diff options
author | Gregory P. Smith <greg@krypto.org> | 2012-07-14 21:12:35 (GMT) |
---|---|---|
committer | Gregory P. Smith <greg@krypto.org> | 2012-07-14 21:12:35 (GMT) |
commit | 7c6309c6afebecd9cefe7c547b87cf23abcbe2a3 (patch) | |
tree | 9852412a7bad47617cd0d42e01d37e3400229b04 /Modules/expat/xmlparse.c | |
parent | 15810673dc33436cbf8d1a6a6f9622249845b205 (diff) | |
download | cpython-7c6309c6afebecd9cefe7c547b87cf23abcbe2a3.zip cpython-7c6309c6afebecd9cefe7c547b87cf23abcbe2a3.tar.gz cpython-7c6309c6afebecd9cefe7c547b87cf23abcbe2a3.tar.bz2 |
Update the embedded copy of the expat XML parser to 2.1.0. It brings
with it a vareity of bug fixes, both security and behavior. See
http://www.libexpat.org/ for the list.
NOTE: I already backported the expat hash randomization fix in March.
Fixes issue #14340.
Diffstat (limited to 'Modules/expat/xmlparse.c')
-rw-r--r-- | Modules/expat/xmlparse.c | 423 |
1 files changed, 248 insertions, 175 deletions
diff --git a/Modules/expat/xmlparse.c b/Modules/expat/xmlparse.c index 46f6507..f35aa36 100644 --- a/Modules/expat/xmlparse.c +++ b/Modules/expat/xmlparse.c @@ -2,24 +2,27 @@ See the file COPYING for copying permission. */ +#include <stddef.h> +#include <string.h> /* memset(), memcpy() */ +#include <assert.h> +#include <limits.h> /* UINT_MAX */ +#include <time.h> /* time() */ + #define XML_BUILDING_EXPAT 1 #ifdef COMPILED_FROM_DSP #include "winconfig.h" #elif defined(MACOS_CLASSIC) #include "macconfig.h" -#elif defined(__amigaos4__) +#elif defined(__amigaos__) #include "amigaconfig.h" +#elif defined(__WATCOMC__) +#include "watcomconfig.h" #elif defined(HAVE_EXPAT_CONFIG_H) #include <expat_config.h> #endif /* ndef COMPILED_FROM_DSP */ -#include <stddef.h> -#include <string.h> /* memset(), memcpy() */ -#include <assert.h> -#include <limits.h> /* UINT_MAX */ -#include <time.h> /* time() */ - +#include "ascii.h" #include "expat.h" #ifdef XML_UNICODE @@ -28,7 +31,8 @@ #define XmlGetInternalEncoding XmlGetUtf16InternalEncoding #define XmlGetInternalEncodingNS XmlGetUtf16InternalEncodingNS #define XmlEncode XmlUtf16Encode -#define MUST_CONVERT(enc, s) (!(enc)->isUtf16 || (((unsigned long)s) & 1)) +/* Using pointer subtraction to convert to integer type. */ +#define MUST_CONVERT(enc, s) (!(enc)->isUtf16 || (((char *)(s) - (char *)NULL) & 1)) typedef unsigned short ICHAR; #else #define XML_ENCODE_MAX XML_UTF8_ENCODE_MAX @@ -325,15 +329,15 @@ processXmlDecl(XML_Parser parser, int isGeneralTextEntity, static enum XML_Error initializeEncoding(XML_Parser parser); static enum XML_Error -doProlog(XML_Parser parser, const ENCODING *enc, const char *s, - const char *end, int tok, const char *next, const char **nextPtr, +doProlog(XML_Parser parser, const ENCODING *enc, const char *s, + const char *end, int tok, const char *next, const char **nextPtr, XML_Bool haveMore); static enum XML_Error -processInternalEntity(XML_Parser parser, ENTITY *entity, +processInternalEntity(XML_Parser parser, ENTITY *entity, XML_Bool betweenDecl); static enum XML_Error doContent(XML_Parser parser, int startTagLevel, const ENCODING *enc, - const char *start, const char *end, const char **endPtr, + const char *start, const char *end, const char **endPtr, XML_Bool haveMore); static enum XML_Error doCdataSection(XML_Parser parser, const ENCODING *, const char **startPtr, @@ -351,7 +355,7 @@ static enum XML_Error addBinding(XML_Parser parser, PREFIX *prefix, const ATTRIBUTE_ID *attId, const XML_Char *uri, BINDING **bindingsPtr); static int -defineAttribute(ELEMENT_TYPE *type, ATTRIBUTE_ID *, XML_Bool isCdata, +defineAttribute(ELEMENT_TYPE *type, ATTRIBUTE_ID *, XML_Bool isCdata, XML_Bool isId, const XML_Char *dfltValue, XML_Parser parser); static enum XML_Error storeAttributeValue(XML_Parser parser, const ENCODING *, XML_Bool isCdata, @@ -436,6 +440,7 @@ parserCreate(const XML_Char *encodingName, const XML_Memory_Handling_Suite *memsuite, const XML_Char *nameSep, DTD *dtd); + static void parserInit(XML_Parser parser, const XML_Char *encodingName); @@ -535,6 +540,9 @@ struct XML_ParserStruct { NS_ATT *m_nsAtts; unsigned long m_nsAttsVersion; unsigned char m_nsAttsPower; +#ifdef XML_ATTR_INFO + XML_AttrInfo *m_attInfo; +#endif POSITION m_position; STRING_POOL m_tempPool; STRING_POOL m_temp2Pool; @@ -643,6 +651,7 @@ struct XML_ParserStruct { #define nsAtts (parser->m_nsAtts) #define nsAttsVersion (parser->m_nsAttsVersion) #define nsAttsPower (parser->m_nsAttsPower) +#define attInfo (parser->m_attInfo) #define tempPool (parser->m_tempPool) #define temp2Pool (parser->m_temp2Pool) #define groupConnector (parser->m_groupConnector) @@ -673,10 +682,12 @@ XML_ParserCreateNS(const XML_Char *encodingName, XML_Char nsSep) } static const XML_Char implicitContext[] = { - 'x', 'm', 'l', '=', 'h', 't', 't', 'p', ':', '/', '/', - 'w', 'w', 'w', '.', 'w', '3', '.', 'o', 'r', 'g', '/', - 'X', 'M', 'L', '/', '1', '9', '9', '8', '/', - 'n', 'a', 'm', 'e', 's', 'p', 'a', 'c', 'e', '\0' + ASCII_x, ASCII_m, ASCII_l, ASCII_EQUALS, ASCII_h, ASCII_t, ASCII_t, ASCII_p, + ASCII_COLON, ASCII_SLASH, ASCII_SLASH, ASCII_w, ASCII_w, ASCII_w, + ASCII_PERIOD, ASCII_w, ASCII_3, ASCII_PERIOD, ASCII_o, ASCII_r, ASCII_g, + ASCII_SLASH, ASCII_X, ASCII_M, ASCII_L, ASCII_SLASH, ASCII_1, ASCII_9, + ASCII_9, ASCII_8, ASCII_SLASH, ASCII_n, ASCII_a, ASCII_m, ASCII_e, + ASCII_s, ASCII_p, ASCII_a, ASCII_c, ASCII_e, '\0' }; static unsigned long @@ -690,23 +701,22 @@ generate_hash_secret_salt(void) static XML_Bool /* only valid for root parser */ startParsing(XML_Parser parser) { - /* hash functions must be initialized before setContext() is called */ - - if (hash_secret_salt == 0) - hash_secret_salt = generate_hash_secret_salt(); - if (ns) { - /* implicit context only set for root parser, since child - parsers (i.e. external entity parsers) will inherit it - */ - return setContext(parser, implicitContext); - } - return XML_TRUE; + /* hash functions must be initialized before setContext() is called */ + if (hash_secret_salt == 0) + hash_secret_salt = generate_hash_secret_salt(); + if (ns) { + /* implicit context only set for root parser, since child + parsers (i.e. external entity parsers) will inherit it + */ + return setContext(parser, implicitContext); + } + return XML_TRUE; } XML_Parser XMLCALL XML_ParserCreate_MM(const XML_Char *encodingName, - const XML_Memory_Handling_Suite *memsuite, - const XML_Char *nameSep) + const XML_Memory_Handling_Suite *memsuite, + const XML_Char *nameSep) { return parserCreate(encodingName, memsuite, nameSep, NULL); } @@ -753,9 +763,20 @@ parserCreate(const XML_Char *encodingName, FREE(parser); return NULL; } +#ifdef XML_ATTR_INFO + attInfo = (XML_AttrInfo*)MALLOC(attsSize * sizeof(XML_AttrInfo)); + if (attInfo == NULL) { + FREE(atts); + FREE(parser); + return NULL; + } +#endif dataBuf = (XML_Char *)MALLOC(INIT_DATA_BUF_SIZE * sizeof(XML_Char)); if (dataBuf == NULL) { FREE(atts); +#ifdef XML_ATTR_INFO + FREE(attInfo); +#endif FREE(parser); return NULL; } @@ -768,6 +789,9 @@ parserCreate(const XML_Char *encodingName, if (_dtd == NULL) { FREE(dataBuf); FREE(atts); +#ifdef XML_ATTR_INFO + FREE(attInfo); +#endif FREE(parser); return NULL; } @@ -783,7 +807,7 @@ parserCreate(const XML_Char *encodingName, unknownEncodingHandler = NULL; unknownEncodingHandlerData = NULL; - namespaceSeparator = '!'; + namespaceSeparator = ASCII_EXCL; ns = XML_FALSE; ns_triplets = XML_FALSE; @@ -1154,6 +1178,9 @@ XML_ParserFree(XML_Parser parser) #endif /* XML_DTD */ dtdDestroy(_dtd, (XML_Bool)!parentParser, &parser->m_mem); FREE((void *)atts); +#ifdef XML_ATTR_INFO + FREE((void *)attInfo); +#endif FREE(groupConnector); FREE(buffer); FREE(dataBuf); @@ -1234,6 +1261,14 @@ XML_GetIdAttributeIndex(XML_Parser parser) return idAttIndex; } +#ifdef XML_ATTR_INFO +const XML_AttrInfo * XMLCALL +XML_GetAttributeInfo(XML_Parser parser) +{ + return attInfo; +} +#endif + void XMLCALL XML_SetElementHandler(XML_Parser parser, XML_StartElementHandler start, @@ -1499,7 +1534,7 @@ XML_Parse(XML_Parser parser, const char *s, int len, int isFinal) XmlUpdatePosition(encoding, positionPtr, bufferPtr, &position); positionPtr = bufferPtr; return XML_STATUS_SUSPENDED; - case XML_INITIALIZED: + case XML_INITIALIZED: case XML_PARSING: ps_parsing = XML_FINISHED; /* fall through */ @@ -1555,15 +1590,11 @@ XML_Parse(XML_Parser parser, const char *s, int len, int isFinal) : (char *)REALLOC(buffer, len * 2)); if (temp == NULL) { errorCode = XML_ERROR_NO_MEMORY; - return XML_STATUS_ERROR; - } - buffer = temp; - if (!buffer) { - errorCode = XML_ERROR_NO_MEMORY; eventPtr = eventEndPtr = NULL; processor = errorProcessor; return XML_STATUS_ERROR; } + buffer = temp; bufferLim = buffer + len * 2; } memcpy(buffer, end, nLeftOver); @@ -1629,7 +1660,7 @@ XML_ParseBuffer(XML_Parser parser, int len, int isFinal) case XML_SUSPENDED: result = XML_STATUS_SUSPENDED; break; - case XML_INITIALIZED: + case XML_INITIALIZED: case XML_PARSING: if (isFinal) { ps_parsing = XML_FINISHED; @@ -1719,6 +1750,8 @@ XML_GetBuffer(XML_Parser parser, int len) bufferPtr = buffer = newBuf; #endif /* not defined XML_CONTEXT_BYTES */ } + eventPtr = eventEndPtr = NULL; + positionPtr = NULL; } return bufferEnd; } @@ -1776,7 +1809,7 @@ XML_ResumeParser(XML_Parser parser) case XML_SUSPENDED: result = XML_STATUS_SUSPENDED; break; - case XML_INITIALIZED: + case XML_INITIALIZED: case XML_PARSING: if (ps_finalBuffer) { ps_parsing = XML_FINISHED; @@ -2001,6 +2034,12 @@ XML_GetFeatureList(void) #ifdef XML_NS {XML_FEATURE_NS, XML_L("XML_NS"), 0}, #endif +#ifdef XML_LARGE_SIZE + {XML_FEATURE_LARGE_SIZE, XML_L("XML_LARGE_SIZE"), 0}, +#endif +#ifdef XML_ATTR_INFO + {XML_FEATURE_ATTR_INFO, XML_L("XML_ATTR_INFO"), 0}, +#endif {XML_FEATURE_END, NULL, 0} }; @@ -2063,7 +2102,7 @@ contentProcessor(XML_Parser parser, const char *end, const char **endPtr) { - enum XML_Error result = doContent(parser, 0, encoding, start, end, + enum XML_Error result = doContent(parser, 0, encoding, start, end, endPtr, (XML_Bool)!ps_finalBuffer); if (result == XML_ERROR_NONE) { if (!storeRawNames(parser)) @@ -2145,7 +2184,7 @@ externalEntityInitProcessor3(XML_Parser parser, if (result != XML_ERROR_NONE) return result; switch (ps_parsing) { - case XML_SUSPENDED: + case XML_SUSPENDED: *endPtr = next; return XML_ERROR_NONE; case XML_FINISHED: @@ -2179,7 +2218,7 @@ externalEntityContentProcessor(XML_Parser parser, const char *end, const char **endPtr) { - enum XML_Error result = doContent(parser, 1, encoding, start, end, + enum XML_Error result = doContent(parser, 1, encoding, start, end, endPtr, (XML_Bool)!ps_finalBuffer); if (result == XML_ERROR_NONE) { if (!storeRawNames(parser)) @@ -2198,7 +2237,7 @@ doContent(XML_Parser parser, XML_Bool haveMore) { /* save one level of indirection */ - DTD * const dtd = _dtd; + DTD * const dtd = _dtd; const char **eventPP; const char **eventEndPP; @@ -2229,8 +2268,8 @@ doContent(XML_Parser parser, } else if (defaultHandler) reportDefault(parser, enc, s, end); - /* We are at the end of the final buffer, should we check for - XML_SUSPENDED, XML_FINISHED? + /* We are at the end of the final buffer, should we check for + XML_SUSPENDED, XML_FINISHED? */ if (startTagLevel == 0) return XML_ERROR_NO_ELEMENTS; @@ -2581,8 +2620,8 @@ doContent(XML_Parser parser, } else if (defaultHandler) reportDefault(parser, enc, s, end); - /* We are at the end of the final buffer, should we check for - XML_SUSPENDED, XML_FINISHED? + /* We are at the end of the final buffer, should we check for + XML_SUSPENDED, XML_FINISHED? */ if (startTagLevel == 0) { *eventPP = end; @@ -2595,26 +2634,29 @@ doContent(XML_Parser parser, *nextPtr = end; return XML_ERROR_NONE; case XML_TOK_DATA_CHARS: - if (characterDataHandler) { - if (MUST_CONVERT(enc, s)) { - for (;;) { - ICHAR *dataPtr = (ICHAR *)dataBuf; - XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd); - *eventEndPP = s; - characterDataHandler(handlerArg, dataBuf, - (int)(dataPtr - (ICHAR *)dataBuf)); - if (s == next) - break; - *eventPP = s; + { + XML_CharacterDataHandler charDataHandler = characterDataHandler; + if (charDataHandler) { + if (MUST_CONVERT(enc, s)) { + for (;;) { + ICHAR *dataPtr = (ICHAR *)dataBuf; + XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd); + *eventEndPP = s; + charDataHandler(handlerArg, dataBuf, + (int)(dataPtr - (ICHAR *)dataBuf)); + if (s == next) + break; + *eventPP = s; + } } + else + charDataHandler(handlerArg, + (XML_Char *)s, + (int)((XML_Char *)next - (XML_Char *)s)); } - else - characterDataHandler(handlerArg, - (XML_Char *)s, - (int)((XML_Char *)next - (XML_Char *)s)); + else if (defaultHandler) + reportDefault(parser, enc, s, next); } - else if (defaultHandler) - reportDefault(parser, enc, s, next); break; case XML_TOK_PI: if (!reportProcessingInstruction(parser, enc, s, next)) @@ -2631,7 +2673,7 @@ doContent(XML_Parser parser, } *eventPP = s = next; switch (ps_parsing) { - case XML_SUSPENDED: + case XML_SUSPENDED: *nextPtr = next; return XML_ERROR_NONE; case XML_FINISHED: @@ -2690,23 +2732,44 @@ storeAtts(XML_Parser parser, const ENCODING *enc, if (n + nDefaultAtts > attsSize) { int oldAttsSize = attsSize; ATTRIBUTE *temp; +#ifdef XML_ATTR_INFO + XML_AttrInfo *temp2; +#endif attsSize = n + nDefaultAtts + INIT_ATTS_SIZE; temp = (ATTRIBUTE *)REALLOC((void *)atts, attsSize * sizeof(ATTRIBUTE)); if (temp == NULL) return XML_ERROR_NO_MEMORY; atts = temp; +#ifdef XML_ATTR_INFO + temp2 = (XML_AttrInfo *)REALLOC((void *)attInfo, attsSize * sizeof(XML_AttrInfo)); + if (temp2 == NULL) + return XML_ERROR_NO_MEMORY; + attInfo = temp2; +#endif if (n > oldAttsSize) XmlGetAttributes(enc, attStr, n, atts); } appAtts = (const XML_Char **)atts; for (i = 0; i < n; i++) { + ATTRIBUTE *currAtt = &atts[i]; +#ifdef XML_ATTR_INFO + XML_AttrInfo *currAttInfo = &attInfo[i]; +#endif /* add the name and value to the attribute list */ - ATTRIBUTE_ID *attId = getAttributeId(parser, enc, atts[i].name, - atts[i].name - + XmlNameLength(enc, atts[i].name)); + ATTRIBUTE_ID *attId = getAttributeId(parser, enc, currAtt->name, + currAtt->name + + XmlNameLength(enc, currAtt->name)); if (!attId) return XML_ERROR_NO_MEMORY; +#ifdef XML_ATTR_INFO + currAttInfo->nameStart = parseEndByteIndex - (parseEndPtr - currAtt->name); + currAttInfo->nameEnd = currAttInfo->nameStart + + XmlNameLength(enc, currAtt->name); + currAttInfo->valueStart = parseEndByteIndex - + (parseEndPtr - currAtt->valuePtr); + currAttInfo->valueEnd = parseEndByteIndex - (parseEndPtr - currAtt->valueEnd); +#endif /* Detect duplicate attributes by their QNames. This does not work when namespace processing is turned on and different prefixes for the same namespace are used. For this case we have a check further down. @@ -2848,8 +2911,6 @@ storeAtts(XML_Parser parser, const ENCODING *enc, unsigned long uriHash = hash_secret_salt; ((XML_Char *)s)[-1] = 0; /* clear flag */ id = (ATTRIBUTE_ID *)lookup(parser, &dtd->attributeIds, s, 0); - if (!id) - return XML_ERROR_NO_MEMORY; b = id->prefix->binding; if (!b) return XML_ERROR_UNBOUND_PREFIX; @@ -2861,7 +2922,7 @@ storeAtts(XML_Parser parser, const ENCODING *enc, return XML_ERROR_NO_MEMORY; uriHash = CHAR_HASH(uriHash, c); } - while (*s++ != XML_T(':')) + while (*s++ != XML_T(ASCII_COLON)) ; do { /* copies null terminator */ const XML_Char c = *s; @@ -2935,7 +2996,7 @@ storeAtts(XML_Parser parser, const ENCODING *enc, if (!binding) return XML_ERROR_UNBOUND_PREFIX; localPart = tagNamePtr->str; - while (*localPart++ != XML_T(':')) + while (*localPart++ != XML_T(ASCII_COLON)) ; } else if (dtd->defaultPrefix.binding) { @@ -2990,25 +3051,29 @@ addBinding(XML_Parser parser, PREFIX *prefix, const ATTRIBUTE_ID *attId, const XML_Char *uri, BINDING **bindingsPtr) { static const XML_Char xmlNamespace[] = { - 'h', 't', 't', 'p', ':', '/', '/', - 'w', 'w', 'w', '.', 'w', '3', '.', 'o', 'r', 'g', '/', - 'X', 'M', 'L', '/', '1', '9', '9', '8', '/', - 'n', 'a', 'm', 'e', 's', 'p', 'a', 'c', 'e', '\0' + ASCII_h, ASCII_t, ASCII_t, ASCII_p, ASCII_COLON, ASCII_SLASH, ASCII_SLASH, + ASCII_w, ASCII_w, ASCII_w, ASCII_PERIOD, ASCII_w, ASCII_3, ASCII_PERIOD, + ASCII_o, ASCII_r, ASCII_g, ASCII_SLASH, ASCII_X, ASCII_M, ASCII_L, + ASCII_SLASH, ASCII_1, ASCII_9, ASCII_9, ASCII_8, ASCII_SLASH, + ASCII_n, ASCII_a, ASCII_m, ASCII_e, ASCII_s, ASCII_p, ASCII_a, ASCII_c, + ASCII_e, '\0' }; - static const int xmlLen = + static const int xmlLen = (int)sizeof(xmlNamespace)/sizeof(XML_Char) - 1; static const XML_Char xmlnsNamespace[] = { - 'h', 't', 't', 'p', ':', '/', '/', - 'w', 'w', 'w', '.', 'w', '3', '.', 'o', 'r', 'g', '/', - '2', '0', '0', '0', '/', 'x', 'm', 'l', 'n', 's', '/', '\0' + ASCII_h, ASCII_t, ASCII_t, ASCII_p, ASCII_COLON, ASCII_SLASH, ASCII_SLASH, + ASCII_w, ASCII_w, ASCII_w, ASCII_PERIOD, ASCII_w, ASCII_3, ASCII_PERIOD, + ASCII_o, ASCII_r, ASCII_g, ASCII_SLASH, ASCII_2, ASCII_0, ASCII_0, + ASCII_0, ASCII_SLASH, ASCII_x, ASCII_m, ASCII_l, ASCII_n, ASCII_s, + ASCII_SLASH, '\0' }; - static const int xmlnsLen = + static const int xmlnsLen = (int)sizeof(xmlnsNamespace)/sizeof(XML_Char) - 1; XML_Bool mustBeXML = XML_FALSE; XML_Bool isXML = XML_TRUE; XML_Bool isXMLNS = XML_TRUE; - + BINDING *b; int len; @@ -3017,13 +3082,13 @@ addBinding(XML_Parser parser, PREFIX *prefix, const ATTRIBUTE_ID *attId, return XML_ERROR_UNDECLARING_PREFIX; if (prefix->name - && prefix->name[0] == XML_T('x') - && prefix->name[1] == XML_T('m') - && prefix->name[2] == XML_T('l')) { + && prefix->name[0] == XML_T(ASCII_x) + && prefix->name[1] == XML_T(ASCII_m) + && prefix->name[2] == XML_T(ASCII_l)) { /* Not allowed to bind xmlns */ - if (prefix->name[3] == XML_T('n') - && prefix->name[4] == XML_T('s') + if (prefix->name[3] == XML_T(ASCII_n) + && prefix->name[4] == XML_T(ASCII_s) && prefix->name[5] == XML_T('\0')) return XML_ERROR_RESERVED_PREFIX_XMLNS; @@ -3035,7 +3100,7 @@ addBinding(XML_Parser parser, PREFIX *prefix, const ATTRIBUTE_ID *attId, if (isXML && (len > xmlLen || uri[len] != xmlNamespace[len])) isXML = XML_FALSE; - if (!mustBeXML && isXMLNS + if (!mustBeXML && isXMLNS && (len > xmlnsLen || uri[len] != xmlnsNamespace[len])) isXMLNS = XML_FALSE; } @@ -3177,26 +3242,29 @@ doCdataSection(XML_Parser parser, reportDefault(parser, enc, s, next); break; case XML_TOK_DATA_CHARS: - if (characterDataHandler) { - if (MUST_CONVERT(enc, s)) { - for (;;) { - ICHAR *dataPtr = (ICHAR *)dataBuf; - XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd); - *eventEndPP = next; - characterDataHandler(handlerArg, dataBuf, - (int)(dataPtr - (ICHAR *)dataBuf)); - if (s == next) - break; - *eventPP = s; + { + XML_CharacterDataHandler charDataHandler = characterDataHandler; + if (charDataHandler) { + if (MUST_CONVERT(enc, s)) { + for (;;) { + ICHAR *dataPtr = (ICHAR *)dataBuf; + XmlConvert(enc, &s, next, &dataPtr, (ICHAR *)dataBufEnd); + *eventEndPP = next; + charDataHandler(handlerArg, dataBuf, + (int)(dataPtr - (ICHAR *)dataBuf)); + if (s == next) + break; + *eventPP = s; + } } + else + charDataHandler(handlerArg, + (XML_Char *)s, + (int)((XML_Char *)next - (XML_Char *)s)); } - else - characterDataHandler(handlerArg, - (XML_Char *)s, - (int)((XML_Char *)next - (XML_Char *)s)); + else if (defaultHandler) + reportDefault(parser, enc, s, next); } - else if (defaultHandler) - reportDefault(parser, enc, s, next); break; case XML_TOK_INVALID: *eventPP = next; @@ -3243,7 +3311,7 @@ ignoreSectionProcessor(XML_Parser parser, const char *end, const char **endPtr) { - enum XML_Error result = doIgnoreSection(parser, encoding, &start, end, + enum XML_Error result = doIgnoreSection(parser, encoding, &start, end, endPtr, (XML_Bool)!ps_finalBuffer); if (result != XML_ERROR_NONE) return result; @@ -3525,7 +3593,7 @@ entityValueInitProcessor(XML_Parser parser, const char *next = start; eventPtr = start; - for (;;) { + for (;;) { tok = XmlPrologTok(encoding, start, end, &next); eventEndPtr = next; if (tok <= 0) { @@ -3553,7 +3621,7 @@ entityValueInitProcessor(XML_Parser parser, if (result != XML_ERROR_NONE) return result; switch (ps_parsing) { - case XML_SUSPENDED: + case XML_SUSPENDED: *nextPtr = next; return XML_ERROR_NONE; case XML_FINISHED: @@ -3618,7 +3686,7 @@ externalParEntProcessor(XML_Parser parser, } processor = prologProcessor; - return doProlog(parser, encoding, s, end, tok, next, + return doProlog(parser, encoding, s, end, tok, next, nextPtr, (XML_Bool)!ps_finalBuffer); } @@ -3668,7 +3736,7 @@ prologProcessor(XML_Parser parser, { const char *next = s; int tok = XmlPrologTok(encoding, s, end, &next); - return doProlog(parser, encoding, s, end, tok, next, + return doProlog(parser, encoding, s, end, tok, next, nextPtr, (XML_Bool)!ps_finalBuffer); } @@ -3683,26 +3751,30 @@ doProlog(XML_Parser parser, XML_Bool haveMore) { #ifdef XML_DTD - static const XML_Char externalSubsetName[] = { '#' , '\0' }; + static const XML_Char externalSubsetName[] = { ASCII_HASH , '\0' }; #endif /* XML_DTD */ - static const XML_Char atypeCDATA[] = { 'C', 'D', 'A', 'T', 'A', '\0' }; - static const XML_Char atypeID[] = { 'I', 'D', '\0' }; - static const XML_Char atypeIDREF[] = { 'I', 'D', 'R', 'E', 'F', '\0' }; - static const XML_Char atypeIDREFS[] = { 'I', 'D', 'R', 'E', 'F', 'S', '\0' }; - static const XML_Char atypeENTITY[] = { 'E', 'N', 'T', 'I', 'T', 'Y', '\0' }; - static const XML_Char atypeENTITIES[] = - { 'E', 'N', 'T', 'I', 'T', 'I', 'E', 'S', '\0' }; + static const XML_Char atypeCDATA[] = + { ASCII_C, ASCII_D, ASCII_A, ASCII_T, ASCII_A, '\0' }; + static const XML_Char atypeID[] = { ASCII_I, ASCII_D, '\0' }; + static const XML_Char atypeIDREF[] = + { ASCII_I, ASCII_D, ASCII_R, ASCII_E, ASCII_F, '\0' }; + static const XML_Char atypeIDREFS[] = + { ASCII_I, ASCII_D, ASCII_R, ASCII_E, ASCII_F, ASCII_S, '\0' }; + static const XML_Char atypeENTITY[] = + { ASCII_E, ASCII_N, ASCII_T, ASCII_I, ASCII_T, ASCII_Y, '\0' }; + static const XML_Char atypeENTITIES[] = { ASCII_E, ASCII_N, + ASCII_T, ASCII_I, ASCII_T, ASCII_I, ASCII_E, ASCII_S, '\0' }; static const XML_Char atypeNMTOKEN[] = { - 'N', 'M', 'T', 'O', 'K', 'E', 'N', '\0' }; - static const XML_Char atypeNMTOKENS[] = { - 'N', 'M', 'T', 'O', 'K', 'E', 'N', 'S', '\0' }; - static const XML_Char notationPrefix[] = { - 'N', 'O', 'T', 'A', 'T', 'I', 'O', 'N', '(', '\0' }; - static const XML_Char enumValueSep[] = { '|', '\0' }; - static const XML_Char enumValueStart[] = { '(', '\0' }; + ASCII_N, ASCII_M, ASCII_T, ASCII_O, ASCII_K, ASCII_E, ASCII_N, '\0' }; + static const XML_Char atypeNMTOKENS[] = { ASCII_N, ASCII_M, ASCII_T, + ASCII_O, ASCII_K, ASCII_E, ASCII_N, ASCII_S, '\0' }; + static const XML_Char notationPrefix[] = { ASCII_N, ASCII_O, ASCII_T, + ASCII_A, ASCII_T, ASCII_I, ASCII_O, ASCII_N, ASCII_LPAREN, '\0' }; + static const XML_Char enumValueSep[] = { ASCII_PIPE, '\0' }; + static const XML_Char enumValueStart[] = { ASCII_LPAREN, '\0' }; /* save one level of indirection */ - DTD * const dtd = _dtd; + DTD * const dtd = _dtd; const char **eventPP; const char **eventEndPP; @@ -3818,15 +3890,17 @@ doProlog(XML_Parser parser, #endif /* XML_DTD */ dtd->hasParamEntityRefs = XML_TRUE; if (startDoctypeDeclHandler) { + XML_Char *pubId; if (!XmlIsPublicId(enc, s, next, eventPP)) return XML_ERROR_PUBLICID; - doctypePubid = poolStoreString(&tempPool, enc, - s + enc->minBytesPerChar, - next - enc->minBytesPerChar); - if (!doctypePubid) + pubId = poolStoreString(&tempPool, enc, + s + enc->minBytesPerChar, + next - enc->minBytesPerChar); + if (!pubId) return XML_ERROR_NO_MEMORY; - normalizePublicId((XML_Char *)doctypePubid); + normalizePublicId(pubId); poolFinish(&tempPool); + doctypePubid = pubId; handleDefault = XML_FALSE; goto alreadyChecked; } @@ -3881,8 +3955,8 @@ doProlog(XML_Parser parser, entity->publicId)) return XML_ERROR_EXTERNAL_ENTITY_HANDLING; if (dtd->paramEntityRead) { - if (!dtd->standalone && - notStandaloneHandler && + if (!dtd->standalone && + notStandaloneHandler && !notStandaloneHandler(handlerArg)) return XML_ERROR_NOT_STANDALONE; } @@ -4010,11 +4084,11 @@ doProlog(XML_Parser parser, 0, parser)) return XML_ERROR_NO_MEMORY; if (attlistDeclHandler && declAttributeType) { - if (*declAttributeType == XML_T('(') - || (*declAttributeType == XML_T('N') - && declAttributeType[1] == XML_T('O'))) { + if (*declAttributeType == XML_T(ASCII_LPAREN) + || (*declAttributeType == XML_T(ASCII_N) + && declAttributeType[1] == XML_T(ASCII_O))) { /* Enumerated or Notation type */ - if (!poolAppendChar(&tempPool, XML_T(')')) + if (!poolAppendChar(&tempPool, XML_T(ASCII_RPAREN)) || !poolAppendChar(&tempPool, XML_T('\0'))) return XML_ERROR_NO_MEMORY; declAttributeType = tempPool.start; @@ -4047,11 +4121,11 @@ doProlog(XML_Parser parser, declAttributeIsCdata, XML_FALSE, attVal, parser)) return XML_ERROR_NO_MEMORY; if (attlistDeclHandler && declAttributeType) { - if (*declAttributeType == XML_T('(') - || (*declAttributeType == XML_T('N') - && declAttributeType[1] == XML_T('O'))) { + if (*declAttributeType == XML_T(ASCII_LPAREN) + || (*declAttributeType == XML_T(ASCII_N) + && declAttributeType[1] == XML_T(ASCII_O))) { /* Enumerated or Notation type */ - if (!poolAppendChar(&tempPool, XML_T(')')) + if (!poolAppendChar(&tempPool, XML_T(ASCII_RPAREN)) || !poolAppendChar(&tempPool, XML_T('\0'))) return XML_ERROR_NO_MEMORY; declAttributeType = tempPool.start; @@ -4321,7 +4395,7 @@ doProlog(XML_Parser parser, switch (tok) { case XML_TOK_PARAM_ENTITY_REF: /* PE references in internal subset are - not allowed within declarations. */ + not allowed within declarations. */ return XML_ERROR_PARAM_ENTITY_REF; case XML_TOK_XML_DECL: return XML_ERROR_MISPLACED_XML_PI; @@ -4379,14 +4453,14 @@ doProlog(XML_Parser parser, } break; case XML_ROLE_GROUP_SEQUENCE: - if (groupConnector[prologState.level] == '|') + if (groupConnector[prologState.level] == ASCII_PIPE) return XML_ERROR_SYNTAX; - groupConnector[prologState.level] = ','; + groupConnector[prologState.level] = ASCII_COMMA; if (dtd->in_eldecl && elementDeclHandler) handleDefault = XML_FALSE; break; case XML_ROLE_GROUP_CHOICE: - if (groupConnector[prologState.level] == ',') + if (groupConnector[prologState.level] == ASCII_COMMA) return XML_ERROR_SYNTAX; if (dtd->in_eldecl && !groupConnector[prologState.level] @@ -4398,7 +4472,7 @@ doProlog(XML_Parser parser, if (elementDeclHandler) handleDefault = XML_FALSE; } - groupConnector[prologState.level] = '|'; + groupConnector[prologState.level] = ASCII_PIPE; break; case XML_ROLE_PARAM_ENTITY_REF: #ifdef XML_DTD @@ -4442,7 +4516,7 @@ doProlog(XML_Parser parser, return XML_ERROR_RECURSIVE_ENTITY_REF; if (entity->textPtr) { enum XML_Error result; - XML_Bool betweenDecl = + XML_Bool betweenDecl = (role == XML_ROLE_PARAM_ENTITY_REF ? XML_TRUE : XML_FALSE); result = processInternalEntity(parser, entity, betweenDecl); if (result != XML_ERROR_NONE) @@ -4637,7 +4711,7 @@ doProlog(XML_Parser parser, reportDefault(parser, enc, s, next); switch (ps_parsing) { - case XML_SUSPENDED: + case XML_SUSPENDED: *nextPtr = next; return XML_ERROR_NONE; case XML_FINISHED: @@ -4707,7 +4781,7 @@ epilogProcessor(XML_Parser parser, } eventPtr = s = next; switch (ps_parsing) { - case XML_SUSPENDED: + case XML_SUSPENDED: *nextPtr = next; return XML_ERROR_NONE; case XML_FINISHED: @@ -4750,12 +4824,12 @@ processInternalEntity(XML_Parser parser, ENTITY *entity, #ifdef XML_DTD if (entity->is_param) { int tok = XmlPrologTok(internalEncoding, textStart, textEnd, &next); - result = doProlog(parser, internalEncoding, textStart, textEnd, tok, + result = doProlog(parser, internalEncoding, textStart, textEnd, tok, next, &next, XML_FALSE); } - else + else #endif /* XML_DTD */ - result = doContent(parser, tagLevel, internalEncoding, textStart, + result = doContent(parser, tagLevel, internalEncoding, textStart, textEnd, &next, XML_FALSE); if (result == XML_ERROR_NONE) { @@ -4795,13 +4869,13 @@ internalEntityProcessor(XML_Parser parser, #ifdef XML_DTD if (entity->is_param) { int tok = XmlPrologTok(internalEncoding, textStart, textEnd, &next); - result = doProlog(parser, internalEncoding, textStart, textEnd, tok, + result = doProlog(parser, internalEncoding, textStart, textEnd, tok, next, &next, XML_FALSE); } else #endif /* XML_DTD */ - result = doContent(parser, openEntity->startTagLevel, internalEncoding, - textStart, textEnd, &next, XML_FALSE); + result = doContent(parser, openEntity->startTagLevel, internalEncoding, + textStart, textEnd, &next, XML_FALSE); if (result != XML_ERROR_NONE) return result; @@ -4822,7 +4896,7 @@ internalEntityProcessor(XML_Parser parser, int tok; processor = prologProcessor; tok = XmlPrologTok(encoding, s, end, &next); - return doProlog(parser, encoding, s, end, tok, next, nextPtr, + return doProlog(parser, encoding, s, end, tok, next, nextPtr, (XML_Bool)!ps_finalBuffer); } else @@ -4831,8 +4905,8 @@ internalEntityProcessor(XML_Parser parser, processor = contentProcessor; /* see externalEntityContentProcessor vs contentProcessor */ return doContent(parser, parentParser ? 1 : 0, encoding, s, end, - nextPtr, (XML_Bool)!ps_finalBuffer); - } + nextPtr, (XML_Bool)!ps_finalBuffer); + } } static enum XML_Error PTRCALL @@ -4985,7 +5059,7 @@ appendAttributeValue(XML_Parser parser, const ENCODING *enc, XML_Bool isCdata, if (!entity->textPtr) { if (enc == encoding) eventPtr = ptr; - return XML_ERROR_ATTRIBUTE_EXTERNAL_ENTITY_REF; + return XML_ERROR_ATTRIBUTE_EXTERNAL_ENTITY_REF; } else { enum XML_Error result; @@ -5328,7 +5402,7 @@ setElementTypePrefix(XML_Parser parser, ELEMENT_TYPE *elementType) DTD * const dtd = _dtd; /* save one level of indirection */ const XML_Char *name; for (name = elementType->name; *name; name++) { - if (*name == XML_T(':')) { + if (*name == XML_T(ASCII_COLON)) { PREFIX *prefix; const XML_Char *s; for (s = elementType->name; s != name; s++) { @@ -5375,12 +5449,12 @@ getAttributeId(XML_Parser parser, const ENCODING *enc, poolFinish(&dtd->pool); if (!ns) ; - else if (name[0] == XML_T('x') - && name[1] == XML_T('m') - && name[2] == XML_T('l') - && name[3] == XML_T('n') - && name[4] == XML_T('s') - && (name[5] == XML_T('\0') || name[5] == XML_T(':'))) { + else if (name[0] == XML_T(ASCII_x) + && name[1] == XML_T(ASCII_m) + && name[2] == XML_T(ASCII_l) + && name[3] == XML_T(ASCII_n) + && name[4] == XML_T(ASCII_s) + && (name[5] == XML_T('\0') || name[5] == XML_T(ASCII_COLON))) { if (name[5] == XML_T('\0')) id->prefix = &dtd->defaultPrefix; else @@ -5391,7 +5465,7 @@ getAttributeId(XML_Parser parser, const ENCODING *enc, int i; for (i = 0; name[i]; i++) { /* attributes without prefix are *not* in the default namespace */ - if (name[i] == XML_T(':')) { + if (name[i] == XML_T(ASCII_COLON)) { int j; for (j = 0; j < i; j++) { if (!poolAppendChar(&dtd->pool, name[j])) @@ -5401,8 +5475,6 @@ getAttributeId(XML_Parser parser, const ENCODING *enc, return NULL; id->prefix = (PREFIX *)lookup(parser, &dtd->prefixes, poolStart(&dtd->pool), sizeof(PREFIX)); - if (!id->prefix) - return NULL; if (id->prefix->name == poolStart(&dtd->pool)) poolFinish(&dtd->pool); else @@ -5415,7 +5487,7 @@ getAttributeId(XML_Parser parser, const ENCODING *enc, return id; } -#define CONTEXT_SEP XML_T('\f') +#define CONTEXT_SEP XML_T(ASCII_FF) static const XML_Char * getContext(XML_Parser parser) @@ -5427,7 +5499,7 @@ getContext(XML_Parser parser) if (dtd->defaultPrefix.binding) { int i; int len; - if (!poolAppendChar(&tempPool, XML_T('='))) + if (!poolAppendChar(&tempPool, XML_T(ASCII_EQUALS))) return NULL; len = dtd->defaultPrefix.binding->uriLen; if (namespaceSeparator) @@ -5453,7 +5525,7 @@ getContext(XML_Parser parser) for (s = prefix->name; *s; s++) if (!poolAppendChar(&tempPool, *s)) return NULL; - if (!poolAppendChar(&tempPool, XML_T('='))) + if (!poolAppendChar(&tempPool, XML_T(ASCII_EQUALS))) return NULL; len = prefix->binding->uriLen; if (namespaceSeparator) @@ -5505,7 +5577,7 @@ setContext(XML_Parser parser, const XML_Char *context) context = s; poolDiscard(&tempPool); } - else if (*s == XML_T('=')) { + else if (*s == XML_T(ASCII_EQUALS)) { PREFIX *prefix; if (poolLength(&tempPool) == 0) prefix = &dtd->defaultPrefix; @@ -6162,12 +6234,13 @@ poolGrow(STRING_POOL *pool) } if (pool->blocks && pool->start == pool->blocks->s) { int blockSize = (int)(pool->end - pool->start)*2; - pool->blocks = (BLOCK *) + BLOCK *temp = (BLOCK *) pool->mem->realloc_fcn(pool->blocks, (offsetof(BLOCK, s) + blockSize * sizeof(XML_Char))); - if (pool->blocks == NULL) + if (temp == NULL) return XML_FALSE; + pool->blocks = temp; pool->blocks->size = blockSize; pool->ptr = pool->blocks->s + (pool->ptr - pool->start); pool->start = pool->blocks->s; |