diff options
Diffstat (limited to 'Modules/expat/xmltok.h')
| -rw-r--r-- | Modules/expat/xmltok.h | 260 |
1 files changed, 138 insertions, 122 deletions
diff --git a/Modules/expat/xmltok.h b/Modules/expat/xmltok.h index 8b02324..3d776be 100644 --- a/Modules/expat/xmltok.h +++ b/Modules/expat/xmltok.h @@ -1,6 +1,5 @@ -/* -Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd -See the file COPYING for copying permission. +/* Copyright (c) 1998, 1999 Thai Open Source Software Center Ltd + See the file COPYING for copying permission. */ #ifndef XmlTok_INCLUDED @@ -11,19 +10,21 @@ extern "C" { #endif /* The following token may be returned by XmlContentTok */ -#define XML_TOK_TRAILING_RSQB -5 /* ] or ]] at the end of the scan; might be start of - illegal ]]> sequence */ -/* The following tokens may be returned by both XmlPrologTok and XmlContentTok */ -#define XML_TOK_NONE -4 /* The string to be scanned is empty */ -#define XML_TOK_TRAILING_CR -3 /* A CR at the end of the scan; - might be part of CRLF sequence */ -#define XML_TOK_PARTIAL_CHAR -2 /* only part of a multibyte sequence */ -#define XML_TOK_PARTIAL -1 /* only part of a token */ +#define XML_TOK_TRAILING_RSQB -5 /* ] or ]] at the end of the scan; might be + start of illegal ]]> sequence */ +/* The following tokens may be returned by both XmlPrologTok and + XmlContentTok. +*/ +#define XML_TOK_NONE -4 /* The string to be scanned is empty */ +#define XML_TOK_TRAILING_CR -3 /* A CR at the end of the scan; + might be part of CRLF sequence */ +#define XML_TOK_PARTIAL_CHAR -2 /* only part of a multibyte sequence */ +#define XML_TOK_PARTIAL -1 /* only part of a token */ #define XML_TOK_INVALID 0 /* The following tokens are returned by XmlContentTok; some are also - returned by XmlAttributeValueTok, XmlEntityTok, XmlCdataSectionTok */ - + returned by XmlAttributeValueTok, XmlEntityTok, XmlCdataSectionTok. +*/ #define XML_TOK_START_TAG_WITH_ATTS 1 #define XML_TOK_START_TAG_NO_ATTS 2 #define XML_TOK_EMPTY_ELEMENT_WITH_ATTS 3 /* empty element tag <e/> */ @@ -33,22 +34,24 @@ extern "C" { #define XML_TOK_DATA_NEWLINE 7 #define XML_TOK_CDATA_SECT_OPEN 8 #define XML_TOK_ENTITY_REF 9 -#define XML_TOK_CHAR_REF 10 /* numeric character reference */ +#define XML_TOK_CHAR_REF 10 /* numeric character reference */ -/* The following tokens may be returned by both XmlPrologTok and XmlContentTok */ -#define XML_TOK_PI 11 /* processing instruction */ -#define XML_TOK_XML_DECL 12 /* XML decl or text decl */ +/* The following tokens may be returned by both XmlPrologTok and + XmlContentTok. +*/ +#define XML_TOK_PI 11 /* processing instruction */ +#define XML_TOK_XML_DECL 12 /* XML decl or text decl */ #define XML_TOK_COMMENT 13 -#define XML_TOK_BOM 14 /* Byte order mark */ +#define XML_TOK_BOM 14 /* Byte order mark */ /* The following tokens are returned only by XmlPrologTok */ #define XML_TOK_PROLOG_S 15 -#define XML_TOK_DECL_OPEN 16 /* <!foo */ -#define XML_TOK_DECL_CLOSE 17 /* > */ +#define XML_TOK_DECL_OPEN 16 /* <!foo */ +#define XML_TOK_DECL_CLOSE 17 /* > */ #define XML_TOK_NAME 18 #define XML_TOK_NMTOKEN 19 -#define XML_TOK_POUND_NAME 20 /* #name */ -#define XML_TOK_OR 21 /* | */ +#define XML_TOK_POUND_NAME 20 /* #name */ +#define XML_TOK_OR 21 /* | */ #define XML_TOK_PERCENT 22 #define XML_TOK_OPEN_PAREN 23 #define XML_TOK_CLOSE_PAREN 24 @@ -59,14 +62,14 @@ extern "C" { #define XML_TOK_INSTANCE_START 29 /* The following occur only in element type declarations */ -#define XML_TOK_NAME_QUESTION 30 /* name? */ -#define XML_TOK_NAME_ASTERISK 31 /* name* */ -#define XML_TOK_NAME_PLUS 32 /* name+ */ -#define XML_TOK_COND_SECT_OPEN 33 /* <![ */ -#define XML_TOK_COND_SECT_CLOSE 34 /* ]]> */ -#define XML_TOK_CLOSE_PAREN_QUESTION 35 /* )? */ -#define XML_TOK_CLOSE_PAREN_ASTERISK 36 /* )* */ -#define XML_TOK_CLOSE_PAREN_PLUS 37 /* )+ */ +#define XML_TOK_NAME_QUESTION 30 /* name? */ +#define XML_TOK_NAME_ASTERISK 31 /* name* */ +#define XML_TOK_NAME_PLUS 32 /* name+ */ +#define XML_TOK_COND_SECT_OPEN 33 /* <![ */ +#define XML_TOK_COND_SECT_CLOSE 34 /* ]]> */ +#define XML_TOK_CLOSE_PAREN_QUESTION 35 /* )? */ +#define XML_TOK_CLOSE_PAREN_ASTERISK 36 /* )* */ +#define XML_TOK_CLOSE_PAREN_PLUS 37 /* )+ */ #define XML_TOK_COMMA 38 /* The following token is returned only by XmlAttributeValueTok */ @@ -75,8 +78,9 @@ extern "C" { /* The following token is returned only by XmlCdataSectionTok */ #define XML_TOK_CDATA_SECT_CLOSE 40 -/* With namespace processing this is returned by XmlPrologTok - for a name with a colon. */ +/* With namespace processing this is returned by XmlPrologTok for a + name with a colon. +*/ #define XML_TOK_PREFIXED_NAME 41 #ifdef XML_DTD @@ -121,64 +125,73 @@ typedef struct { struct encoding; typedef struct encoding ENCODING; +typedef int (PTRCALL *SCANNER)(const ENCODING *, + const char *, + const char *, + const char **); + struct encoding { - int (*scanners[XML_N_STATES])(const ENCODING *, - const char *, - const char *, - const char **); - int (*literalScanners[XML_N_LITERAL_TYPES])(const ENCODING *, - const char *, - const char *, - const char **); - int (*sameName)(const ENCODING *, - const char *, const char *); - int (*nameMatchesAscii)(const ENCODING *, - const char *, const char *, const char *); - int (*nameLength)(const ENCODING *, const char *); - const char *(*skipS)(const ENCODING *, const char *); - int (*getAtts)(const ENCODING *enc, const char *ptr, - int attsMax, ATTRIBUTE *atts); - int (*charRefNumber)(const ENCODING *enc, const char *ptr); - int (*predefinedEntityName)(const ENCODING *, const char *, const char *); - void (*updatePosition)(const ENCODING *, - const char *ptr, - const char *end, - POSITION *); - int (*isPublicId)(const ENCODING *enc, const char *ptr, const char *end, - const char **badPtr); - void (*utf8Convert)(const ENCODING *enc, - const char **fromP, - const char *fromLim, - char **toP, - const char *toLim); - void (*utf16Convert)(const ENCODING *enc, - const char **fromP, - const char *fromLim, - unsigned short **toP, - const unsigned short *toLim); + SCANNER scanners[XML_N_STATES]; + SCANNER literalScanners[XML_N_LITERAL_TYPES]; + int (PTRCALL *sameName)(const ENCODING *, + const char *, + const char *); + int (PTRCALL *nameMatchesAscii)(const ENCODING *, + const char *, + const char *, + const char *); + int (PTRFASTCALL *nameLength)(const ENCODING *, const char *); + const char *(PTRFASTCALL *skipS)(const ENCODING *, const char *); + int (PTRCALL *getAtts)(const ENCODING *enc, + const char *ptr, + int attsMax, + ATTRIBUTE *atts); + int (PTRFASTCALL *charRefNumber)(const ENCODING *enc, const char *ptr); + int (PTRCALL *predefinedEntityName)(const ENCODING *, + const char *, + const char *); + void (PTRCALL *updatePosition)(const ENCODING *, + const char *ptr, + const char *end, + POSITION *); + int (PTRCALL *isPublicId)(const ENCODING *enc, + const char *ptr, + const char *end, + const char **badPtr); + void (PTRCALL *utf8Convert)(const ENCODING *enc, + const char **fromP, + const char *fromLim, + char **toP, + const char *toLim); + void (PTRCALL *utf16Convert)(const ENCODING *enc, + const char **fromP, + const char *fromLim, + unsigned short **toP, + const unsigned short *toLim); int minBytesPerChar; char isUtf8; char isUtf16; }; -/* -Scan the string starting at ptr until the end of the next complete token, -but do not scan past eptr. Return an integer giving the type of token. +/* Scan the string starting at ptr until the end of the next complete + token, but do not scan past eptr. Return an integer giving the + type of token. -Return XML_TOK_NONE when ptr == eptr; nextTokPtr will not be set. + Return XML_TOK_NONE when ptr == eptr; nextTokPtr will not be set. -Return XML_TOK_PARTIAL when the string does not contain a complete token; -nextTokPtr will not be set. + Return XML_TOK_PARTIAL when the string does not contain a complete + token; nextTokPtr will not be set. -Return XML_TOK_INVALID when the string does not start a valid token; nextTokPtr -will be set to point to the character which made the token invalid. + Return XML_TOK_INVALID when the string does not start a valid + token; nextTokPtr will be set to point to the character which made + the token invalid. -Otherwise the string starts with a valid token; nextTokPtr will be set to point -to the character following the end of that token. + Otherwise the string starts with a valid token; nextTokPtr will be + set to point to the character following the end of that token. -Each data character counts as a single token, but adjacent data characters -may be returned together. Similarly for characters in the prolog outside -literals, comments and processing instructions. + Each data character counts as a single token, but adjacent data + characters may be returned together. Similarly for characters in + the prolog outside literals, comments and processing instructions. */ @@ -201,9 +214,9 @@ literals, comments and processing instructions. #endif /* XML_DTD */ -/* This is used for performing a 2nd-level tokenization on -the content of a literal that has already been returned by XmlTok. */ - +/* This is used for performing a 2nd-level tokenization on the content + of a literal that has already been returned by XmlTok. +*/ #define XmlLiteralTok(enc, literalType, ptr, end, nextTokPtr) \ (((enc)->literalScanners[literalType])(enc, ptr, end, nextTokPtr)) @@ -250,48 +263,51 @@ typedef struct { const ENCODING **encPtr; } INIT_ENCODING; -int XmlParseXmlDecl(int isGeneralTextEntity, - const ENCODING *enc, - const char *ptr, - const char *end, - const char **badPtr, - const char **versionPtr, - const char **versionEndPtr, - const char **encodingNamePtr, - const ENCODING **namedEncodingPtr, - int *standalonePtr); - -int XmlInitEncoding(INIT_ENCODING *, const ENCODING **, const char *name); -const ENCODING *XmlGetUtf8InternalEncoding(void); -const ENCODING *XmlGetUtf16InternalEncoding(void); -int XmlUtf8Encode(int charNumber, char *buf); -int XmlUtf16Encode(int charNumber, unsigned short *buf); - -int XmlSizeOfUnknownEncoding(void); -ENCODING * +int XmlParseXmlDecl(int isGeneralTextEntity, + const ENCODING *enc, + const char *ptr, + const char *end, + const char **badPtr, + const char **versionPtr, + const char **versionEndPtr, + const char **encodingNamePtr, + const ENCODING **namedEncodingPtr, + int *standalonePtr); + +int XmlInitEncoding(INIT_ENCODING *, const ENCODING **, const char *name); +const ENCODING *XmlGetUtf8InternalEncoding(void); +const ENCODING *XmlGetUtf16InternalEncoding(void); +int FASTCALL XmlUtf8Encode(int charNumber, char *buf); +int FASTCALL XmlUtf16Encode(int charNumber, unsigned short *buf); +int XmlSizeOfUnknownEncoding(void); + +typedef int (*CONVERTER)(void *userData, const char *p); + +ENCODING * XmlInitUnknownEncoding(void *mem, - int *table, - int (*conv)(void *userData, const char *p), - void *userData); - -int XmlParseXmlDeclNS(int isGeneralTextEntity, - const ENCODING *enc, - const char *ptr, - const char *end, - const char **badPtr, - const char **versionPtr, - const char **versionEndPtr, - const char **encodingNamePtr, - const ENCODING **namedEncodingPtr, - int *standalonePtr); -int XmlInitEncodingNS(INIT_ENCODING *, const ENCODING **, const char *name); -const ENCODING *XmlGetUtf8InternalEncodingNS(void); -const ENCODING *XmlGetUtf16InternalEncodingNS(void); -ENCODING * + int *table, + CONVERTER convert, + void *userData); + +int XmlParseXmlDeclNS(int isGeneralTextEntity, + const ENCODING *enc, + const char *ptr, + const char *end, + const char **badPtr, + const char **versionPtr, + const char **versionEndPtr, + const char **encodingNamePtr, + const ENCODING **namedEncodingPtr, + int *standalonePtr); + +int XmlInitEncodingNS(INIT_ENCODING *, const ENCODING **, const char *name); +const ENCODING *XmlGetUtf8InternalEncodingNS(void); +const ENCODING *XmlGetUtf16InternalEncodingNS(void); +ENCODING * XmlInitUnknownEncodingNS(void *mem, - int *table, - int (*conv)(void *userData, const char *p), - void *userData); + int *table, + CONVERTER convert, + void *userData); #ifdef __cplusplus } #endif |