buf->raw != NULL) &&
- (xmlBufIsEmpty(ctxt->input->buf->raw) == 0)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer,
- ctxt->input);
- size_t current = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, 0, "");
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input,
- base, current);
- }
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- }
- if (avail < 1)
- goto done;
- switch (ctxt->instate) {
- case XML_PARSER_EOF:
- /*
- * Document parsing is done !
- */
- goto done;
- case XML_PARSER_START:
- if (ctxt->charset == XML_CHAR_ENCODING_NONE) {
- xmlChar start[4];
- xmlCharEncoding enc;
- /*
- * Very first chars read from the document flow.
- */
- if (avail < 4)
- goto done;
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines,
- * else xmlSwitchEncoding will set to (default)
- * UTF8.
- */
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- xmlSwitchEncoding(ctxt, enc);
- break;
- }
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if (cur == 0) {
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData,
- &xmlDefaultSAXLocator);
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- xmlHaltParser(ctxt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering EOF\n");
- #endif
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- if ((cur == '<') && (next == '?')) {
- /* PI or XML decl */
- if (avail < 5) return(ret);
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0))
- return(ret);
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData,
- &xmlDefaultSAXLocator);
- if ((ctxt->input->cur[2] == 'x') &&
- (ctxt->input->cur[3] == 'm') &&
- (ctxt->input->cur[4] == 'l') &&
- (IS_BLANK_CH(ctxt->input->cur[5]))) {
- ret += 5;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing XML Decl\n");
- #endif
- xmlParseXMLDecl(ctxt);
- if (ctxt->errNo == XML_ERR_UNSUPPORTED_ENCODING) {
- /*
- * The XML REC instructs us to stop parsing right
- * here
- */
- xmlHaltParser(ctxt);
- return(0);
- }
- ctxt->standalone = ctxt->input->standalone;
- if ((ctxt->encoding == NULL) &&
- (ctxt->input->encoding != NULL))
- ctxt->encoding = xmlStrdup(ctxt->input->encoding);
- if ((ctxt->sax) && (ctxt->sax->startDocument) &&
- (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- ctxt->instate = XML_PARSER_MISC;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering MISC\n");
- #endif
- } else {
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- if ((ctxt->sax) && (ctxt->sax->startDocument) &&
- (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- ctxt->instate = XML_PARSER_MISC;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering MISC\n");
- #endif
- }
- } else {
- if ((ctxt->sax) && (ctxt->sax->setDocumentLocator))
- ctxt->sax->setDocumentLocator(ctxt->userData,
- &xmlDefaultSAXLocator);
- ctxt->version = xmlCharStrdup(XML_DEFAULT_VERSION);
- if (ctxt->version == NULL) {
- xmlErrMemory(ctxt, NULL);
- break;
- }
- if ((ctxt->sax) && (ctxt->sax->startDocument) &&
- (!ctxt->disableSAX))
- ctxt->sax->startDocument(ctxt->userData);
- ctxt->instate = XML_PARSER_MISC;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering MISC\n");
- #endif
- }
- break;
- case XML_PARSER_START_TAG: {
- const xmlChar *name;
- const xmlChar *prefix = NULL;
- const xmlChar *URI = NULL;
- int line = ctxt->input->line;
- int nsNr = ctxt->nsNr;
- if ((avail < 2) && (ctxt->inputNr == 1))
- goto done;
- cur = ctxt->input->cur[0];
- if (cur != '<') {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_EMPTY, NULL);
- xmlHaltParser(ctxt);
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- if (!terminate) {
- if (ctxt->progressive) {
- /* > can be found unescaped in attribute values */
- if ((lastgt == NULL) || (ctxt->input->cur >= lastgt))
- goto done;
- } else if (xmlParseLookupSequence(ctxt, '>', 0, 0) < 0) {
- goto done;
- }
- }
- if (ctxt->spaceNr == 0)
- spacePush(ctxt, -1);
- else if (*ctxt->space == -2)
- spacePush(ctxt, -1);
- else
- spacePush(ctxt, *ctxt->space);
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax2)
- #endif /* LIBXML_SAX1_ENABLED */
- name = xmlParseStartTag2(ctxt, &prefix, &URI, &tlen);
- #ifdef LIBXML_SAX1_ENABLED
- else
- name = xmlParseStartTag(ctxt);
- #endif /* LIBXML_SAX1_ENABLED */
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- if (name == NULL) {
- spacePop(ctxt);
- xmlHaltParser(ctxt);
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- #ifdef LIBXML_VALID_ENABLED
- /*
- * [ VC: Root Element Type ]
- * The Name in the document type declaration must match
- * the element type of the root element.
- */
- if (ctxt->validate && ctxt->wellFormed && ctxt->myDoc &&
- ctxt->node && (ctxt->node == ctxt->myDoc->children))
- ctxt->valid &= xmlValidateRoot(&ctxt->vctxt, ctxt->myDoc);
- #endif /* LIBXML_VALID_ENABLED */
- /*
- * Check for an Empty Element.
- */
- if ((RAW == '/') && (NXT(1) == '>')) {
- SKIP(2);
- if (ctxt->sax2) {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->endElementNs != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElementNs(ctxt->userData, name,
- prefix, URI);
- if (ctxt->nsNr - nsNr > 0)
- nsPop(ctxt, ctxt->nsNr - nsNr);
- #ifdef LIBXML_SAX1_ENABLED
- } else {
- if ((ctxt->sax != NULL) &&
- (ctxt->sax->endElement != NULL) &&
- (!ctxt->disableSAX))
- ctxt->sax->endElement(ctxt->userData, name);
- #endif /* LIBXML_SAX1_ENABLED */
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- spacePop(ctxt);
- if (ctxt->nameNr == 0) {
- ctxt->instate = XML_PARSER_EPILOG;
- } else {
- ctxt->instate = XML_PARSER_CONTENT;
- }
- ctxt->progressive = 1;
- break;
- }
- if (RAW == '>') {
- NEXT;
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_GT_REQUIRED,
- "Couldn't find end of Start Tag %s\n",
- name);
- nodePop(ctxt);
- spacePop(ctxt);
- }
- nameNsPush(ctxt, name, prefix, URI, line, ctxt->nsNr - nsNr);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->progressive = 1;
- break;
- }
- case XML_PARSER_CONTENT: {
- const xmlChar *test;
- unsigned int cons;
- if ((avail < 2) && (ctxt->inputNr == 1))
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- test = CUR_PTR;
- cons = ctxt->input->consumed;
- if ((cur == '<') && (next == '/')) {
- ctxt->instate = XML_PARSER_END_TAG;
- break;
- } else if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- xmlParsePI(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next != '!')) {
- ctxt->instate = XML_PARSER_START_TAG;
- break;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') &&
- (ctxt->input->cur[3] == '-')) {
- int term;
- if (avail < 4)
- goto done;
- ctxt->input->cur += 4;
- term = xmlParseLookupSequence(ctxt, '-', '-', '>');
- ctxt->input->cur -= 4;
- if ((!terminate) && (term < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- xmlParseComment(ctxt);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (ctxt->input->cur[1] == '!') &&
- (ctxt->input->cur[2] == '[') &&
- (ctxt->input->cur[3] == 'C') &&
- (ctxt->input->cur[4] == 'D') &&
- (ctxt->input->cur[5] == 'A') &&
- (ctxt->input->cur[6] == 'T') &&
- (ctxt->input->cur[7] == 'A') &&
- (ctxt->input->cur[8] == '[')) {
- SKIP(9);
- ctxt->instate = XML_PARSER_CDATA_SECTION;
- break;
- } else if ((cur == '<') && (next == '!') &&
- (avail < 9)) {
- goto done;
- } else if (cur == '&') {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, ';', 0, 0) < 0))
- goto done;
- xmlParseReference(ctxt);
- } else {
- /* TODO Avoid the extra copy, handle directly !!! */
- /*
- * Goal of the following test is:
- * - minimize calls to the SAX 'character' callback
- * when they are mergeable
- * - handle an problem for isBlank when we only parse
- * a sequence of blank chars and the next one is
- * not available to check against '<' presence.
- * - tries to homogenize the differences in SAX
- * callbacks between the push and pull versions
- * of the parser.
- */
- if ((ctxt->inputNr == 1) &&
- (avail < XML_PARSER_BIG_BUFFER_SIZE)) {
- if (!terminate) {
- if (ctxt->progressive) {
- if ((lastlt == NULL) ||
- (ctxt->input->cur > lastlt))
- goto done;
- } else if (xmlParseLookupSequence(ctxt,
- '<', 0, 0) < 0) {
- goto done;
- }
- }
- }
- ctxt->checkIndex = 0;
- xmlParseCharData(ctxt, 0);
- }
- if ((cons == ctxt->input->consumed) && (test == CUR_PTR)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR,
- "detected an error in element content\n");
- xmlHaltParser(ctxt);
- break;
- }
- break;
- }
- case XML_PARSER_END_TAG:
- if (avail < 2)
- goto done;
- if (!terminate) {
- if (ctxt->progressive) {
- /* > can be found unescaped in attribute values */
- if ((lastgt == NULL) || (ctxt->input->cur >= lastgt))
- goto done;
- } else if (xmlParseLookupSequence(ctxt, '>', 0, 0) < 0) {
- goto done;
- }
- }
- if (ctxt->sax2) {
- xmlParseEndTag2(ctxt, &ctxt->pushTab[ctxt->nameNr - 1]);
- nameNsPop(ctxt);
- }
- #ifdef LIBXML_SAX1_ENABLED
- else
- xmlParseEndTag1(ctxt, 0);
- #endif /* LIBXML_SAX1_ENABLED */
- if (ctxt->instate == XML_PARSER_EOF) {
- /* Nothing */
- } else if (ctxt->nameNr == 0) {
- ctxt->instate = XML_PARSER_EPILOG;
- } else {
- ctxt->instate = XML_PARSER_CONTENT;
- }
- break;
- case XML_PARSER_CDATA_SECTION: {
- /*
- * The Push mode need to have the SAX callback for
- * cdataBlock merge back contiguous callbacks.
- */
- int base;
- base = xmlParseLookupSequence(ctxt, ']', ']', '>');
- if (base < 0) {
- if (avail >= XML_PARSER_BIG_BUFFER_SIZE + 2) {
- int tmp;
- tmp = xmlCheckCdataPush(ctxt->input->cur,
- XML_PARSER_BIG_BUFFER_SIZE, 0);
- if (tmp < 0) {
- tmp = -tmp;
- ctxt->input->cur += tmp;
- goto encoding_error;
- }
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX)) {
- if (ctxt->sax->cdataBlock != NULL)
- ctxt->sax->cdataBlock(ctxt->userData,
- ctxt->input->cur, tmp);
- else if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- ctxt->input->cur, tmp);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- SKIPL(tmp);
- ctxt->checkIndex = 0;
- }
- goto done;
- } else {
- int tmp;
- tmp = xmlCheckCdataPush(ctxt->input->cur, base, 1);
- if ((tmp < 0) || (tmp != base)) {
- tmp = -tmp;
- ctxt->input->cur += tmp;
- goto encoding_error;
- }
- if ((ctxt->sax != NULL) && (base == 0) &&
- (ctxt->sax->cdataBlock != NULL) &&
- (!ctxt->disableSAX)) {
- /*
- * Special case to provide identical behaviour
- * between pull and push parsers on enpty CDATA
- * sections
- */
- if ((ctxt->input->cur - ctxt->input->base >= 9) &&
- (!strncmp((const char *)&ctxt->input->cur[-9],
- "<![CDATA[", 9)))
- ctxt->sax->cdataBlock(ctxt->userData,
- BAD_CAST "", 0);
- } else if ((ctxt->sax != NULL) && (base > 0) &&
- (!ctxt->disableSAX)) {
- if (ctxt->sax->cdataBlock != NULL)
- ctxt->sax->cdataBlock(ctxt->userData,
- ctxt->input->cur, base);
- else if (ctxt->sax->characters != NULL)
- ctxt->sax->characters(ctxt->userData,
- ctxt->input->cur, base);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- SKIPL(base + 3);
- ctxt->checkIndex = 0;
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering CONTENT\n");
- #endif
- }
- break;
- }
- case XML_PARSER_MISC:
- SKIP_BLANKS;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length -
- (ctxt->input->cur - ctxt->input->base);
- else
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing PI\n");
- #endif
- xmlParsePI(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_MISC;
- ctxt->progressive = 1;
- ctxt->checkIndex = 0;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') &&
- (ctxt->input->cur[3] == '-')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '-', '-', '>') < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing Comment\n");
- #endif
- xmlParseComment(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_MISC;
- ctxt->progressive = 1;
- ctxt->checkIndex = 0;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == 'D') &&
- (ctxt->input->cur[3] == 'O') &&
- (ctxt->input->cur[4] == 'C') &&
- (ctxt->input->cur[5] == 'T') &&
- (ctxt->input->cur[6] == 'Y') &&
- (ctxt->input->cur[7] == 'P') &&
- (ctxt->input->cur[8] == 'E')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '>', 0, 0) < 0)) {
- ctxt->progressive = XML_PARSER_DTD;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing internal subset\n");
- #endif
- ctxt->inSubset = 1;
- ctxt->progressive = 0;
- ctxt->checkIndex = 0;
- xmlParseDocTypeDecl(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- if (RAW == '[') {
- ctxt->instate = XML_PARSER_DTD;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- } else {
- /*
- * Create and update the external subset.
- */
- ctxt->inSubset = 2;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->externalSubset != NULL))
- ctxt->sax->externalSubset(ctxt->userData,
- ctxt->intSubName, ctxt->extSubSystem,
- ctxt->extSubURI);
- ctxt->inSubset = 0;
- xmlCleanSpecialAttr(ctxt);
- ctxt->instate = XML_PARSER_PROLOG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering PROLOG\n");
- #endif
- }
- } else if ((cur == '<') && (next == '!') &&
- (avail < 9)) {
- goto done;
- } else {
- ctxt->instate = XML_PARSER_START_TAG;
- ctxt->progressive = XML_PARSER_START_TAG;
- xmlParseGetLasts(ctxt, &lastlt, &lastgt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- }
- break;
- case XML_PARSER_PROLOG:
- SKIP_BLANKS;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length - (ctxt->input->cur - ctxt->input->base);
- else
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing PI\n");
- #endif
- xmlParsePI(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_PROLOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') && (ctxt->input->cur[3] == '-')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '-', '-', '>') < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing Comment\n");
- #endif
- xmlParseComment(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_PROLOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (avail < 4)) {
- goto done;
- } else {
- ctxt->instate = XML_PARSER_START_TAG;
- if (ctxt->progressive == 0)
- ctxt->progressive = XML_PARSER_START_TAG;
- xmlParseGetLasts(ctxt, &lastlt, &lastgt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- }
- break;
- case XML_PARSER_EPILOG:
- SKIP_BLANKS;
- if (ctxt->input->buf == NULL)
- avail = ctxt->input->length - (ctxt->input->cur - ctxt->input->base);
- else
- avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- if (avail < 2)
- goto done;
- cur = ctxt->input->cur[0];
- next = ctxt->input->cur[1];
- if ((cur == '<') && (next == '?')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '?', '>', 0) < 0)) {
- ctxt->progressive = XML_PARSER_PI;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing PI\n");
- #endif
- xmlParsePI(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_EPILOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (ctxt->input->cur[2] == '-') && (ctxt->input->cur[3] == '-')) {
- if ((!terminate) &&
- (xmlParseLookupSequence(ctxt, '-', '-', '>') < 0)) {
- ctxt->progressive = XML_PARSER_COMMENT;
- goto done;
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: Parsing Comment\n");
- #endif
- xmlParseComment(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_EPILOG;
- ctxt->progressive = 1;
- } else if ((cur == '<') && (next == '!') &&
- (avail < 4)) {
- goto done;
- } else {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- xmlHaltParser(ctxt);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering EOF\n");
- #endif
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- goto done;
- }
- break;
- case XML_PARSER_DTD: {
- /*
- * Sorry but progressive parsing of the internal subset
- * is not expected to be supported. We first check that
- * the full content of the internal subset is available and
- * the parsing is launched only at that point.
- * Internal subset ends up with "']' S? '>'" in an unescaped
- * section and not in a ']]>' sequence which are conditional
- * sections (whoever argued to keep that crap in XML deserve
- * a place in hell !).
- */
- int base, i;
- xmlChar *buf;
- xmlChar quote = 0;
- size_t use;
- base = ctxt->input->cur - ctxt->input->base;
- if (base < 0) return(0);
- if (ctxt->checkIndex > base)
- base = ctxt->checkIndex;
- buf = xmlBufContent(ctxt->input->buf->buffer);
- use = xmlBufUse(ctxt->input->buf->buffer);
- for (;(unsigned int) base < use; base++) {
- if (quote != 0) {
- if (buf[base] == quote)
- quote = 0;
- continue;
- }
- if ((quote == 0) && (buf[base] == '<')) {
- int found = 0;
- /* special handling of comments */
- if (((unsigned int) base + 4 < use) &&
- (buf[base + 1] == '!') &&
- (buf[base + 2] == '-') &&
- (buf[base + 3] == '-')) {
- for (;(unsigned int) base + 3 < use; base++) {
- if ((buf[base] == '-') &&
- (buf[base + 1] == '-') &&
- (buf[base + 2] == '>')) {
- found = 1;
- base += 2;
- break;
- }
- }
- if (!found) {
- #if 0
- fprintf(stderr, "unfinished comment\n");
- #endif
- break; /* for */
- }
- continue;
- }
- }
- if (buf[base] == '"') {
- quote = '"';
- continue;
- }
- if (buf[base] == '\'') {
- quote = '\'';
- continue;
- }
- if (buf[base] == ']') {
- #if 0
- fprintf(stderr, "%c%c%c%c: ", buf[base],
- buf[base + 1], buf[base + 2], buf[base + 3]);
- #endif
- if ((unsigned int) base +1 >= use)
- break;
- if (buf[base + 1] == ']') {
- /* conditional crap, skip both ']' ! */
- base++;
- continue;
- }
- for (i = 1; (unsigned int) base + i < use; i++) {
- if (buf[base + i] == '>') {
- #if 0
- fprintf(stderr, "found\n");
- #endif
- goto found_end_int_subset;
- }
- if (!IS_BLANK_CH(buf[base + i])) {
- #if 0
- fprintf(stderr, "not found\n");
- #endif
- goto not_end_of_int_subset;
- }
- }
- #if 0
- fprintf(stderr, "end of stream\n");
- #endif
- break;
- }
- not_end_of_int_subset:
- continue; /* for */
- }
- /*
- * We didn't found the end of the Internal subset
- */
- if (quote == 0)
- ctxt->checkIndex = base;
- else
- ctxt->checkIndex = 0;
- #ifdef DEBUG_PUSH
- if (next == 0)
- xmlGenericError(xmlGenericErrorContext,
- "PP: lookup of int subset end filed\n");
- #endif
- goto done;
- found_end_int_subset:
- ctxt->checkIndex = 0;
- xmlParseInternalSubset(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->inSubset = 2;
- if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
- (ctxt->sax->externalSubset != NULL))
- ctxt->sax->externalSubset(ctxt->userData, ctxt->intSubName,
- ctxt->extSubSystem, ctxt->extSubURI);
- ctxt->inSubset = 0;
- xmlCleanSpecialAttr(ctxt);
- if (ctxt->instate == XML_PARSER_EOF)
- goto done;
- ctxt->instate = XML_PARSER_PROLOG;
- ctxt->checkIndex = 0;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering PROLOG\n");
- #endif
- break;
- }
- case XML_PARSER_COMMENT:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == COMMENT\n");
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering CONTENT\n");
- #endif
- break;
- case XML_PARSER_IGNORE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == IGNORE");
- ctxt->instate = XML_PARSER_DTD;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- break;
- case XML_PARSER_PI:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == PI\n");
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering CONTENT\n");
- #endif
- break;
- case XML_PARSER_ENTITY_DECL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == ENTITY_DECL\n");
- ctxt->instate = XML_PARSER_DTD;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- break;
- case XML_PARSER_ENTITY_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == ENTITY_VALUE\n");
- ctxt->instate = XML_PARSER_CONTENT;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering DTD\n");
- #endif
- break;
- case XML_PARSER_ATTRIBUTE_VALUE:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == ATTRIBUTE_VALUE\n");
- ctxt->instate = XML_PARSER_START_TAG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- break;
- case XML_PARSER_SYSTEM_LITERAL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == SYSTEM_LITERAL\n");
- ctxt->instate = XML_PARSER_START_TAG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- break;
- case XML_PARSER_PUBLIC_LITERAL:
- xmlGenericError(xmlGenericErrorContext,
- "PP: internal error, state == PUBLIC_LITERAL\n");
- ctxt->instate = XML_PARSER_START_TAG;
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext,
- "PP: entering START_TAG\n");
- #endif
- break;
- }
- }
- done:
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: done %d\n", ret);
- #endif
- return(ret);
- encoding_error:
- {
- char buffer[150];
- snprintf(buffer, 149, "Bytes: 0x%02X 0x%02X 0x%02X 0x%02X\n",
- ctxt->input->cur[0], ctxt->input->cur[1],
- ctxt->input->cur[2], ctxt->input->cur[3]);
- __xmlErrEncoding(ctxt, XML_ERR_INVALID_CHAR,
- "Input is not proper UTF-8, indicate encoding !\n%s",
- BAD_CAST buffer, NULL);
- }
- return(0);
- }
- /**
- * xmlParseCheckTransition:
- * @ctxt: an XML parser context
- * @chunk: a char array
- * @size: the size in byte of the chunk
- *
- * Check depending on the current parser state if the chunk given must be
- * processed immediately or one need more data to advance on parsing.
- *
- * Returns -1 in case of error, 0 if the push is not needed and 1 if needed
- */
- static int
- xmlParseCheckTransition(xmlParserCtxtPtr ctxt, const char *chunk, int size) {
- if ((ctxt == NULL) || (chunk == NULL) || (size < 0))
- return(-1);
- if (ctxt->instate == XML_PARSER_START_TAG) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->progressive == XML_PARSER_COMMENT) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->instate == XML_PARSER_CDATA_SECTION) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->progressive == XML_PARSER_PI) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if (ctxt->instate == XML_PARSER_END_TAG) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- if ((ctxt->progressive == XML_PARSER_DTD) ||
- (ctxt->instate == XML_PARSER_DTD)) {
- if (memchr(chunk, '>', size) != NULL)
- return(1);
- return(0);
- }
- return(1);
- }
- /**
- * xmlParseChunk:
- * @ctxt: an XML parser context
- * @chunk: an char array
- * @size: the size in byte of the chunk
- * @terminate: last chunk indicator
- *
- * Parse a Chunk of memory
- *
- * Returns zero if no error, the xmlParserErrors otherwise.
- */
- int
- xmlParseChunk(xmlParserCtxtPtr ctxt, const char *chunk, int size,
- int terminate) {
- int end_in_lf = 0;
- int remain = 0;
- size_t old_avail = 0;
- size_t avail = 0;
- if (ctxt == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- if ((ctxt->errNo != XML_ERR_OK) && (ctxt->disableSAX == 1))
- return(ctxt->errNo);
- if (ctxt->instate == XML_PARSER_EOF)
- return(-1);
- if (ctxt->instate == XML_PARSER_START)
- xmlDetectSAX2(ctxt);
- if ((size > 0) && (chunk != NULL) && (!terminate) &&
- (chunk[size - 1] == '\r')) {
- end_in_lf = 1;
- size--;
- }
- xmldecl_done:
- if ((size > 0) && (chunk != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL) && (ctxt->instate != XML_PARSER_EOF)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer, ctxt->input);
- size_t cur = ctxt->input->cur - ctxt->input->base;
- int res;
- old_avail = xmlBufUse(ctxt->input->buf->buffer);
- /*
- * Specific handling if we autodetected an encoding, we should not
- * push more than the first line ... which depend on the encoding
- * And only push the rest once the final encoding was detected
- */
- if ((ctxt->instate == XML_PARSER_START) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL) && (ctxt->input->buf->encoder != NULL)) {
- unsigned int len = 45;
- if ((xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UTF-16")) ||
- (xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UTF16")))
- len = 90;
- else if ((xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UCS-4")) ||
- (xmlStrcasestr(BAD_CAST ctxt->input->buf->encoder->name,
- BAD_CAST "UCS4")))
- len = 180;
- if (ctxt->input->buf->rawconsumed < len)
- len -= ctxt->input->buf->rawconsumed;
- /*
- * Change size for reading the initial declaration only
- * if size is greater than len. Otherwise, memmove in xmlBufferAdd
- * will blindly copy extra bytes from memory.
- */
- if ((unsigned int) size > len) {
- remain = size - len;
- size = len;
- } else {
- remain = 0;
- }
- }
- res = xmlParserInputBufferPush(ctxt->input->buf, size, chunk);
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input, base, cur);
- if (res < 0) {
- ctxt->errNo = XML_PARSER_EOF;
- xmlHaltParser(ctxt);
- return (XML_PARSER_EOF);
- }
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: pushed %d\n", size);
- #endif
- } else if (ctxt->instate != XML_PARSER_EOF) {
- if ((ctxt->input != NULL) && ctxt->input->buf != NULL) {
- xmlParserInputBufferPtr in = ctxt->input->buf;
- if ((in->encoder != NULL) && (in->buffer != NULL) &&
- (in->raw != NULL)) {
- int nbchars;
- size_t base = xmlBufGetInputBase(in->buffer, ctxt->input);
- size_t current = ctxt->input->cur - ctxt->input->base;
- nbchars = xmlCharEncInput(in, terminate);
- xmlBufSetInputBaseCur(in->buffer, ctxt->input, base, current);
- if (nbchars < 0) {
- /* TODO 2.6.0 */
- xmlGenericError(xmlGenericErrorContext,
- "xmlParseChunk: encoder error\n");
- xmlHaltParser(ctxt);
- return(XML_ERR_INVALID_ENCODING);
- }
- }
- }
- }
- if (remain != 0) {
- xmlParseTryOrFinish(ctxt, 0);
- } else {
- if ((ctxt->input != NULL) && (ctxt->input->buf != NULL))
- avail = xmlBufUse(ctxt->input->buf->buffer);
- /*
- * Depending on the current state it may not be such
- * a good idea to try parsing if there is nothing in the chunk
- * which would be worth doing a parser state transition and we
- * need to wait for more data
- */
- if ((terminate) || (avail > XML_MAX_TEXT_LENGTH) ||
- (old_avail == 0) || (avail == 0) ||
- (xmlParseCheckTransition(ctxt,
- (const char *)&ctxt->input->base[old_avail],
- avail - old_avail)))
- xmlParseTryOrFinish(ctxt, terminate);
- }
- if (ctxt->instate == XML_PARSER_EOF)
- return(ctxt->errNo);
- if ((ctxt->input != NULL) &&
- (((ctxt->input->end - ctxt->input->cur) > XML_MAX_LOOKUP_LIMIT) ||
- ((ctxt->input->cur - ctxt->input->base) > XML_MAX_LOOKUP_LIMIT)) &&
- ((ctxt->options & XML_PARSE_HUGE) == 0)) {
- xmlFatalErr(ctxt, XML_ERR_INTERNAL_ERROR, "Huge input lookup");
- xmlHaltParser(ctxt);
- }
- if ((ctxt->errNo != XML_ERR_OK) && (ctxt->disableSAX == 1))
- return(ctxt->errNo);
- if (remain != 0) {
- chunk += size;
- size = remain;
- remain = 0;
- goto xmldecl_done;
- }
- if ((end_in_lf == 1) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer,
- ctxt->input);
- size_t current = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, 1, "\r");
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input,
- base, current);
- }
- if (terminate) {
- /*
- * Check for termination
- */
- int cur_avail = 0;
- if (ctxt->input != NULL) {
- if (ctxt->input->buf == NULL)
- cur_avail = ctxt->input->length -
- (ctxt->input->cur - ctxt->input->base);
- else
- cur_avail = xmlBufUse(ctxt->input->buf->buffer) -
- (ctxt->input->cur - ctxt->input->base);
- }
- if ((ctxt->instate != XML_PARSER_EOF) &&
- (ctxt->instate != XML_PARSER_EPILOG)) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- }
- if ((ctxt->instate == XML_PARSER_EPILOG) && (cur_avail > 0)) {
- xmlFatalErr(ctxt, XML_ERR_DOCUMENT_END, NULL);
- }
- if (ctxt->instate != XML_PARSER_EOF) {
- if ((ctxt->sax) && (ctxt->sax->endDocument != NULL))
- ctxt->sax->endDocument(ctxt->userData);
- }
- ctxt->instate = XML_PARSER_EOF;
- }
- if (ctxt->wellFormed == 0)
- return((xmlParserErrors) ctxt->errNo);
- else
- return(0);
- }
- /************************************************************************
- * *
- * I/O front end functions to the parser *
- * *
- ************************************************************************/
- /**
- * xmlCreatePushParserCtxt:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @chunk: a pointer to an array of chars
- * @size: number of chars in the array
- * @filename: an optional file name or URI
- *
- * Create a parser context for using the XML parser in push mode.
- * If @buffer and @size are non-NULL, the data is used to detect
- * the encoding. The remaining characters will be parsed so they
- * don't need to be fed in again through xmlParseChunk.
- * To allow content encoding detection, @size should be >= 4
- * The value of @filename is used for fetching external entities
- * and error/warning reports.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreatePushParserCtxt(xmlSAXHandlerPtr sax, void *user_data,
- const char *chunk, int size, const char *filename) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- xmlParserInputBufferPtr buf;
- xmlCharEncoding enc = XML_CHAR_ENCODING_NONE;
- /*
- * plug some encoding conversion routines
- */
- if ((chunk != NULL) && (size >= 4))
- enc = xmlDetectCharEncoding((const xmlChar *) chunk, size);
- buf = xmlAllocParserInputBuffer(enc);
- if (buf == NULL) return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlErrMemory(NULL, "creating parser: out of memory\n");
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- ctxt->dictNames = 1;
- if (sax != NULL) {
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- #endif /* LIBXML_SAX1_ENABLED */
- xmlFree(ctxt->sax);
- ctxt->sax = (xmlSAXHandlerPtr) xmlMalloc(sizeof(xmlSAXHandler));
- if (ctxt->sax == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFreeParserInputBuffer(buf);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- memset(ctxt->sax, 0, sizeof(xmlSAXHandler));
- if (sax->initialized == XML_SAX2_MAGIC)
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandler));
- else
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandlerV1));
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- if (filename == NULL) {
- ctxt->directory = NULL;
- } else {
- ctxt->directory = xmlParserGetDirectory(filename);
- }
- inputStream = xmlNewInputStream(ctxt);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- if (filename == NULL)
- inputStream->filename = NULL;
- else {
- inputStream->filename = (char *)
- xmlCanonicPath((const xmlChar *) filename);
- if (inputStream->filename == NULL) {
- xmlFreeParserCtxt(ctxt);
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- }
- inputStream->buf = buf;
- xmlBufResetInput(inputStream->buf->buffer, inputStream);
- inputPush(ctxt, inputStream);
- /*
- * If the caller didn't provide an initial 'chunk' for determining
- * the encoding, we set the context to XML_CHAR_ENCODING_NONE so
- * that it can be automatically determined later
- */
- if ((size == 0) || (chunk == NULL)) {
- ctxt->charset = XML_CHAR_ENCODING_NONE;
- } else if ((ctxt->input != NULL) && (ctxt->input->buf != NULL)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer, ctxt->input);
- size_t cur = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, size, chunk);
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input, base, cur);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: pushed %d\n", size);
- #endif
- }
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- return(ctxt);
- }
- #endif /* LIBXML_PUSH_ENABLED */
- /**
- * xmlHaltParser:
- * @ctxt: an XML parser context
- *
- * Blocks further parser processing don't override error
- * for internal use
- */
- static void
- xmlHaltParser(xmlParserCtxtPtr ctxt) {
- if (ctxt == NULL)
- return;
- ctxt->instate = XML_PARSER_EOF;
- ctxt->disableSAX = 1;
- while (ctxt->inputNr > 1)
- xmlFreeInputStream(inputPop(ctxt));
- if (ctxt->input != NULL) {
- /*
- * in case there was a specific allocation deallocate before
- * overriding base
- */
- if (ctxt->input->free != NULL) {
- ctxt->input->free((xmlChar *) ctxt->input->base);
- ctxt->input->free = NULL;
- }
- if (ctxt->input->buf != NULL) {
- xmlFreeParserInputBuffer(ctxt->input->buf);
- ctxt->input->buf = NULL;
- }
- ctxt->input->cur = BAD_CAST"";
- ctxt->input->length = 0;
- ctxt->input->base = ctxt->input->cur;
- ctxt->input->end = ctxt->input->cur;
- }
- }
- /**
- * xmlStopParser:
- * @ctxt: an XML parser context
- *
- * Blocks further parser processing
- */
- void
- xmlStopParser(xmlParserCtxtPtr ctxt) {
- if (ctxt == NULL)
- return;
- xmlHaltParser(ctxt);
- ctxt->errNo = XML_ERR_USER_STOP;
- }
- /**
- * xmlCreateIOParserCtxt:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @ioread: an I/O read function
- * @ioclose: an I/O close function
- * @ioctx: an I/O handler
- * @enc: the charset encoding if known
- *
- * Create a parser context for using the XML parser with an existing
- * I/O stream
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateIOParserCtxt(xmlSAXHandlerPtr sax, void *user_data,
- xmlInputReadCallback ioread, xmlInputCloseCallback ioclose,
- void *ioctx, xmlCharEncoding enc) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- xmlParserInputBufferPtr buf;
- if (ioread == NULL) return(NULL);
- buf = xmlParserInputBufferCreateIO(ioread, ioclose, ioctx, enc);
- if (buf == NULL) {
- if (ioclose != NULL)
- ioclose(ioctx);
- return (NULL);
- }
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(buf);
- return(NULL);
- }
- if (sax != NULL) {
- #ifdef LIBXML_SAX1_ENABLED
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- #endif /* LIBXML_SAX1_ENABLED */
- xmlFree(ctxt->sax);
- ctxt->sax = (xmlSAXHandlerPtr) xmlMalloc(sizeof(xmlSAXHandler));
- if (ctxt->sax == NULL) {
- xmlErrMemory(ctxt, NULL);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- memset(ctxt->sax, 0, sizeof(xmlSAXHandler));
- if (sax->initialized == XML_SAX2_MAGIC)
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandler));
- else
- memcpy(ctxt->sax, sax, sizeof(xmlSAXHandlerV1));
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- inputStream = xmlNewIOInputStream(ctxt, buf, enc);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- return(ctxt);
- }
- #ifdef LIBXML_VALID_ENABLED
- /************************************************************************
- * *
- * Front ends when parsing a DTD *
- * *
- ************************************************************************/
- /**
- * xmlIOParseDTD:
- * @sax: the SAX handler block or NULL
- * @input: an Input Buffer
- * @enc: the charset encoding if known
- *
- * Load and parse a DTD
- *
- * Returns the resulting xmlDtdPtr or NULL in case of error.
- * @input will be freed by the function in any case.
- */
- xmlDtdPtr
- xmlIOParseDTD(xmlSAXHandlerPtr sax, xmlParserInputBufferPtr input,
- xmlCharEncoding enc) {
- xmlDtdPtr ret = NULL;
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr pinput = NULL;
- xmlChar start[4];
- if (input == NULL)
- return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(input);
- return(NULL);
- }
- /* We are loading a DTD */
- ctxt->options |= XML_PARSE_DTDLOAD;
- /*
- * Set-up the SAX context
- */
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- ctxt->userData = ctxt;
- }
- xmlDetectSAX2(ctxt);
- /*
- * generate a parser input from the I/O handler
- */
- pinput = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (pinput == NULL) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserInputBuffer(input);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- /*
- * plug some encoding conversion routines here.
- */
- if (xmlPushInput(ctxt, pinput) < 0) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- pinput->filename = NULL;
- pinput->line = 1;
- pinput->col = 1;
- pinput->base = ctxt->input->cur;
- pinput->cur = ctxt->input->cur;
- pinput->free = NULL;
- /*
- * let's parse that entity knowing it's an external subset.
- */
- ctxt->inSubset = 2;
- ctxt->myDoc = xmlNewDoc(BAD_CAST "1.0");
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- return(NULL);
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- ctxt->myDoc->extSubset = xmlNewDtd(ctxt->myDoc, BAD_CAST "none",
- BAD_CAST "none", BAD_CAST "none");
- if ((enc == XML_CHAR_ENCODING_NONE) &&
- ((ctxt->input->end - ctxt->input->cur) >= 4)) {
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- xmlParseExternalSubset(ctxt, BAD_CAST "none", BAD_CAST "none");
- if (ctxt->myDoc != NULL) {
- if (ctxt->wellFormed) {
- ret = ctxt->myDoc->extSubset;
- ctxt->myDoc->extSubset = NULL;
- if (ret != NULL) {
- xmlNodePtr tmp;
- ret->doc = NULL;
- tmp = ret->children;
- while (tmp != NULL) {
- tmp->doc = NULL;
- tmp = tmp->next;
- }
- }
- } else {
- ret = NULL;
- }
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlSAXParseDTD:
- * @sax: the SAX handler block
- * @ExternalID: a NAME* containing the External ID of the DTD
- * @SystemID: a NAME* containing the URL to the DTD
- *
- * Load and parse an external subset.
- *
- * Returns the resulting xmlDtdPtr or NULL in case of error.
- */
- xmlDtdPtr
- xmlSAXParseDTD(xmlSAXHandlerPtr sax, const xmlChar *ExternalID,
- const xmlChar *SystemID) {
- xmlDtdPtr ret = NULL;
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr input = NULL;
- xmlCharEncoding enc;
- xmlChar* systemIdCanonic;
- if ((ExternalID == NULL) && (SystemID == NULL)) return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- return(NULL);
- }
- /* We are loading a DTD */
- ctxt->options |= XML_PARSE_DTDLOAD;
- /*
- * Set-up the SAX context
- */
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- ctxt->userData = ctxt;
- }
- /*
- * Canonicalise the system ID
- */
- systemIdCanonic = xmlCanonicPath(SystemID);
- if ((SystemID != NULL) && (systemIdCanonic == NULL)) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- /*
- * Ask the Entity resolver to load the damn thing
- */
- if ((ctxt->sax != NULL) && (ctxt->sax->resolveEntity != NULL))
- input = ctxt->sax->resolveEntity(ctxt->userData, ExternalID,
- systemIdCanonic);
- if (input == NULL) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- if (systemIdCanonic != NULL)
- xmlFree(systemIdCanonic);
- return(NULL);
- }
- /*
- * plug some encoding conversion routines here.
- */
- if (xmlPushInput(ctxt, input) < 0) {
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- if (systemIdCanonic != NULL)
- xmlFree(systemIdCanonic);
- return(NULL);
- }
- if ((ctxt->input->end - ctxt->input->cur) >= 4) {
- enc = xmlDetectCharEncoding(ctxt->input->cur, 4);
- xmlSwitchEncoding(ctxt, enc);
- }
- if (input->filename == NULL)
- input->filename = (char *) systemIdCanonic;
- else
- xmlFree(systemIdCanonic);
- input->line = 1;
- input->col = 1;
- input->base = ctxt->input->cur;
- input->cur = ctxt->input->cur;
- input->free = NULL;
- /*
- * let's parse that entity knowing it's an external subset.
- */
- ctxt->inSubset = 2;
- ctxt->myDoc = xmlNewDoc(BAD_CAST "1.0");
- if (ctxt->myDoc == NULL) {
- xmlErrMemory(ctxt, "New Doc failed");
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- ctxt->myDoc->properties = XML_DOC_INTERNAL;
- ctxt->myDoc->extSubset = xmlNewDtd(ctxt->myDoc, BAD_CAST "none",
- ExternalID, SystemID);
- xmlParseExternalSubset(ctxt, ExternalID, SystemID);
- if (ctxt->myDoc != NULL) {
- if (ctxt->wellFormed) {
- ret = ctxt->myDoc->extSubset;
- ctxt->myDoc->extSubset = NULL;
- if (ret != NULL) {
- xmlNodePtr tmp;
- ret->doc = NULL;
- tmp = ret->children;
- while (tmp != NULL) {
- tmp->doc = NULL;
- tmp = tmp->next;
- }
- }
- } else {
- ret = NULL;
- }
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL) ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlParseDTD:
- * @ExternalID: a NAME* containing the External ID of the DTD
- * @SystemID: a NAME* containing the URL to the DTD
- *
- * Load and parse an external subset.
- *
- * Returns the resulting xmlDtdPtr or NULL in case of error.
- */
- xmlDtdPtr
- xmlParseDTD(const xmlChar *ExternalID, const xmlChar *SystemID) {
- return(xmlSAXParseDTD(NULL, ExternalID, SystemID));
- }
- #endif /* LIBXML_VALID_ENABLED */
- /************************************************************************
- * *
- * Front ends when parsing an Entity *
- * *
- ************************************************************************/
- /**
- * xmlParseCtxtExternalEntity:
- * @ctx: the existing parsing context
- * @URL: the URL for the entity to load
- * @ID: the System ID for the entity to load
- * @lst: the return value for the set of parsed nodes
- *
- * Parse an external general entity within an existing parsing context
- * An external general parsed entity is well-formed if it matches the
- * production labeled extParsedEnt.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * Returns 0 if the entity is well formed, -1 in case of args problem and
- * the parser error code otherwise
- */
- int
- xmlParseCtxtExternalEntity(xmlParserCtxtPtr ctx, const xmlChar *URL,
- const xmlChar *ID, xmlNodePtr *lst) {
- void *userData;
- if (ctx == NULL) return(-1);
- /*
- * If the user provided their own SAX callbacks, then reuse the
- * userData callback field, otherwise the expected setup in a
- * DOM builder is to have userData == ctxt
- */
- if (ctx->userData == ctx)
- userData = NULL;
- else
- userData = ctx->userData;
- return xmlParseExternalEntityPrivate(ctx->myDoc, ctx, ctx->sax,
- userData, ctx->depth + 1,
- URL, ID, lst);
- }
- /**
- * xmlParseExternalEntityPrivate:
- * @doc: the document the chunk pertains to
- * @oldctxt: the previous parser context if available
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @URL: the URL for the entity to load
- * @ID: the System ID for the entity to load
- * @list: the return value for the set of parsed nodes
- *
- * Private version of xmlParseExternalEntity()
- *
- * Returns 0 if the entity is well formed, -1 in case of args problem and
- * the parser error code otherwise
- */
- static xmlParserErrors
- xmlParseExternalEntityPrivate(xmlDocPtr doc, xmlParserCtxtPtr oldctxt,
- xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *URL,
- const xmlChar *ID, xmlNodePtr *list) {
- xmlParserCtxtPtr ctxt;
- xmlDocPtr newDoc;
- xmlNodePtr newRoot;
- xmlSAXHandlerPtr oldsax = NULL;
- xmlParserErrors ret = XML_ERR_OK;
- xmlChar start[4];
- xmlCharEncoding enc;
- if (((depth > 40) &&
- ((oldctxt == NULL) || (oldctxt->options & XML_PARSE_HUGE) == 0)) ||
- (depth > 1024)) {
- return(XML_ERR_ENTITY_LOOP);
- }
- if (list != NULL)
- *list = NULL;
- if ((URL == NULL) && (ID == NULL))
- return(XML_ERR_INTERNAL_ERROR);
- if (doc == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- ctxt = xmlCreateEntityParserCtxtInternal(URL, ID, NULL, oldctxt);
- if (ctxt == NULL) return(XML_WAR_UNDECLARED_ENTITY);
- ctxt->userData = ctxt;
- if (sax != NULL) {
- oldsax = ctxt->sax;
- ctxt->sax = sax;
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- xmlDetectSAX2(ctxt);
- newDoc = xmlNewDoc(BAD_CAST "1.0");
- if (newDoc == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(XML_ERR_INTERNAL_ERROR);
- }
- newDoc->properties = XML_DOC_INTERNAL;
- if (doc) {
- newDoc->intSubset = doc->intSubset;
- newDoc->extSubset = doc->extSubset;
- if (doc->dict) {
- newDoc->dict = doc->dict;
- xmlDictReference(newDoc->dict);
- }
- if (doc->URL != NULL) {
- newDoc->URL = xmlStrdup(doc->URL);
- }
- }
- newRoot = xmlNewDocNode(newDoc, NULL, BAD_CAST "pseudoroot", NULL);
- if (newRoot == NULL) {
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- xmlFreeDoc(newDoc);
- return(XML_ERR_INTERNAL_ERROR);
- }
- xmlAddChild((xmlNodePtr) newDoc, newRoot);
- nodePush(ctxt, newDoc->children);
- if (doc == NULL) {
- ctxt->myDoc = newDoc;
- } else {
- ctxt->myDoc = doc;
- newRoot->doc = doc;
- }
- /*
- * Get the 4 first bytes and decode the charset
- * if enc != XML_CHAR_ENCODING_NONE
- * plug some encoding conversion routines.
- */
- GROW;
- if ((ctxt->input->end - ctxt->input->cur) >= 4) {
- start[0] = RAW;
- start[1] = NXT(1);
- start[2] = NXT(2);
- start[3] = NXT(3);
- enc = xmlDetectCharEncoding(start, 4);
- if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- }
- /*
- * Parse a possible text declaration first
- */
- if ((CMP5(CUR_PTR, '<', '?', 'x', 'm', 'l')) && (IS_BLANK_CH(NXT(5)))) {
- xmlParseTextDecl(ctxt);
- /*
- * An XML-1.0 document can't reference an entity not XML-1.0
- */
- if ((xmlStrEqual(oldctxt->version, BAD_CAST "1.0")) &&
- (!xmlStrEqual(ctxt->input->version, BAD_CAST "1.0"))) {
- xmlFatalErrMsg(ctxt, XML_ERR_VERSION_MISMATCH,
- "Version mismatch between document and entity\n");
- }
- }
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->depth = depth;
- if (oldctxt != NULL) {
- ctxt->_private = oldctxt->_private;
- ctxt->loadsubset = oldctxt->loadsubset;
- ctxt->validate = oldctxt->validate;
- ctxt->valid = oldctxt->valid;
- ctxt->replaceEntities = oldctxt->replaceEntities;
- if (oldctxt->validate) {
- ctxt->vctxt.error = oldctxt->vctxt.error;
- ctxt->vctxt.warning = oldctxt->vctxt.warning;
- ctxt->vctxt.userData = oldctxt->vctxt.userData;
- }
- ctxt->external = oldctxt->external;
- if (ctxt->dict) xmlDictFree(ctxt->dict);
- ctxt->dict = oldctxt->dict;
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- ctxt->dictNames = oldctxt->dictNames;
- ctxt->attsDefault = oldctxt->attsDefault;
- ctxt->attsSpecial = oldctxt->attsSpecial;
- ctxt->linenumbers = oldctxt->linenumbers;
- ctxt->record_info = oldctxt->record_info;
- ctxt->node_seq.maximum = oldctxt->node_seq.maximum;
- ctxt->node_seq.length = oldctxt->node_seq.length;
- ctxt->node_seq.buffer = oldctxt->node_seq.buffer;
- } else {
- /*
- * Doing validity checking on chunk without context
- * doesn't make sense
- */
- ctxt->_private = NULL;
- ctxt->validate = 0;
- ctxt->external = 2;
- ctxt->loadsubset = 0;
- }
- xmlParseContent(ctxt);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if (ctxt->node != newDoc->children) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = XML_ERR_INTERNAL_ERROR;
- else
- ret = (xmlParserErrors)ctxt->errNo;
- } else {
- if (list != NULL) {
- xmlNodePtr cur;
- /*
- * Return the newly created nodeset after unlinking it from
- * they pseudo parent.
- */
- cur = newDoc->children->children;
- *list = cur;
- while (cur != NULL) {
- cur->parent = NULL;
- cur = cur->next;
- }
- newDoc->children->children = NULL;
- }
- ret = XML_ERR_OK;
- }
- /*
- * Record in the parent context the number of entities replacement
- * done when parsing that reference.
- */
- if (oldctxt != NULL)
- oldctxt->nbentities += ctxt->nbentities;
- /*
- * Also record the size of the entity parsed
- */
- if (ctxt->input != NULL && oldctxt != NULL) {
- oldctxt->sizeentities += ctxt->input->consumed;
- oldctxt->sizeentities += (ctxt->input->cur - ctxt->input->base);
- }
- /*
- * And record the last error if any
- */
- if ((oldctxt != NULL) && (ctxt->lastError.code != XML_ERR_OK))
- xmlCopyError(&ctxt->lastError, &oldctxt->lastError);
- if (sax != NULL)
- ctxt->sax = oldsax;
- if (oldctxt != NULL) {
- ctxt->dict = NULL;
- ctxt->attsDefault = NULL;
- ctxt->attsSpecial = NULL;
- oldctxt->validate = ctxt->validate;
- oldctxt->valid = ctxt->valid;
- oldctxt->node_seq.maximum = ctxt->node_seq.maximum;
- oldctxt->node_seq.length = ctxt->node_seq.length;
- oldctxt->node_seq.buffer = ctxt->node_seq.buffer;
- }
- ctxt->node_seq.maximum = 0;
- ctxt->node_seq.length = 0;
- ctxt->node_seq.buffer = NULL;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- xmlFreeDoc(newDoc);
- return(ret);
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlParseExternalEntity:
- * @doc: the document the chunk pertains to
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @URL: the URL for the entity to load
- * @ID: the System ID for the entity to load
- * @lst: the return value for the set of parsed nodes
- *
- * Parse an external general entity
- * An external general parsed entity is well-formed if it matches the
- * production labeled extParsedEnt.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * Returns 0 if the entity is well formed, -1 in case of args problem and
- * the parser error code otherwise
- */
- int
- xmlParseExternalEntity(xmlDocPtr doc, xmlSAXHandlerPtr sax, void *user_data,
- int depth, const xmlChar *URL, const xmlChar *ID, xmlNodePtr *lst) {
- return(xmlParseExternalEntityPrivate(doc, NULL, sax, user_data, depth, URL,
- ID, lst));
- }
- /**
- * xmlParseBalancedChunkMemory:
- * @doc: the document the chunk pertains to (must not be NULL)
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @string: the input string in UTF8 or ISO-Latin (zero terminated)
- * @lst: the return value for the set of parsed nodes
- *
- * Parse a well-balanced chunk of an XML document
- * called by the parser
- * The allowed sequence for the Well Balanced Chunk is the one defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns 0 if the chunk is well balanced, -1 in case of args problem and
- * the parser error code otherwise
- */
- int
- xmlParseBalancedChunkMemory(xmlDocPtr doc, xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *string, xmlNodePtr *lst) {
- return xmlParseBalancedChunkMemoryRecover( doc, sax, user_data,
- depth, string, lst, 0 );
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /**
- * xmlParseBalancedChunkMemoryInternal:
- * @oldctxt: the existing parsing context
- * @string: the input string in UTF8 or ISO-Latin (zero terminated)
- * @user_data: the user data field for the parser context
- * @lst: the return value for the set of parsed nodes
- *
- *
- * Parse a well-balanced chunk of an XML document
- * called by the parser
- * The allowed sequence for the Well Balanced Chunk is the one defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns XML_ERR_OK if the chunk is well balanced, and the parser
- * error code otherwise
- *
- * In case recover is set to 1, the nodelist will not be empty even if
- * the parsed chunk is not well balanced.
- */
- static xmlParserErrors
- xmlParseBalancedChunkMemoryInternal(xmlParserCtxtPtr oldctxt,
- const xmlChar *string, void *user_data, xmlNodePtr *lst) {
- xmlParserCtxtPtr ctxt;
- xmlDocPtr newDoc = NULL;
- xmlNodePtr newRoot;
- xmlSAXHandlerPtr oldsax = NULL;
- xmlNodePtr content = NULL;
- xmlNodePtr last = NULL;
- int size;
- xmlParserErrors ret = XML_ERR_OK;
- #ifdef SAX2
- int i;
- #endif
- if (((oldctxt->depth > 40) && ((oldctxt->options & XML_PARSE_HUGE) == 0)) ||
- (oldctxt->depth > 1024)) {
- return(XML_ERR_ENTITY_LOOP);
- }
- if (lst != NULL)
- *lst = NULL;
- if (string == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- size = xmlStrlen(string);
- ctxt = xmlCreateMemoryParserCtxt((char *) string, size);
- if (ctxt == NULL) return(XML_WAR_UNDECLARED_ENTITY);
- if (user_data != NULL)
- ctxt->userData = user_data;
- else
- ctxt->userData = ctxt;
- if (ctxt->dict != NULL) xmlDictFree(ctxt->dict);
- ctxt->dict = oldctxt->dict;
- ctxt->input_id = oldctxt->input_id + 1;
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- #ifdef SAX2
- /* propagate namespaces down the entity */
- for (i = 0;i < oldctxt->nsNr;i += 2) {
- nsPush(ctxt, oldctxt->nsTab[i], oldctxt->nsTab[i+1]);
- }
- #endif
- oldsax = ctxt->sax;
- ctxt->sax = oldctxt->sax;
- xmlDetectSAX2(ctxt);
- ctxt->replaceEntities = oldctxt->replaceEntities;
- ctxt->options = oldctxt->options;
- ctxt->_private = oldctxt->_private;
- if (oldctxt->myDoc == NULL) {
- newDoc = xmlNewDoc(BAD_CAST "1.0");
- if (newDoc == NULL) {
- ctxt->sax = oldsax;
- ctxt->dict = NULL;
- xmlFreeParserCtxt(ctxt);
- return(XML_ERR_INTERNAL_ERROR);
- }
- newDoc->properties = XML_DOC_INTERNAL;
- newDoc->dict = ctxt->dict;
- xmlDictReference(newDoc->dict);
- ctxt->myDoc = newDoc;
- } else {
- ctxt->myDoc = oldctxt->myDoc;
- content = ctxt->myDoc->children;
- last = ctxt->myDoc->last;
- }
- newRoot = xmlNewDocNode(ctxt->myDoc, NULL, BAD_CAST "pseudoroot", NULL);
- if (newRoot == NULL) {
- ctxt->sax = oldsax;
- ctxt->dict = NULL;
- xmlFreeParserCtxt(ctxt);
- if (newDoc != NULL) {
- xmlFreeDoc(newDoc);
- }
- return(XML_ERR_INTERNAL_ERROR);
- }
- ctxt->myDoc->children = NULL;
- ctxt->myDoc->last = NULL;
- xmlAddChild((xmlNodePtr) ctxt->myDoc, newRoot);
- nodePush(ctxt, ctxt->myDoc->children);
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->depth = oldctxt->depth + 1;
- ctxt->validate = 0;
- ctxt->loadsubset = oldctxt->loadsubset;
- if ((oldctxt->validate) || (oldctxt->replaceEntities != 0)) {
- /*
- * ID/IDREF registration will be done in xmlValidateElement below
- */
- ctxt->loadsubset |= XML_SKIP_IDS;
- }
- ctxt->dictNames = oldctxt->dictNames;
- ctxt->attsDefault = oldctxt->attsDefault;
- ctxt->attsSpecial = oldctxt->attsSpecial;
- xmlParseContent(ctxt);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if (ctxt->node != ctxt->myDoc->children) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = XML_ERR_INTERNAL_ERROR;
- else
- ret = (xmlParserErrors)ctxt->errNo;
- } else {
- ret = XML_ERR_OK;
- }
- if ((lst != NULL) && (ret == XML_ERR_OK)) {
- xmlNodePtr cur;
- /*
- * Return the newly created nodeset after unlinking it from
- * they pseudo parent.
- */
- cur = ctxt->myDoc->children->children;
- *lst = cur;
- while (cur != NULL) {
- #ifdef LIBXML_VALID_ENABLED
- if ((oldctxt->validate) && (oldctxt->wellFormed) &&
- (oldctxt->myDoc) && (oldctxt->myDoc->intSubset) &&
- (cur->type == XML_ELEMENT_NODE)) {
- oldctxt->valid &= xmlValidateElement(&oldctxt->vctxt,
- oldctxt->myDoc, cur);
- }
- #endif /* LIBXML_VALID_ENABLED */
- cur->parent = NULL;
- cur = cur->next;
- }
- ctxt->myDoc->children->children = NULL;
- }
- if (ctxt->myDoc != NULL) {
- xmlFreeNode(ctxt->myDoc->children);
- ctxt->myDoc->children = content;
- ctxt->myDoc->last = last;
- }
- /*
- * Record in the parent context the number of entities replacement
- * done when parsing that reference.
- */
- if (oldctxt != NULL)
- oldctxt->nbentities += ctxt->nbentities;
- /*
- * Also record the last error if any
- */
- if (ctxt->lastError.code != XML_ERR_OK)
- xmlCopyError(&ctxt->lastError, &oldctxt->lastError);
- ctxt->sax = oldsax;
- ctxt->dict = NULL;
- ctxt->attsDefault = NULL;
- ctxt->attsSpecial = NULL;
- xmlFreeParserCtxt(ctxt);
- if (newDoc != NULL) {
- xmlFreeDoc(newDoc);
- }
- return(ret);
- }
- /**
- * xmlParseInNodeContext:
- * @node: the context node
- * @data: the input string
- * @datalen: the input string length in bytes
- * @options: a combination of xmlParserOption
- * @lst: the return value for the set of parsed nodes
- *
- * Parse a well-balanced chunk of an XML document
- * within the context (DTD, namespaces, etc ...) of the given node.
- *
- * The allowed sequence for the data is a Well Balanced Chunk defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns XML_ERR_OK if the chunk is well balanced, and the parser
- * error code otherwise
- */
- xmlParserErrors
- xmlParseInNodeContext(xmlNodePtr node, const char *data, int datalen,
- int options, xmlNodePtr *lst) {
- #ifdef SAX2
- xmlParserCtxtPtr ctxt;
- xmlDocPtr doc = NULL;
- xmlNodePtr fake, cur;
- int nsnr = 0;
- xmlParserErrors ret = XML_ERR_OK;
- /*
- * check all input parameters, grab the document
- */
- if ((lst == NULL) || (node == NULL) || (data == NULL) || (datalen < 0))
- return(XML_ERR_INTERNAL_ERROR);
- switch (node->type) {
- case XML_ELEMENT_NODE:
- case XML_ATTRIBUTE_NODE:
- case XML_TEXT_NODE:
- case XML_CDATA_SECTION_NODE:
- case XML_ENTITY_REF_NODE:
- case XML_PI_NODE:
- case XML_COMMENT_NODE:
- case XML_DOCUMENT_NODE:
- case XML_HTML_DOCUMENT_NODE:
- break;
- default:
- return(XML_ERR_INTERNAL_ERROR);
- }
- while ((node != NULL) && (node->type != XML_ELEMENT_NODE) &&
- (node->type != XML_DOCUMENT_NODE) &&
- (node->type != XML_HTML_DOCUMENT_NODE))
- node = node->parent;
- if (node == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- if (node->type == XML_ELEMENT_NODE)
- doc = node->doc;
- else
- doc = (xmlDocPtr) node;
- if (doc == NULL)
- return(XML_ERR_INTERNAL_ERROR);
- /*
- * allocate a context and set-up everything not related to the
- * node position in the tree
- */
- if (doc->type == XML_DOCUMENT_NODE)
- ctxt = xmlCreateMemoryParserCtxt((char *) data, datalen);
- #ifdef LIBXML_HTML_ENABLED
- else if (doc->type == XML_HTML_DOCUMENT_NODE) {
- ctxt = htmlCreateMemoryParserCtxt((char *) data, datalen);
- /*
- * When parsing in context, it makes no sense to add implied
- * elements like html/body/etc...
- */
- options |= HTML_PARSE_NOIMPLIED;
- }
- #endif
- else
- return(XML_ERR_INTERNAL_ERROR);
- if (ctxt == NULL)
- return(XML_ERR_NO_MEMORY);
- /*
- * Use input doc's dict if present, else assure XML_PARSE_NODICT is set.
- * We need a dictionary for xmlDetectSAX2, so if there's no doc dict
- * we must wait until the last moment to free the original one.
- */
- if (doc->dict != NULL) {
- if (ctxt->dict != NULL)
- xmlDictFree(ctxt->dict);
- ctxt->dict = doc->dict;
- } else
- options |= XML_PARSE_NODICT;
- if (doc->encoding != NULL) {
- xmlCharEncodingHandlerPtr hdlr;
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = xmlStrdup((const xmlChar *) doc->encoding);
- hdlr = xmlFindCharEncodingHandler((const char *) doc->encoding);
- if (hdlr != NULL) {
- xmlSwitchToEncoding(ctxt, hdlr);
- } else {
- return(XML_ERR_UNSUPPORTED_ENCODING);
- }
- }
- xmlCtxtUseOptionsInternal(ctxt, options, NULL);
- xmlDetectSAX2(ctxt);
- ctxt->myDoc = doc;
- /* parsing in context, i.e. as within existing content */
- ctxt->input_id = 2;
- ctxt->instate = XML_PARSER_CONTENT;
- fake = xmlNewComment(NULL);
- if (fake == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(XML_ERR_NO_MEMORY);
- }
- xmlAddChild(node, fake);
- if (node->type == XML_ELEMENT_NODE) {
- nodePush(ctxt, node);
- /*
- * initialize the SAX2 namespaces stack
- */
- cur = node;
- while ((cur != NULL) && (cur->type == XML_ELEMENT_NODE)) {
- xmlNsPtr ns = cur->nsDef;
- const xmlChar *iprefix, *ihref;
- while (ns != NULL) {
- if (ctxt->dict) {
- iprefix = xmlDictLookup(ctxt->dict, ns->prefix, -1);
- ihref = xmlDictLookup(ctxt->dict, ns->href, -1);
- } else {
- iprefix = ns->prefix;
- ihref = ns->href;
- }
- if (xmlGetNamespace(ctxt, iprefix) == NULL) {
- nsPush(ctxt, iprefix, ihref);
- nsnr++;
- }
- ns = ns->next;
- }
- cur = cur->parent;
- }
- }
- if ((ctxt->validate) || (ctxt->replaceEntities != 0)) {
- /*
- * ID/IDREF registration will be done in xmlValidateElement below
- */
- ctxt->loadsubset |= XML_SKIP_IDS;
- }
- #ifdef LIBXML_HTML_ENABLED
- if (doc->type == XML_HTML_DOCUMENT_NODE)
- __htmlParseContent(ctxt);
- else
- #endif
- xmlParseContent(ctxt);
- nsPop(ctxt, nsnr);
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if ((ctxt->node != NULL) && (ctxt->node != node)) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- ctxt->wellFormed = 0;
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = XML_ERR_INTERNAL_ERROR;
- else
- ret = (xmlParserErrors)ctxt->errNo;
- } else {
- ret = XML_ERR_OK;
- }
- /*
- * Return the newly created nodeset after unlinking it from
- * the pseudo sibling.
- */
- cur = fake->next;
- fake->next = NULL;
- node->last = fake;
- if (cur != NULL) {
- cur->prev = NULL;
- }
- *lst = cur;
- while (cur != NULL) {
- cur->parent = NULL;
- cur = cur->next;
- }
- xmlUnlinkNode(fake);
- xmlFreeNode(fake);
- if (ret != XML_ERR_OK) {
- xmlFreeNodeList(*lst);
- *lst = NULL;
- }
- if (doc->dict != NULL)
- ctxt->dict = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- #else /* !SAX2 */
- return(XML_ERR_INTERNAL_ERROR);
- #endif
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlParseBalancedChunkMemoryRecover:
- * @doc: the document the chunk pertains to (must not be NULL)
- * @sax: the SAX handler block (possibly NULL)
- * @user_data: The user data returned on SAX callbacks (possibly NULL)
- * @depth: Used for loop detection, use 0
- * @string: the input string in UTF8 or ISO-Latin (zero terminated)
- * @lst: the return value for the set of parsed nodes
- * @recover: return nodes even if the data is broken (use 0)
- *
- *
- * Parse a well-balanced chunk of an XML document
- * called by the parser
- * The allowed sequence for the Well Balanced Chunk is the one defined by
- * the content production in the XML grammar:
- *
- * [43] content ::= (element | CharData | Reference | CDSect | PI | Comment)*
- *
- * Returns 0 if the chunk is well balanced, -1 in case of args problem and
- * the parser error code otherwise
- *
- * In case recover is set to 1, the nodelist will not be empty even if
- * the parsed chunk is not well balanced, assuming the parsing succeeded to
- * some extent.
- */
- int
- xmlParseBalancedChunkMemoryRecover(xmlDocPtr doc, xmlSAXHandlerPtr sax,
- void *user_data, int depth, const xmlChar *string, xmlNodePtr *lst,
- int recover) {
- xmlParserCtxtPtr ctxt;
- xmlDocPtr newDoc;
- xmlSAXHandlerPtr oldsax = NULL;
- xmlNodePtr content, newRoot;
- int size;
- int ret = 0;
- if (depth > 40) {
- return(XML_ERR_ENTITY_LOOP);
- }
- if (lst != NULL)
- *lst = NULL;
- if (string == NULL)
- return(-1);
- size = xmlStrlen(string);
- ctxt = xmlCreateMemoryParserCtxt((char *) string, size);
- if (ctxt == NULL) return(-1);
- ctxt->userData = ctxt;
- if (sax != NULL) {
- oldsax = ctxt->sax;
- ctxt->sax = sax;
- if (user_data != NULL)
- ctxt->userData = user_data;
- }
- newDoc = xmlNewDoc(BAD_CAST "1.0");
- if (newDoc == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(-1);
- }
- newDoc->properties = XML_DOC_INTERNAL;
- if ((doc != NULL) && (doc->dict != NULL)) {
- xmlDictFree(ctxt->dict);
- ctxt->dict = doc->dict;
- xmlDictReference(ctxt->dict);
- ctxt->str_xml = xmlDictLookup(ctxt->dict, BAD_CAST "xml", 3);
- ctxt->str_xmlns = xmlDictLookup(ctxt->dict, BAD_CAST "xmlns", 5);
- ctxt->str_xml_ns = xmlDictLookup(ctxt->dict, XML_XML_NAMESPACE, 36);
- ctxt->dictNames = 1;
- } else {
- xmlCtxtUseOptionsInternal(ctxt, XML_PARSE_NODICT, NULL);
- }
- /* doc == NULL is only supported for historic reasons */
- if (doc != NULL) {
- newDoc->intSubset = doc->intSubset;
- newDoc->extSubset = doc->extSubset;
- }
- newRoot = xmlNewDocNode(newDoc, NULL, BAD_CAST "pseudoroot", NULL);
- if (newRoot == NULL) {
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- xmlFreeDoc(newDoc);
- return(-1);
- }
- xmlAddChild((xmlNodePtr) newDoc, newRoot);
- nodePush(ctxt, newRoot);
- /* doc == NULL is only supported for historic reasons */
- if (doc == NULL) {
- ctxt->myDoc = newDoc;
- } else {
- ctxt->myDoc = newDoc;
- newDoc->children->doc = doc;
- /* Ensure that doc has XML spec namespace */
- xmlSearchNsByHref(doc, (xmlNodePtr)doc, XML_XML_NAMESPACE);
- newDoc->oldNs = doc->oldNs;
- }
- ctxt->instate = XML_PARSER_CONTENT;
- ctxt->input_id = 2;
- ctxt->depth = depth;
- /*
- * Doing validity checking on chunk doesn't make sense
- */
- ctxt->validate = 0;
- ctxt->loadsubset = 0;
- xmlDetectSAX2(ctxt);
- if ( doc != NULL ){
- content = doc->children;
- doc->children = NULL;
- xmlParseContent(ctxt);
- doc->children = content;
- }
- else {
- xmlParseContent(ctxt);
- }
- if ((RAW == '<') && (NXT(1) == '/')) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- } else if (RAW != 0) {
- xmlFatalErr(ctxt, XML_ERR_EXTRA_CONTENT, NULL);
- }
- if (ctxt->node != newDoc->children) {
- xmlFatalErr(ctxt, XML_ERR_NOT_WELL_BALANCED, NULL);
- }
- if (!ctxt->wellFormed) {
- if (ctxt->errNo == 0)
- ret = 1;
- else
- ret = ctxt->errNo;
- } else {
- ret = 0;
- }
- if ((lst != NULL) && ((ret == 0) || (recover == 1))) {
- xmlNodePtr cur;
- /*
- * Return the newly created nodeset after unlinking it from
- * they pseudo parent.
- */
- cur = newDoc->children->children;
- *lst = cur;
- while (cur != NULL) {
- xmlSetTreeDoc(cur, doc);
- cur->parent = NULL;
- cur = cur->next;
- }
- newDoc->children->children = NULL;
- }
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- newDoc->intSubset = NULL;
- newDoc->extSubset = NULL;
- /* This leaks the namespace list if doc == NULL */
- newDoc->oldNs = NULL;
- xmlFreeDoc(newDoc);
- return(ret);
- }
- /**
- * xmlSAXParseEntity:
- * @sax: the SAX handler block
- * @filename: the filename
- *
- * parse an XML external entity out of context and build a tree.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * This correspond to a "Well Balanced" chunk
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseEntity(xmlSAXHandlerPtr sax, const char *filename) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- ctxt = xmlCreateFileParserCtxt(filename);
- if (ctxt == NULL) {
- return(NULL);
- }
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- ctxt->userData = NULL;
- }
- xmlParseExtParsedEnt(ctxt);
- if (ctxt->wellFormed)
- ret = ctxt->myDoc;
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlParseEntity:
- * @filename: the filename
- *
- * parse an XML external entity out of context and build a tree.
- *
- * [78] extParsedEnt ::= TextDecl? content
- *
- * This correspond to a "Well Balanced" chunk
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlParseEntity(const char *filename) {
- return(xmlSAXParseEntity(NULL, filename));
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /**
- * xmlCreateEntityParserCtxtInternal:
- * @URL: the entity URL
- * @ID: the entity PUBLIC ID
- * @base: a possible base for the target URI
- * @pctx: parser context used to set options on new context
- *
- * Create a parser context for an external entity
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time.
- *
- * Returns the new parser context or NULL
- */
- static xmlParserCtxtPtr
- xmlCreateEntityParserCtxtInternal(const xmlChar *URL, const xmlChar *ID,
- const xmlChar *base, xmlParserCtxtPtr pctx) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- char *directory = NULL;
- xmlChar *uri;
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- return(NULL);
- }
- if (pctx != NULL) {
- ctxt->options = pctx->options;
- ctxt->_private = pctx->_private;
- /*
- * this is a subparser of pctx, so the input_id should be
- * incremented to distinguish from main entity
- */
- ctxt->input_id = pctx->input_id + 1;
- }
- /* Don't read from stdin. */
- if (xmlStrcmp(URL, BAD_CAST "-") == 0)
- URL = BAD_CAST "./-";
- uri = xmlBuildURI(URL, base);
- if (uri == NULL) {
- inputStream = xmlLoadExternalEntity((char *)URL, (char *)ID, ctxt);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- if ((ctxt->directory == NULL) && (directory == NULL))
- directory = xmlParserGetDirectory((char *)URL);
- if ((ctxt->directory == NULL) && (directory != NULL))
- ctxt->directory = directory;
- } else {
- inputStream = xmlLoadExternalEntity((char *)uri, (char *)ID, ctxt);
- if (inputStream == NULL) {
- xmlFree(uri);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- if ((ctxt->directory == NULL) && (directory == NULL))
- directory = xmlParserGetDirectory((char *)uri);
- if ((ctxt->directory == NULL) && (directory != NULL))
- ctxt->directory = directory;
- xmlFree(uri);
- }
- return(ctxt);
- }
- /**
- * xmlCreateEntityParserCtxt:
- * @URL: the entity URL
- * @ID: the entity PUBLIC ID
- * @base: a possible base for the target URI
- *
- * Create a parser context for an external entity
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateEntityParserCtxt(const xmlChar *URL, const xmlChar *ID,
- const xmlChar *base) {
- return xmlCreateEntityParserCtxtInternal(URL, ID, base, NULL);
- }
- /************************************************************************
- * *
- * Front ends when parsing from a file *
- * *
- ************************************************************************/
- /**
- * xmlCreateURLParserCtxt:
- * @filename: the filename or URL
- * @options: a combination of xmlParserOption
- *
- * Create a parser context for a file or URL content.
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time and for file accesses
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateURLParserCtxt(const char *filename, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr inputStream;
- char *directory = NULL;
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlErrMemory(NULL, "cannot allocate parser context");
- return(NULL);
- }
- if (options)
- xmlCtxtUseOptionsInternal(ctxt, options, NULL);
- ctxt->linenumbers = 1;
- inputStream = xmlLoadExternalEntity(filename, NULL, ctxt);
- if (inputStream == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- inputPush(ctxt, inputStream);
- if ((ctxt->directory == NULL) && (directory == NULL))
- directory = xmlParserGetDirectory(filename);
- if ((ctxt->directory == NULL) && (directory != NULL))
- ctxt->directory = directory;
- return(ctxt);
- }
- /**
- * xmlCreateFileParserCtxt:
- * @filename: the filename
- *
- * Create a parser context for a file content.
- * Automatic support for ZLIB/Compress compressed document is provided
- * by default if found at compile-time.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateFileParserCtxt(const char *filename)
- {
- return(xmlCreateURLParserCtxt(filename, 0));
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlSAXParseFileWithData:
- * @sax: the SAX handler block
- * @filename: the filename
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- * @data: the userdata
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * User data (void *) is stored within the parser context in the
- * context's _private member, so it is available nearly everywhere in libxml
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseFileWithData(xmlSAXHandlerPtr sax, const char *filename,
- int recovery, void *data) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateFileParserCtxt(filename);
- if (ctxt == NULL) {
- return(NULL);
- }
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- }
- xmlDetectSAX2(ctxt);
- if (data!=NULL) {
- ctxt->_private = data;
- }
- if (ctxt->directory == NULL)
- ctxt->directory = xmlParserGetDirectory(filename);
- ctxt->recovery = recovery;
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || recovery) {
- ret = ctxt->myDoc;
- if ((ret != NULL) && (ctxt->input->buf != NULL)) {
- if (ctxt->input->buf->compressed > 0)
- ret->compression = 9;
- else
- ret->compression = ctxt->input->buf->compressed;
- }
- }
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlSAXParseFile:
- * @sax: the SAX handler block
- * @filename: the filename
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseFile(xmlSAXHandlerPtr sax, const char *filename,
- int recovery) {
- return(xmlSAXParseFileWithData(sax,filename,recovery,NULL));
- }
- /**
- * xmlRecoverDoc:
- * @cur: a pointer to an array of xmlChar
- *
- * parse an XML in-memory document and build a tree.
- * In the case the document is not Well Formed, a attempt to build a
- * tree is tried anyway
- *
- * Returns the resulting document tree or NULL in case of failure
- */
- xmlDocPtr
- xmlRecoverDoc(const xmlChar *cur) {
- return(xmlSAXParseDoc(NULL, cur, 1));
- }
- /**
- * xmlParseFile:
- * @filename: the filename
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- *
- * Returns the resulting document tree if the file was wellformed,
- * NULL otherwise.
- */
- xmlDocPtr
- xmlParseFile(const char *filename) {
- return(xmlSAXParseFile(NULL, filename, 0));
- }
- /**
- * xmlRecoverFile:
- * @filename: the filename
- *
- * parse an XML file and build a tree. Automatic support for ZLIB/Compress
- * compressed document is provided by default if found at compile-time.
- * In the case the document is not Well Formed, it attempts to build
- * a tree anyway
- *
- * Returns the resulting document tree or NULL in case of failure
- */
- xmlDocPtr
- xmlRecoverFile(const char *filename) {
- return(xmlSAXParseFile(NULL, filename, 1));
- }
- /**
- * xmlSetupParserForBuffer:
- * @ctxt: an XML parser context
- * @buffer: a xmlChar * buffer
- * @filename: a file name
- *
- * Setup the parser context to parse a new buffer; Clears any prior
- * contents from the parser context. The buffer parameter must not be
- * NULL, but the filename parameter can be
- */
- void
- xmlSetupParserForBuffer(xmlParserCtxtPtr ctxt, const xmlChar* buffer,
- const char* filename)
- {
- xmlParserInputPtr input;
- if ((ctxt == NULL) || (buffer == NULL))
- return;
- input = xmlNewInputStream(ctxt);
- if (input == NULL) {
- xmlErrMemory(NULL, "parsing new buffer: out of memory\n");
- xmlClearParserCtxt(ctxt);
- return;
- }
- xmlClearParserCtxt(ctxt);
- if (filename != NULL)
- input->filename = (char *) xmlCanonicPath((const xmlChar *)filename);
- input->base = buffer;
- input->cur = buffer;
- input->end = &buffer[xmlStrlen(buffer)];
- inputPush(ctxt, input);
- }
- /**
- * xmlSAXUserParseFile:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @filename: a file name
- *
- * parse an XML file and call the given SAX handler routines.
- * Automatic support for ZLIB/Compress compressed document is provided
- *
- * Returns 0 in case of success or a error number otherwise
- */
- int
- xmlSAXUserParseFile(xmlSAXHandlerPtr sax, void *user_data,
- const char *filename) {
- int ret = 0;
- xmlParserCtxtPtr ctxt;
- ctxt = xmlCreateFileParserCtxt(filename);
- if (ctxt == NULL) return -1;
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- xmlDetectSAX2(ctxt);
- if (user_data != NULL)
- ctxt->userData = user_data;
- xmlParseDocument(ctxt);
- if (ctxt->wellFormed)
- ret = 0;
- else {
- if (ctxt->errNo != 0)
- ret = ctxt->errNo;
- else
- ret = -1;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- if (ctxt->myDoc != NULL) {
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- xmlFreeParserCtxt(ctxt);
- return ret;
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /************************************************************************
- * *
- * Front ends when parsing from memory *
- * *
- ************************************************************************/
- /**
- * xmlCreateMemoryParserCtxt:
- * @buffer: a pointer to a char array
- * @size: the size of the array
- *
- * Create a parser context for an XML in-memory document.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateMemoryParserCtxt(const char *buffer, int size) {
- xmlParserCtxtPtr ctxt;
- xmlParserInputPtr input;
- xmlParserInputBufferPtr buf;
- if (buffer == NULL)
- return(NULL);
- if (size <= 0)
- return(NULL);
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL)
- return(NULL);
- /* TODO: xmlParserInputBufferCreateStatic, requires some serious changes */
- buf = xmlParserInputBufferCreateMem(buffer, size, XML_CHAR_ENCODING_NONE);
- if (buf == NULL) {
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- input = xmlNewInputStream(ctxt);
- if (input == NULL) {
- xmlFreeParserInputBuffer(buf);
- xmlFreeParserCtxt(ctxt);
- return(NULL);
- }
- input->filename = NULL;
- input->buf = buf;
- xmlBufResetInput(input->buf->buffer, input);
- inputPush(ctxt, input);
- return(ctxt);
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlSAXParseMemoryWithData:
- * @sax: the SAX handler block
- * @buffer: an pointer to a char array
- * @size: the size of the array
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- * @data: the userdata
- *
- * parse an XML in-memory block and use the given SAX function block
- * to handle the parsing callback. If sax is NULL, fallback to the default
- * DOM tree building routines.
- *
- * User data (void *) is stored within the parser context in the
- * context's _private member, so it is available nearly everywhere in libxml
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseMemoryWithData(xmlSAXHandlerPtr sax, const char *buffer,
- int size, int recovery, void *data) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateMemoryParserCtxt(buffer, size);
- if (ctxt == NULL) return(NULL);
- if (sax != NULL) {
- if (ctxt->sax != NULL)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- }
- xmlDetectSAX2(ctxt);
- if (data!=NULL) {
- ctxt->_private=data;
- }
- ctxt->recovery = recovery;
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || recovery) ret = ctxt->myDoc;
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlSAXParseMemory:
- * @sax: the SAX handler block
- * @buffer: an pointer to a char array
- * @size: the size of the array
- * @recovery: work in recovery mode, i.e. tries to read not Well Formed
- * documents
- *
- * parse an XML in-memory block and use the given SAX function block
- * to handle the parsing callback. If sax is NULL, fallback to the default
- * DOM tree building routines.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseMemory(xmlSAXHandlerPtr sax, const char *buffer,
- int size, int recovery) {
- return xmlSAXParseMemoryWithData(sax, buffer, size, recovery, NULL);
- }
- /**
- * xmlParseMemory:
- * @buffer: an pointer to a char array
- * @size: the size of the array
- *
- * parse an XML in-memory block and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr xmlParseMemory(const char *buffer, int size) {
- return(xmlSAXParseMemory(NULL, buffer, size, 0));
- }
- /**
- * xmlRecoverMemory:
- * @buffer: an pointer to a char array
- * @size: the size of the array
- *
- * parse an XML in-memory block and build a tree.
- * In the case the document is not Well Formed, an attempt to
- * build a tree is tried anyway
- *
- * Returns the resulting document tree or NULL in case of error
- */
- xmlDocPtr xmlRecoverMemory(const char *buffer, int size) {
- return(xmlSAXParseMemory(NULL, buffer, size, 1));
- }
- /**
- * xmlSAXUserParseMemory:
- * @sax: a SAX handler
- * @user_data: The user data returned on SAX callbacks
- * @buffer: an in-memory XML document input
- * @size: the length of the XML document in bytes
- *
- * A better SAX parsing routine.
- * parse an XML in-memory buffer and call the given SAX handler routines.
- *
- * Returns 0 in case of success or a error number otherwise
- */
- int xmlSAXUserParseMemory(xmlSAXHandlerPtr sax, void *user_data,
- const char *buffer, int size) {
- int ret = 0;
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateMemoryParserCtxt(buffer, size);
- if (ctxt == NULL) return -1;
- if (ctxt->sax != (xmlSAXHandlerPtr) &xmlDefaultSAXHandler)
- xmlFree(ctxt->sax);
- ctxt->sax = sax;
- xmlDetectSAX2(ctxt);
- if (user_data != NULL)
- ctxt->userData = user_data;
- xmlParseDocument(ctxt);
- if (ctxt->wellFormed)
- ret = 0;
- else {
- if (ctxt->errNo != 0)
- ret = ctxt->errNo;
- else
- ret = -1;
- }
- if (sax != NULL)
- ctxt->sax = NULL;
- if (ctxt->myDoc != NULL) {
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- xmlFreeParserCtxt(ctxt);
- return ret;
- }
- #endif /* LIBXML_SAX1_ENABLED */
- /**
- * xmlCreateDocParserCtxt:
- * @cur: a pointer to an array of xmlChar
- *
- * Creates a parser context for an XML in-memory document.
- *
- * Returns the new parser context or NULL
- */
- xmlParserCtxtPtr
- xmlCreateDocParserCtxt(const xmlChar *cur) {
- int len;
- if (cur == NULL)
- return(NULL);
- len = xmlStrlen(cur);
- return(xmlCreateMemoryParserCtxt((const char *)cur, len));
- }
- #ifdef LIBXML_SAX1_ENABLED
- /**
- * xmlSAXParseDoc:
- * @sax: the SAX handler block
- * @cur: a pointer to an array of xmlChar
- * @recovery: work in recovery mode, i.e. tries to read no Well Formed
- * documents
- *
- * parse an XML in-memory document and build a tree.
- * It use the given SAX function block to handle the parsing callback.
- * If sax is NULL, fallback to the default DOM tree building routines.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlSAXParseDoc(xmlSAXHandlerPtr sax, const xmlChar *cur, int recovery) {
- xmlDocPtr ret;
- xmlParserCtxtPtr ctxt;
- xmlSAXHandlerPtr oldsax = NULL;
- if (cur == NULL) return(NULL);
- ctxt = xmlCreateDocParserCtxt(cur);
- if (ctxt == NULL) return(NULL);
- if (sax != NULL) {
- oldsax = ctxt->sax;
- ctxt->sax = sax;
- ctxt->userData = NULL;
- }
- xmlDetectSAX2(ctxt);
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || recovery) ret = ctxt->myDoc;
- else {
- ret = NULL;
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- }
- if (sax != NULL)
- ctxt->sax = oldsax;
- xmlFreeParserCtxt(ctxt);
- return(ret);
- }
- /**
- * xmlParseDoc:
- * @cur: a pointer to an array of xmlChar
- *
- * parse an XML in-memory document and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlParseDoc(const xmlChar *cur) {
- return(xmlSAXParseDoc(NULL, cur, 0));
- }
- #endif /* LIBXML_SAX1_ENABLED */
- #ifdef LIBXML_LEGACY_ENABLED
- /************************************************************************
- * *
- * Specific function to keep track of entities references *
- * and used by the XSLT debugger *
- * *
- ************************************************************************/
- static xmlEntityReferenceFunc xmlEntityRefFunc = NULL;
- /**
- * xmlAddEntityReference:
- * @ent : A valid entity
- * @firstNode : A valid first node for children of entity
- * @lastNode : A valid last node of children entity
- *
- * Notify of a reference to an entity of type XML_EXTERNAL_GENERAL_PARSED_ENTITY
- */
- static void
- xmlAddEntityReference(xmlEntityPtr ent, xmlNodePtr firstNode,
- xmlNodePtr lastNode)
- {
- if (xmlEntityRefFunc != NULL) {
- (*xmlEntityRefFunc) (ent, firstNode, lastNode);
- }
- }
- /**
- * xmlSetEntityReferenceFunc:
- * @func: A valid function
- *
- * Set the function to call call back when a xml reference has been made
- */
- void
- xmlSetEntityReferenceFunc(xmlEntityReferenceFunc func)
- {
- xmlEntityRefFunc = func;
- }
- #endif /* LIBXML_LEGACY_ENABLED */
- /************************************************************************
- * *
- * Miscellaneous *
- * *
- ************************************************************************/
- #ifdef LIBXML_XPATH_ENABLED
- #include <libxml/xpath.h>
- #endif
- extern void XMLCDECL xmlGenericErrorDefaultFunc(void *ctx, const char *msg, ...);
- static int xmlParserInitialized = 0;
- /**
- * xmlInitParser:
- *
- * Initialization function for the XML parser.
- * This is not reentrant. Call once before processing in case of
- * use in multithreaded programs.
- */
- void
- xmlInitParser(void) {
- if (xmlParserInitialized != 0)
- return;
- #if defined(_WIN32) && (!defined(LIBXML_STATIC) || defined(LIBXML_STATIC_FOR_DLL))
- atexit(xmlCleanupParser);
- #endif
- #ifdef LIBXML_THREAD_ENABLED
- __xmlGlobalInitMutexLock();
- if (xmlParserInitialized == 0) {
- #endif
- xmlInitThreads();
- xmlInitGlobals();
- if ((xmlGenericError == xmlGenericErrorDefaultFunc) ||
- (xmlGenericError == NULL))
- initGenericErrorDefaultFunc(NULL);
- xmlInitMemory();
- xmlInitializeDict();
- xmlInitCharEncodingHandlers();
- xmlDefaultSAXHandlerInit();
- xmlRegisterDefaultInputCallbacks();
- #ifdef LIBXML_OUTPUT_ENABLED
- xmlRegisterDefaultOutputCallbacks();
- #endif /* LIBXML_OUTPUT_ENABLED */
- #ifdef LIBXML_HTML_ENABLED
- htmlInitAutoClose();
- htmlDefaultSAXHandlerInit();
- #endif
- #ifdef LIBXML_XPATH_ENABLED
- xmlXPathInit();
- #endif
- xmlParserInitialized = 1;
- #ifdef LIBXML_THREAD_ENABLED
- }
- __xmlGlobalInitMutexUnlock();
- #endif
- }
- /**
- * xmlCleanupParser:
- *
- * This function name is somewhat misleading. It does not clean up
- * parser state, it cleans up memory allocated by the library itself.
- * It is a cleanup function for the XML library. It tries to reclaim all
- * related global memory allocated for the library processing.
- * It doesn't deallocate any document related memory. One should
- * call xmlCleanupParser() only when the process has finished using
- * the library and all XML/HTML documents built with it.
- * See also xmlInitParser() which has the opposite function of preparing
- * the library for operations.
- *
- * WARNING: if your application is multithreaded or has plugin support
- * calling this may crash the application if another thread or
- * a plugin is still using libxml2. It's sometimes very hard to
- * guess if libxml2 is in use in the application, some libraries
- * or plugins may use it without notice. In case of doubt abstain
- * from calling this function or do it just before calling exit()
- * to avoid leak reports from valgrind !
- */
- void
- xmlCleanupParser(void) {
- if (!xmlParserInitialized)
- return;
- xmlCleanupCharEncodingHandlers();
- #ifdef LIBXML_CATALOG_ENABLED
- xmlCatalogCleanup();
- #endif
- xmlDictCleanup();
- xmlCleanupInputCallbacks();
- #ifdef LIBXML_OUTPUT_ENABLED
- xmlCleanupOutputCallbacks();
- #endif
- #ifdef LIBXML_SCHEMAS_ENABLED
- xmlSchemaCleanupTypes();
- xmlRelaxNGCleanupTypes();
- #endif
- xmlResetLastError();
- xmlCleanupGlobals();
- xmlCleanupThreads(); /* must be last if called not from the main thread */
- xmlCleanupMemory();
- xmlParserInitialized = 0;
- }
- #if defined(HAVE_ATTRIBUTE_DESTRUCTOR) && !defined(LIBXML_STATIC) && \
- !defined(_WIN32)
- static void
- ATTRIBUTE_DESTRUCTOR
- xmlDestructor(void) {
- /*
- * Calling custom deallocation functions in a destructor can cause
- * problems, for example with Nokogiri.
- */
- if (xmlFree == free)
- xmlCleanupParser();
- }
- #endif
- /************************************************************************
- * *
- * New set (2.6.0) of simpler and more flexible APIs *
- * *
- ************************************************************************/
- /**
- * DICT_FREE:
- * @str: a string
- *
- * Free a string if it is not owned by the "dict" dictionary in the
- * current scope
- */
- #define DICT_FREE(str) \
- if ((str) && ((!dict) || \
- (xmlDictOwns(dict, (const xmlChar *)(str)) == 0))) \
- xmlFree((char *)(str));
- /**
- * xmlCtxtReset:
- * @ctxt: an XML parser context
- *
- * Reset a parser context
- */
- void
- xmlCtxtReset(xmlParserCtxtPtr ctxt)
- {
- xmlParserInputPtr input;
- xmlDictPtr dict;
- if (ctxt == NULL)
- return;
- dict = ctxt->dict;
- while ((input = inputPop(ctxt)) != NULL) { /* Non consuming */
- xmlFreeInputStream(input);
- }
- ctxt->inputNr = 0;
- ctxt->input = NULL;
- ctxt->spaceNr = 0;
- if (ctxt->spaceTab != NULL) {
- ctxt->spaceTab[0] = -1;
- ctxt->space = &ctxt->spaceTab[0];
- } else {
- ctxt->space = NULL;
- }
- ctxt->nodeNr = 0;
- ctxt->node = NULL;
- ctxt->nameNr = 0;
- ctxt->name = NULL;
- DICT_FREE(ctxt->version);
- ctxt->version = NULL;
- DICT_FREE(ctxt->encoding);
- ctxt->encoding = NULL;
- DICT_FREE(ctxt->directory);
- ctxt->directory = NULL;
- DICT_FREE(ctxt->extSubURI);
- ctxt->extSubURI = NULL;
- DICT_FREE(ctxt->extSubSystem);
- ctxt->extSubSystem = NULL;
- if (ctxt->myDoc != NULL)
- xmlFreeDoc(ctxt->myDoc);
- ctxt->myDoc = NULL;
- ctxt->standalone = -1;
- ctxt->hasExternalSubset = 0;
- ctxt->hasPErefs = 0;
- ctxt->html = 0;
- ctxt->external = 0;
- ctxt->instate = XML_PARSER_START;
- ctxt->token = 0;
- ctxt->wellFormed = 1;
- ctxt->nsWellFormed = 1;
- ctxt->disableSAX = 0;
- ctxt->valid = 1;
- #if 0
- ctxt->vctxt.userData = ctxt;
- ctxt->vctxt.error = xmlParserValidityError;
- ctxt->vctxt.warning = xmlParserValidityWarning;
- #endif
- ctxt->record_info = 0;
- ctxt->checkIndex = 0;
- ctxt->inSubset = 0;
- ctxt->errNo = XML_ERR_OK;
- ctxt->depth = 0;
- ctxt->charset = XML_CHAR_ENCODING_UTF8;
- ctxt->catalogs = NULL;
- ctxt->nbentities = 0;
- ctxt->sizeentities = 0;
- ctxt->sizeentcopy = 0;
- xmlInitNodeInfoSeq(&ctxt->node_seq);
- if (ctxt->attsDefault != NULL) {
- xmlHashFree(ctxt->attsDefault, xmlHashDefaultDeallocator);
- ctxt->attsDefault = NULL;
- }
- if (ctxt->attsSpecial != NULL) {
- xmlHashFree(ctxt->attsSpecial, NULL);
- ctxt->attsSpecial = NULL;
- }
- #ifdef LIBXML_CATALOG_ENABLED
- if (ctxt->catalogs != NULL)
- xmlCatalogFreeLocal(ctxt->catalogs);
- #endif
- if (ctxt->lastError.code != XML_ERR_OK)
- xmlResetError(&ctxt->lastError);
- }
- /**
- * xmlCtxtResetPush:
- * @ctxt: an XML parser context
- * @chunk: a pointer to an array of chars
- * @size: number of chars in the array
- * @filename: an optional file name or URI
- * @encoding: the document encoding, or NULL
- *
- * Reset a push parser context
- *
- * Returns 0 in case of success and 1 in case of error
- */
- int
- xmlCtxtResetPush(xmlParserCtxtPtr ctxt, const char *chunk,
- int size, const char *filename, const char *encoding)
- {
- xmlParserInputPtr inputStream;
- xmlParserInputBufferPtr buf;
- xmlCharEncoding enc = XML_CHAR_ENCODING_NONE;
- if (ctxt == NULL)
- return(1);
- if ((encoding == NULL) && (chunk != NULL) && (size >= 4))
- enc = xmlDetectCharEncoding((const xmlChar *) chunk, size);
- buf = xmlAllocParserInputBuffer(enc);
- if (buf == NULL)
- return(1);
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(buf);
- return(1);
- }
- xmlCtxtReset(ctxt);
- if (filename == NULL) {
- ctxt->directory = NULL;
- } else {
- ctxt->directory = xmlParserGetDirectory(filename);
- }
- inputStream = xmlNewInputStream(ctxt);
- if (inputStream == NULL) {
- xmlFreeParserInputBuffer(buf);
- return(1);
- }
- if (filename == NULL)
- inputStream->filename = NULL;
- else
- inputStream->filename = (char *)
- xmlCanonicPath((const xmlChar *) filename);
- inputStream->buf = buf;
- xmlBufResetInput(buf->buffer, inputStream);
- inputPush(ctxt, inputStream);
- if ((size > 0) && (chunk != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->buf != NULL)) {
- size_t base = xmlBufGetInputBase(ctxt->input->buf->buffer, ctxt->input);
- size_t cur = ctxt->input->cur - ctxt->input->base;
- xmlParserInputBufferPush(ctxt->input->buf, size, chunk);
- xmlBufSetInputBaseCur(ctxt->input->buf->buffer, ctxt->input, base, cur);
- #ifdef DEBUG_PUSH
- xmlGenericError(xmlGenericErrorContext, "PP: pushed %d\n", size);
- #endif
- }
- if (encoding != NULL) {
- xmlCharEncodingHandlerPtr hdlr;
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = xmlStrdup((const xmlChar *) encoding);
- hdlr = xmlFindCharEncodingHandler(encoding);
- if (hdlr != NULL) {
- xmlSwitchToEncoding(ctxt, hdlr);
- } else {
- xmlFatalErrMsgStr(ctxt, XML_ERR_UNSUPPORTED_ENCODING,
- "Unsupported encoding %s\n", BAD_CAST encoding);
- }
- } else if (enc != XML_CHAR_ENCODING_NONE) {
- xmlSwitchEncoding(ctxt, enc);
- }
- return(0);
- }
- /**
- * xmlCtxtUseOptionsInternal:
- * @ctxt: an XML parser context
- * @options: a combination of xmlParserOption
- * @encoding: the user provided encoding to use
- *
- * Applies the options to the parser context
- *
- * Returns 0 in case of success, the set of unknown or unimplemented options
- * in case of error.
- */
- static int
- xmlCtxtUseOptionsInternal(xmlParserCtxtPtr ctxt, int options, const char *encoding)
- {
- if (ctxt == NULL)
- return(-1);
- if (encoding != NULL) {
- if (ctxt->encoding != NULL)
- xmlFree((xmlChar *) ctxt->encoding);
- ctxt->encoding = xmlStrdup((const xmlChar *) encoding);
- }
- if (options & XML_PARSE_RECOVER) {
- ctxt->recovery = 1;
- options -= XML_PARSE_RECOVER;
- ctxt->options |= XML_PARSE_RECOVER;
- } else
- ctxt->recovery = 0;
- if (options & XML_PARSE_DTDLOAD) {
- ctxt->loadsubset = XML_DETECT_IDS;
- options -= XML_PARSE_DTDLOAD;
- ctxt->options |= XML_PARSE_DTDLOAD;
- } else
- ctxt->loadsubset = 0;
- if (options & XML_PARSE_DTDATTR) {
- ctxt->loadsubset |= XML_COMPLETE_ATTRS;
- options -= XML_PARSE_DTDATTR;
- ctxt->options |= XML_PARSE_DTDATTR;
- }
- if (options & XML_PARSE_NOENT) {
- ctxt->replaceEntities = 1;
- /* ctxt->loadsubset |= XML_DETECT_IDS; */
- options -= XML_PARSE_NOENT;
- ctxt->options |= XML_PARSE_NOENT;
- } else
- ctxt->replaceEntities = 0;
- if (options & XML_PARSE_PEDANTIC) {
- ctxt->pedantic = 1;
- options -= XML_PARSE_PEDANTIC;
- ctxt->options |= XML_PARSE_PEDANTIC;
- } else
- ctxt->pedantic = 0;
- if (options & XML_PARSE_NOBLANKS) {
- ctxt->keepBlanks = 0;
- ctxt->sax->ignorableWhitespace = xmlSAX2IgnorableWhitespace;
- options -= XML_PARSE_NOBLANKS;
- ctxt->options |= XML_PARSE_NOBLANKS;
- } else
- ctxt->keepBlanks = 1;
- if (options & XML_PARSE_DTDVALID) {
- ctxt->validate = 1;
- if (options & XML_PARSE_NOWARNING)
- ctxt->vctxt.warning = NULL;
- if (options & XML_PARSE_NOERROR)
- ctxt->vctxt.error = NULL;
- options -= XML_PARSE_DTDVALID;
- ctxt->options |= XML_PARSE_DTDVALID;
- } else
- ctxt->validate = 0;
- if (options & XML_PARSE_NOWARNING) {
- ctxt->sax->warning = NULL;
- options -= XML_PARSE_NOWARNING;
- }
- if (options & XML_PARSE_NOERROR) {
- ctxt->sax->error = NULL;
- ctxt->sax->fatalError = NULL;
- options -= XML_PARSE_NOERROR;
- }
- #ifdef LIBXML_SAX1_ENABLED
- if (options & XML_PARSE_SAX1) {
- ctxt->sax->startElement = xmlSAX2StartElement;
- ctxt->sax->endElement = xmlSAX2EndElement;
- ctxt->sax->startElementNs = NULL;
- ctxt->sax->endElementNs = NULL;
- ctxt->sax->initialized = 1;
- options -= XML_PARSE_SAX1;
- ctxt->options |= XML_PARSE_SAX1;
- }
- #endif /* LIBXML_SAX1_ENABLED */
- if (options & XML_PARSE_NODICT) {
- ctxt->dictNames = 0;
- options -= XML_PARSE_NODICT;
- ctxt->options |= XML_PARSE_NODICT;
- } else {
- ctxt->dictNames = 1;
- }
- if (options & XML_PARSE_NOCDATA) {
- ctxt->sax->cdataBlock = NULL;
- options -= XML_PARSE_NOCDATA;
- ctxt->options |= XML_PARSE_NOCDATA;
- }
- if (options & XML_PARSE_NSCLEAN) {
- ctxt->options |= XML_PARSE_NSCLEAN;
- options -= XML_PARSE_NSCLEAN;
- }
- if (options & XML_PARSE_NONET) {
- ctxt->options |= XML_PARSE_NONET;
- options -= XML_PARSE_NONET;
- }
- if (options & XML_PARSE_COMPACT) {
- ctxt->options |= XML_PARSE_COMPACT;
- options -= XML_PARSE_COMPACT;
- }
- if (options & XML_PARSE_OLD10) {
- ctxt->options |= XML_PARSE_OLD10;
- options -= XML_PARSE_OLD10;
- }
- if (options & XML_PARSE_NOBASEFIX) {
- ctxt->options |= XML_PARSE_NOBASEFIX;
- options -= XML_PARSE_NOBASEFIX;
- }
- if (options & XML_PARSE_HUGE) {
- ctxt->options |= XML_PARSE_HUGE;
- options -= XML_PARSE_HUGE;
- if (ctxt->dict != NULL)
- xmlDictSetLimit(ctxt->dict, 0);
- }
- if (options & XML_PARSE_OLDSAX) {
- ctxt->options |= XML_PARSE_OLDSAX;
- options -= XML_PARSE_OLDSAX;
- }
- if (options & XML_PARSE_IGNORE_ENC) {
- ctxt->options |= XML_PARSE_IGNORE_ENC;
- options -= XML_PARSE_IGNORE_ENC;
- }
- if (options & XML_PARSE_BIG_LINES) {
- ctxt->options |= XML_PARSE_BIG_LINES;
- options -= XML_PARSE_BIG_LINES;
- }
- ctxt->linenumbers = 1;
- return (options);
- }
- /**
- * xmlCtxtUseOptions:
- * @ctxt: an XML parser context
- * @options: a combination of xmlParserOption
- *
- * Applies the options to the parser context
- *
- * Returns 0 in case of success, the set of unknown or unimplemented options
- * in case of error.
- */
- int
- xmlCtxtUseOptions(xmlParserCtxtPtr ctxt, int options)
- {
- return(xmlCtxtUseOptionsInternal(ctxt, options, NULL));
- }
- /**
- * xmlDoRead:
- * @ctxt: an XML parser context
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- * @reuse: keep the context for reuse
- *
- * Common front-end for the xmlRead functions
- *
- * Returns the resulting document tree or NULL
- */
- static xmlDocPtr
- xmlDoRead(xmlParserCtxtPtr ctxt, const char *URL, const char *encoding,
- int options, int reuse)
- {
- xmlDocPtr ret;
- xmlCtxtUseOptionsInternal(ctxt, options, encoding);
- if (encoding != NULL) {
- xmlCharEncodingHandlerPtr hdlr;
- hdlr = xmlFindCharEncodingHandler(encoding);
- if (hdlr != NULL)
- xmlSwitchToEncoding(ctxt, hdlr);
- }
- if ((URL != NULL) && (ctxt->input != NULL) &&
- (ctxt->input->filename == NULL))
- ctxt->input->filename = (char *) xmlStrdup((const xmlChar *) URL);
- xmlParseDocument(ctxt);
- if ((ctxt->wellFormed) || ctxt->recovery)
- ret = ctxt->myDoc;
- else {
- ret = NULL;
- if (ctxt->myDoc != NULL) {
- xmlFreeDoc(ctxt->myDoc);
- }
- }
- ctxt->myDoc = NULL;
- if (!reuse) {
- xmlFreeParserCtxt(ctxt);
- }
- return (ret);
- }
- /**
- * xmlReadDoc:
- * @cur: a pointer to a zero terminated string
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadDoc(const xmlChar * cur, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- if (cur == NULL)
- return (NULL);
- xmlInitParser();
- ctxt = xmlCreateDocParserCtxt(cur);
- if (ctxt == NULL)
- return (NULL);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlReadFile:
- * @filename: a file or URL
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML file from the filesystem or the network.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadFile(const char *filename, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateURLParserCtxt(filename, options);
- if (ctxt == NULL)
- return (NULL);
- return (xmlDoRead(ctxt, NULL, encoding, options, 0));
- }
- /**
- * xmlReadMemory:
- * @buffer: a pointer to a char array
- * @size: the size of the array
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadMemory(const char *buffer, int size, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlInitParser();
- ctxt = xmlCreateMemoryParserCtxt(buffer, size);
- if (ctxt == NULL)
- return (NULL);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlReadFd:
- * @fd: an open file descriptor
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML from a file descriptor and build a tree.
- * NOTE that the file descriptor will not be closed when the
- * reader is closed or reset.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadFd(int fd, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (fd < 0)
- return (NULL);
- xmlInitParser();
- input = xmlParserInputBufferCreateFd(fd, XML_CHAR_ENCODING_NONE);
- if (input == NULL)
- return (NULL);
- input->closecallback = NULL;
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- xmlFreeParserCtxt(ctxt);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlReadIO:
- * @ioread: an I/O read function
- * @ioclose: an I/O close function
- * @ioctx: an I/O handler
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML document from I/O functions and source and build a tree.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlReadIO(xmlInputReadCallback ioread, xmlInputCloseCallback ioclose,
- void *ioctx, const char *URL, const char *encoding, int options)
- {
- xmlParserCtxtPtr ctxt;
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (ioread == NULL)
- return (NULL);
- xmlInitParser();
- input = xmlParserInputBufferCreateIO(ioread, ioclose, ioctx,
- XML_CHAR_ENCODING_NONE);
- if (input == NULL) {
- if (ioclose != NULL)
- ioclose(ioctx);
- return (NULL);
- }
- ctxt = xmlNewParserCtxt();
- if (ctxt == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- xmlFreeParserCtxt(ctxt);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 0));
- }
- /**
- * xmlCtxtReadDoc:
- * @ctxt: an XML parser context
- * @cur: a pointer to a zero terminated string
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadDoc(xmlParserCtxtPtr ctxt, const xmlChar * cur,
- const char *URL, const char *encoding, int options)
- {
- xmlParserInputPtr stream;
- if (cur == NULL)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- stream = xmlNewStringInputStream(ctxt, cur);
- if (stream == NULL) {
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadFile:
- * @ctxt: an XML parser context
- * @filename: a file or URL
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML file from the filesystem or the network.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadFile(xmlParserCtxtPtr ctxt, const char *filename,
- const char *encoding, int options)
- {
- xmlParserInputPtr stream;
- if (filename == NULL)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- stream = xmlLoadExternalEntity(filename, NULL, ctxt);
- if (stream == NULL) {
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, NULL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadMemory:
- * @ctxt: an XML parser context
- * @buffer: a pointer to a char array
- * @size: the size of the array
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML in-memory document and build a tree.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadMemory(xmlParserCtxtPtr ctxt, const char *buffer, int size,
- const char *URL, const char *encoding, int options)
- {
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (ctxt == NULL)
- return (NULL);
- if (buffer == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- input = xmlParserInputBufferCreateMem(buffer, size, XML_CHAR_ENCODING_NONE);
- if (input == NULL) {
- return(NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- return(NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadFd:
- * @ctxt: an XML parser context
- * @fd: an open file descriptor
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML from a file descriptor and build a tree.
- * This reuses the existing @ctxt parser context
- * NOTE that the file descriptor will not be closed when the
- * reader is closed or reset.
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadFd(xmlParserCtxtPtr ctxt, int fd,
- const char *URL, const char *encoding, int options)
- {
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (fd < 0)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- input = xmlParserInputBufferCreateFd(fd, XML_CHAR_ENCODING_NONE);
- if (input == NULL)
- return (NULL);
- input->closecallback = NULL;
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- /**
- * xmlCtxtReadIO:
- * @ctxt: an XML parser context
- * @ioread: an I/O read function
- * @ioclose: an I/O close function
- * @ioctx: an I/O handler
- * @URL: the base URL to use for the document
- * @encoding: the document encoding, or NULL
- * @options: a combination of xmlParserOption
- *
- * parse an XML document from I/O functions and source and build a tree.
- * This reuses the existing @ctxt parser context
- *
- * Returns the resulting document tree
- */
- xmlDocPtr
- xmlCtxtReadIO(xmlParserCtxtPtr ctxt, xmlInputReadCallback ioread,
- xmlInputCloseCallback ioclose, void *ioctx,
- const char *URL,
- const char *encoding, int options)
- {
- xmlParserInputBufferPtr input;
- xmlParserInputPtr stream;
- if (ioread == NULL)
- return (NULL);
- if (ctxt == NULL)
- return (NULL);
- xmlInitParser();
- xmlCtxtReset(ctxt);
- input = xmlParserInputBufferCreateIO(ioread, ioclose, ioctx,
- XML_CHAR_ENCODING_NONE);
- if (input == NULL) {
- if (ioclose != NULL)
- ioclose(ioctx);
- return (NULL);
- }
- stream = xmlNewIOInputStream(ctxt, input, XML_CHAR_ENCODING_NONE);
- if (stream == NULL) {
- xmlFreeParserInputBuffer(input);
- return (NULL);
- }
- inputPush(ctxt, stream);
- return (xmlDoRead(ctxt, URL, encoding, options, 1));
- }
- #define bottom_parser
- #include "elfgcchack.h"
|