mirror of
https://github.com/GNOME/libxml2.git
synced 2025-01-30 17:30:01 +08:00
bc765307ff
- HTMLparser.c: fixed htmlStartCloseIndexinitialized init - entities.h: exported xmlInitializePredefinedEntities - parser.[ch] : added xmlInitParser() - parserInternals.h : had to export htmlInitAutoClose() Daniel
301 lines
9.1 KiB
C
301 lines
9.1 KiB
C
/*
|
|
* parserInternals.h : internals routines exported by the parser.
|
|
*
|
|
* See Copyright for the status of this software.
|
|
*
|
|
* Daniel.Veillard@w3.org
|
|
*/
|
|
|
|
#ifndef __XML_PARSER_INTERNALS_H__
|
|
#define __XML_PARSER_INTERNALS_H__
|
|
|
|
#include <libxml/parser.h>
|
|
|
|
#ifdef __cplusplus
|
|
extern "C" {
|
|
#endif
|
|
|
|
/*
|
|
* Identifiers can be longer, but this will be more costly
|
|
* at runtime.
|
|
*/
|
|
#define XML_MAX_NAMELEN 100
|
|
|
|
/*
|
|
* The parser tries to always have that amount of input ready
|
|
* one of the point is providing context when reporting errors
|
|
*/
|
|
#define INPUT_CHUNK 250
|
|
|
|
/************************************************************************
|
|
* *
|
|
* UNICODE version of the macros. *
|
|
* *
|
|
************************************************************************/
|
|
/*
|
|
* [2] Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD]
|
|
* | [#x10000-#x10FFFF]
|
|
* any Unicode character, excluding the surrogate blocks, FFFE, and FFFF.
|
|
*/
|
|
#define IS_CHAR(c) \
|
|
(((c) == 0x09) || ((c) == 0x0A) || ((c) == 0x0D) || \
|
|
(((c) >= 0x20) && ((c) <= 0xD7FF)) || \
|
|
(((c) >= 0xE000) && ((c) <= 0xFFFD)) || \
|
|
(((c) >= 0x10000) && ((c) <= 0x10FFFF)))
|
|
|
|
/*
|
|
* [3] S ::= (#x20 | #x9 | #xD | #xA)+
|
|
*/
|
|
#define IS_BLANK(c) (((c) == 0x20) || ((c) == 0x09) || ((c) == 0xA) || \
|
|
((c) == 0x0D))
|
|
|
|
/*
|
|
* [85] BaseChar ::= ... long list see REC ...
|
|
*/
|
|
#define IS_BASECHAR(c) xmlIsBaseChar(c)
|
|
|
|
/*
|
|
* [88] Digit ::= ... long list see REC ...
|
|
*/
|
|
#define IS_DIGIT(c) xmlIsDigit(c)
|
|
|
|
/*
|
|
* [87] CombiningChar ::= ... long list see REC ...
|
|
*/
|
|
#define IS_COMBINING(c) xmlIsCombining(c)
|
|
|
|
/*
|
|
* [89] Extender ::= #x00B7 | #x02D0 | #x02D1 | #x0387 | #x0640 |
|
|
* #x0E46 | #x0EC6 | #x3005 | [#x3031-#x3035] |
|
|
* [#x309D-#x309E] | [#x30FC-#x30FE]
|
|
*/
|
|
#define IS_EXTENDER(c) xmlIsExtender(c)
|
|
|
|
/*
|
|
* [86] Ideographic ::= [#x4E00-#x9FA5] | #x3007 | [#x3021-#x3029]
|
|
*/
|
|
#define IS_IDEOGRAPHIC(c) xmlIsIdeographic(c)
|
|
|
|
/*
|
|
* [84] Letter ::= BaseChar | Ideographic
|
|
*/
|
|
#define IS_LETTER(c) (IS_BASECHAR(c) || IS_IDEOGRAPHIC(c))
|
|
|
|
|
|
/*
|
|
* [13] PubidChar ::= #x20 | #xD | #xA | [a-zA-Z0-9] | [-'()+,./:=?;!*#@$_%]
|
|
*/
|
|
#define IS_PUBIDCHAR(c) xmlIsPubidChar(c)
|
|
|
|
#define SKIP_EOL(p) \
|
|
if (*(p) == 0x13) { p++ ; if (*(p) == 0x10) p++; } \
|
|
if (*(p) == 0x10) { p++ ; if (*(p) == 0x13) p++; }
|
|
|
|
#define MOVETO_ENDTAG(p) \
|
|
while ((*p) && (*(p) != '>')) (p)++
|
|
|
|
#define MOVETO_STARTTAG(p) \
|
|
while ((*p) && (*(p) != '<')) (p)++
|
|
|
|
/**
|
|
* Global vaiables affecting the default parser behaviour.
|
|
*/
|
|
|
|
extern int xmlParserDebugEntities;
|
|
extern int xmlGetWarningsDefaultValue;
|
|
extern int xmlParserDebugEntities;
|
|
extern int xmlSubstituteEntitiesDefaultValue;
|
|
extern int xmlDoValidityCheckingDefaultValue;
|
|
extern int xmlPedanticParserDefaultValue;
|
|
extern int xmlKeepBlanksDefaultValue;
|
|
|
|
/*
|
|
* Function to finish teh work of the macros where needed
|
|
*/
|
|
int xmlIsBaseChar (int c);
|
|
int xmlIsBlank (int c);
|
|
int xmlIsPubidChar (int c);
|
|
int xmlIsLetter (int c);
|
|
int xmlIsDigit (int c);
|
|
int xmlIsIdeographic(int c);
|
|
int xmlIsCombining (int c);
|
|
int xmlIsExtender (int c);
|
|
int xmlIsCombining (int c);
|
|
int xmlIsChar (int c);
|
|
|
|
/**
|
|
* Parser context
|
|
*/
|
|
xmlParserCtxtPtr xmlCreateDocParserCtxt (xmlChar *cur);
|
|
xmlParserCtxtPtr xmlCreateFileParserCtxt (const char *filename);
|
|
xmlParserCtxtPtr xmlCreateMemoryParserCtxt(char *buffer,
|
|
int size);
|
|
xmlParserCtxtPtr xmlNewParserCtxt (void);
|
|
xmlParserCtxtPtr xmlCreateEntityParserCtxt(const xmlChar *URL,
|
|
const xmlChar *ID,
|
|
const xmlChar *base);
|
|
int xmlSwitchEncoding (xmlParserCtxtPtr ctxt,
|
|
xmlCharEncoding enc);
|
|
int xmlSwitchToEncoding (xmlParserCtxtPtr ctxt,
|
|
xmlCharEncodingHandlerPtr handler);
|
|
void xmlFreeParserCtxt (xmlParserCtxtPtr ctxt);
|
|
|
|
/**
|
|
* Entities
|
|
*/
|
|
void xmlHandleEntity (xmlParserCtxtPtr ctxt,
|
|
xmlEntityPtr entity);
|
|
|
|
/**
|
|
* Input Streams
|
|
*/
|
|
xmlParserInputPtr xmlNewEntityInputStream (xmlParserCtxtPtr ctxt,
|
|
xmlEntityPtr entity);
|
|
void xmlPushInput (xmlParserCtxtPtr ctxt,
|
|
xmlParserInputPtr input);
|
|
xmlChar xmlPopInput (xmlParserCtxtPtr ctxt);
|
|
void xmlFreeInputStream (xmlParserInputPtr input);
|
|
xmlParserInputPtr xmlNewInputFromFile (xmlParserCtxtPtr ctxt,
|
|
const char *filename);
|
|
xmlParserInputPtr xmlNewInputStream (xmlParserCtxtPtr ctxt);
|
|
|
|
/**
|
|
* Namespaces.
|
|
*/
|
|
xmlChar * xmlSplitQName (xmlParserCtxtPtr ctxt,
|
|
const xmlChar *name,
|
|
xmlChar **prefix);
|
|
xmlChar * xmlNamespaceParseNCName (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlNamespaceParseQName (xmlParserCtxtPtr ctxt,
|
|
xmlChar **prefix);
|
|
xmlChar * xmlNamespaceParseNSDef (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseQuotedString (xmlParserCtxtPtr ctxt);
|
|
void xmlParseNamespace (xmlParserCtxtPtr ctxt);
|
|
|
|
/**
|
|
* Generic production rules
|
|
*/
|
|
xmlChar * xmlScanName (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseName (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseNmtoken (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseEntityValue (xmlParserCtxtPtr ctxt,
|
|
xmlChar **orig);
|
|
xmlChar * xmlParseAttValue (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseSystemLiteral (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParsePubidLiteral (xmlParserCtxtPtr ctxt);
|
|
void xmlParseCharData (xmlParserCtxtPtr ctxt,
|
|
int cdata);
|
|
xmlChar * xmlParseExternalID (xmlParserCtxtPtr ctxt,
|
|
xmlChar **publicID,
|
|
int strict);
|
|
void xmlParseComment (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParsePITarget (xmlParserCtxtPtr ctxt);
|
|
void xmlParsePI (xmlParserCtxtPtr ctxt);
|
|
void xmlParseNotationDecl (xmlParserCtxtPtr ctxt);
|
|
void xmlParseEntityDecl (xmlParserCtxtPtr ctxt);
|
|
int xmlParseDefaultDecl (xmlParserCtxtPtr ctxt,
|
|
xmlChar **value);
|
|
xmlEnumerationPtr xmlParseNotationType (xmlParserCtxtPtr ctxt);
|
|
xmlEnumerationPtr xmlParseEnumerationType (xmlParserCtxtPtr ctxt);
|
|
int xmlParseEnumeratedType (xmlParserCtxtPtr ctxt,
|
|
xmlEnumerationPtr *tree);
|
|
int xmlParseAttributeType (xmlParserCtxtPtr ctxt,
|
|
xmlEnumerationPtr *tree);
|
|
void xmlParseAttributeListDecl(xmlParserCtxtPtr ctxt);
|
|
xmlElementContentPtr xmlParseElementMixedContentDecl
|
|
(xmlParserCtxtPtr ctxt);
|
|
xmlElementContentPtr xmlParseElementChildrenContentDecl
|
|
(xmlParserCtxtPtr ctxt);
|
|
int xmlParseElementContentDecl(xmlParserCtxtPtr ctxt,
|
|
xmlChar *name,
|
|
xmlElementContentPtr *result);
|
|
int xmlParseElementDecl (xmlParserCtxtPtr ctxt);
|
|
void xmlParseMarkupDecl (xmlParserCtxtPtr ctxt);
|
|
int xmlParseCharRef (xmlParserCtxtPtr ctxt);
|
|
xmlEntityPtr xmlParseEntityRef (xmlParserCtxtPtr ctxt);
|
|
void xmlParseReference (xmlParserCtxtPtr ctxt);
|
|
void xmlParsePEReference (xmlParserCtxtPtr ctxt);
|
|
void xmlParseDocTypeDecl (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseAttribute (xmlParserCtxtPtr ctxt,
|
|
xmlChar **value);
|
|
xmlChar * xmlParseStartTag (xmlParserCtxtPtr ctxt);
|
|
void xmlParseEndTag (xmlParserCtxtPtr ctxt);
|
|
void xmlParseCDSect (xmlParserCtxtPtr ctxt);
|
|
void xmlParseContent (xmlParserCtxtPtr ctxt);
|
|
void xmlParseElement (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseVersionNum (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseVersionInfo (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseEncName (xmlParserCtxtPtr ctxt);
|
|
xmlChar * xmlParseEncodingDecl (xmlParserCtxtPtr ctxt);
|
|
int xmlParseSDDecl (xmlParserCtxtPtr ctxt);
|
|
void xmlParseXMLDecl (xmlParserCtxtPtr ctxt);
|
|
void xmlParseTextDecl (xmlParserCtxtPtr ctxt);
|
|
void xmlParseMisc (xmlParserCtxtPtr ctxt);
|
|
void xmlParseExternalSubset (xmlParserCtxtPtr ctxt,
|
|
const xmlChar *ExternalID,
|
|
const xmlChar *SystemID);
|
|
/*
|
|
* Entities substitution
|
|
*/
|
|
#define XML_SUBSTITUTE_NONE 0
|
|
#define XML_SUBSTITUTE_REF 1
|
|
#define XML_SUBSTITUTE_PEREF 2
|
|
#define XML_SUBSTITUTE_BOTH 3
|
|
|
|
xmlChar * xmlDecodeEntities (xmlParserCtxtPtr ctxt,
|
|
int len,
|
|
int what,
|
|
xmlChar end,
|
|
xmlChar end2,
|
|
xmlChar end3);
|
|
xmlChar * xmlStringDecodeEntities (xmlParserCtxtPtr ctxt,
|
|
const xmlChar *str,
|
|
int what,
|
|
xmlChar end,
|
|
xmlChar end2,
|
|
xmlChar end3);
|
|
|
|
/*
|
|
* Generated by MACROS on top of parser.c c.f. PUSH_AND_POP
|
|
*/
|
|
int nodePush (xmlParserCtxtPtr ctxt,
|
|
xmlNodePtr value);
|
|
xmlNodePtr nodePop (xmlParserCtxtPtr ctxt);
|
|
int inputPush (xmlParserCtxtPtr ctxt,
|
|
xmlParserInputPtr value);
|
|
xmlParserInputPtr inputPop (xmlParserCtxtPtr ctxt);
|
|
|
|
/*
|
|
* other comodities shared between parser.c and parserInternals
|
|
*/
|
|
int xmlSkipBlankChars (xmlParserCtxtPtr ctxt);
|
|
int xmlStringCurrentChar (xmlParserCtxtPtr ctxt,
|
|
const xmlChar *cur,
|
|
int *len);
|
|
void xmlParserHandlePEReference(xmlParserCtxtPtr ctxt);
|
|
void xmlParserHandleReference(xmlParserCtxtPtr ctxt);
|
|
xmlChar *namePop (xmlParserCtxtPtr ctxt);
|
|
int xmlCheckLanguageID (const xmlChar *lang);
|
|
|
|
/*
|
|
* Really core function shared with HTML parser
|
|
*/
|
|
int xmlCurrentChar (xmlParserCtxtPtr ctxt,
|
|
int *len);
|
|
int xmlCopyChar (int len,
|
|
xmlChar *out,
|
|
int val);
|
|
void xmlNextChar (xmlParserCtxtPtr ctxt);
|
|
void xmlParserInputShrink (xmlParserInputPtr in);
|
|
|
|
#ifdef LIBXML_HTML_ENABLED
|
|
/*
|
|
* Actually comes from the HTML parser but launched from the init stuff
|
|
*/
|
|
void htmlInitAutoClose (void);
|
|
#endif
|
|
#ifdef __cplusplus
|
|
}
|
|
#endif
|
|
#endif /* __XML_PARSER_INTERNALS_H__ */
|