Annotation of win32/gnome/gnome-xml/include/libxml/HTMLparser.h, revision 1.2
1.1 paf 1: /*
2: * HTMLparser.h : interface for an HTML 4.0 non-verifying parser
3: *
4: * See Copyright for the status of this software.
5: *
6: * daniel@veillard.com
7: */
8:
9: #ifndef __HTML_PARSER_H__
10: #define __HTML_PARSER_H__
11: #include <libxml/parser.h>
12:
13: #ifdef __cplusplus
14: extern "C" {
15: #endif
16:
17: /*
18: * Most of the back-end structures from XML and HTML are shared.
19: */
20: typedef xmlParserCtxt htmlParserCtxt;
21: typedef xmlParserCtxtPtr htmlParserCtxtPtr;
22: typedef xmlParserNodeInfo htmlParserNodeInfo;
23: typedef xmlSAXHandler htmlSAXHandler;
24: typedef xmlSAXHandlerPtr htmlSAXHandlerPtr;
25: typedef xmlParserInput htmlParserInput;
26: typedef xmlParserInputPtr htmlParserInputPtr;
27: typedef xmlDocPtr htmlDocPtr;
28: typedef xmlNodePtr htmlNodePtr;
29:
30: /*
31: * Internal description of an HTML element, representing HTML 4.01
32: * and XHTML 1.0 (which share the same structure).
33: */
34: typedef struct _htmlElemDesc htmlElemDesc;
35: typedef htmlElemDesc *htmlElemDescPtr;
36: struct _htmlElemDesc {
37: const char *name; /* The tag name */
38: char startTag; /* Whether the start tag can be implied */
39: char endTag; /* Whether the end tag can be implied */
40: char saveEndTag; /* Whether the end tag should be saved */
41: char empty; /* Is this an empty element ? */
42: char depr; /* Is this a deprecated element ? */
43: char dtd; /* 1: only in Loose DTD, 2: only Frameset one */
44: char isinline; /* is this a block 0 or inline 1 element */
45: const char *desc; /* the description */
46:
47: /* NRK Jan.2003
48: * New fields encapsulating HTML structure
49: *
50: * Bugs:
51: * This is a very limited representation. It fails to tell us when
52: * an element *requires* subelements (we only have whether they're
53: * allowed or not), and it doesn't tell us where CDATA and PCDATA
54: * are allowed. Some element relationships are not fully represented:
55: * these are flagged with the word MODIFIER
56: */
57: const char** subelts; /* allowed sub-elements of this element */
58: const char* defaultsubelt; /* subelement for suggested auto-repair
59: if necessary or NULL */
60: const char** attrs_opt; /* Optional Attributes */
61: const char** attrs_depr; /* Additional deprecated attributes */
62: const char** attrs_req; /* Required attributes */
63: };
64:
65: /*
66: * Internal description of an HTML entity.
67: */
68: typedef struct _htmlEntityDesc htmlEntityDesc;
69: typedef htmlEntityDesc *htmlEntityDescPtr;
70: struct _htmlEntityDesc {
71: unsigned int value; /* the UNICODE value for the character */
72: const char *name; /* The entity name */
73: const char *desc; /* the description */
74: };
75:
76: /*
77: * There is only few public functions.
78: */
79: const htmlElemDesc * htmlTagLookup (const xmlChar *tag);
80: const htmlEntityDesc * htmlEntityLookup(const xmlChar *name);
81: const htmlEntityDesc * htmlEntityValueLookup(unsigned int value);
82:
83: int htmlIsAutoClosed(htmlDocPtr doc,
84: htmlNodePtr elem);
85: int htmlAutoCloseTag(htmlDocPtr doc,
86: const xmlChar *name,
87: htmlNodePtr elem);
88: const htmlEntityDesc * htmlParseEntityRef(htmlParserCtxtPtr ctxt,
89: xmlChar **str);
90: int htmlParseCharRef(htmlParserCtxtPtr ctxt);
91: void htmlParseElement(htmlParserCtxtPtr ctxt);
92:
1.2 ! paf 93: htmlParserCtxtPtr htmlCreateMemoryParserCtxt(const char *buffer,
! 94: int size);
! 95:
1.1 paf 96: int htmlParseDocument(htmlParserCtxtPtr ctxt);
97: htmlDocPtr htmlSAXParseDoc (xmlChar *cur,
98: const char *encoding,
99: htmlSAXHandlerPtr sax,
100: void *userData);
101: htmlDocPtr htmlParseDoc (xmlChar *cur,
102: const char *encoding);
103: htmlDocPtr htmlSAXParseFile(const char *filename,
104: const char *encoding,
105: htmlSAXHandlerPtr sax,
106: void *userData);
107: htmlDocPtr htmlParseFile (const char *filename,
108: const char *encoding);
109: int UTF8ToHtml (unsigned char *out,
110: int *outlen,
111: const unsigned char *in,
112: int *inlen);
113: int htmlEncodeEntities(unsigned char *out,
114: int *outlen,
115: const unsigned char *in,
116: int *inlen, int quoteChar);
117: int htmlIsScriptAttribute(const xmlChar *name);
118: int htmlHandleOmittedElem(int val);
119:
120: /**
121: * Interfaces for the Push mode.
122: */
123: void htmlFreeParserCtxt (htmlParserCtxtPtr ctxt);
124: htmlParserCtxtPtr htmlCreatePushParserCtxt(htmlSAXHandlerPtr sax,
125: void *user_data,
126: const char *chunk,
127: int size,
128: const char *filename,
129: xmlCharEncoding enc);
130: int htmlParseChunk (htmlParserCtxtPtr ctxt,
131: const char *chunk,
132: int size,
133: int terminate);
134:
135: /* NRK/Jan2003: further knowledge of HTML structure
136: */
137: typedef enum {
138: HTML_NA = 0 , /* something we don't check at all */
139: HTML_INVALID = 0x1 ,
140: HTML_DEPRECATED = 0x2 ,
141: HTML_VALID = 0x4 ,
142: HTML_REQUIRED = 0xc /* VALID bit set so ( & HTML_VALID ) is TRUE */
143: } htmlStatus ;
144:
145: /* Using htmlElemDesc rather than name here, to emphasise the fact
146: that otherwise there's a lookup overhead
147: */
148: htmlStatus htmlAttrAllowed(const htmlElemDesc*, const xmlChar*, int) ;
149: int htmlElementAllowedHere(const htmlElemDesc*, const xmlChar*) ;
150: htmlStatus htmlElementStatusHere(const htmlElemDesc*, const htmlElemDesc*) ;
151: htmlStatus htmlNodeStatus(const htmlNodePtr, int) ;
152: #define htmlDefaultSubelement(elt) elt->defaultsubelt
153: #define htmlElementAllowedHereDesc(parent,elt) \
154: htmlElementAllowedHere((parent), (elt)->name)
155: #define htmlRequiredAttrs(elt) (elt)->attrs_req
156:
157:
158: #ifdef __cplusplus
159: }
160: #endif
161:
162: #endif /* __HTML_PARSER_H__ */
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to HTMLparser.h CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / win32 / gnome / gnome-xml / include / libxml