Gumbo  0.9.2
A C library for parsing HTML.
 All Data Structures Files Functions Variables Typedefs Enumerations Enumerator Pages
gumbo.h
Go to the documentation of this file.
1 // Copyright 2010 Google Inc. All Rights Reserved.
2 //
3 // Licensed under the Apache License, Version 2.0 (the "License");
4 // you may not use this file except in compliance with the License.
5 // You may obtain a copy of the License at
6 //
7 // http://www.apache.org/licenses/LICENSE-2.0
8 //
9 // Unless required by applicable law or agreed to in writing, software
10 // distributed under the License is distributed on an "AS IS" BASIS,
11 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 // See the License for the specific language governing permissions and
13 // limitations under the License.
14 //
15 // Author: jdtang@google.com (Jonathan Tang)
16 //
17 // We use Gumbo as a prefix for types, gumbo_ as a prefix for functions, and
18 // GUMBO_ as a prefix for enum constants (static constants get the Google-style
19 // kGumbo prefix).
20 
42 #ifndef GUMBO_GUMBO_H_
43 #define GUMBO_GUMBO_H_
44 
45 #ifdef _MSC_VER
46 #define _CRT_SECURE_NO_WARNINGS
47 #define fileno _fileno
48 #endif
49 
50 #include <stdbool.h>
51 #include <stddef.h>
52 
53 #ifdef __cplusplus
54 extern "C" {
55 #endif
56 
67 typedef struct {
68  unsigned int line;
69  unsigned int column;
70  unsigned int offset;
72 
78 
88 typedef struct {
90  const char* data;
91 
93  size_t length;
95 
98 
104  const GumboStringPiece* str1, const GumboStringPiece* str2);
105 
111  const GumboStringPiece* str1, const GumboStringPiece* str2);
112 
122 typedef struct {
126  void** data;
127 
129  unsigned int length;
130 
132  unsigned int capacity;
133 } GumboVector;
134 
136 extern const GumboVector kGumboEmptyVector;
137 
142 int gumbo_vector_index_of(GumboVector* vector, const void* element);
143 
156 typedef enum {
157 // Load all the tags from an external source, generated from tag.in.
158 #include "tag_enum.h"
159  // Used for all tags that don't have special handling in HTML. Add new tags
160  // to the end of tag.in so as to preserve backwards-compatibility.
161  GUMBO_TAG_UNKNOWN,
162  // A marker value to indicate the end of the enum, for iterating over it.
163  // Also used as the terminator for varargs functions that take tags.
164  GUMBO_TAG_LAST,
165 } GumboTag;
166 
172 const char* gumbo_normalized_tagname(GumboTag tag);
173 
185 
198 const char* gumbo_normalize_svg_tagname(const GumboStringPiece* tagname);
199 
204 GumboTag gumbo_tag_enum(const char* tagname);
205 GumboTag gumbo_tagn_enum(const char* tagname, unsigned int length);
206 
212 typedef enum {
213  GUMBO_ATTR_NAMESPACE_NONE,
214  GUMBO_ATTR_NAMESPACE_XLINK,
215  GUMBO_ATTR_NAMESPACE_XML,
216  GUMBO_ATTR_NAMESPACE_XMLNS,
218 
224 typedef struct {
232 
237  const char* name;
238 
244 
251  const char* value;
252 
262 
265 
272 
275 
279 
285 GumboAttribute* gumbo_get_attribute(const GumboVector* attrs, const char* name);
286 
291 typedef enum {
310 } GumboNodeType;
311 
316 typedef struct GumboInternalNode GumboNode;
317 
321 typedef enum {
322  GUMBO_DOCTYPE_NO_QUIRKS,
323  GUMBO_DOCTYPE_QUIRKS,
324  GUMBO_DOCTYPE_LIMITED_QUIRKS
326 
334 typedef enum {
335  GUMBO_NAMESPACE_HTML,
336  GUMBO_NAMESPACE_SVG,
337  GUMBO_NAMESPACE_MATHML
339 
348 typedef enum {
354 
362 
375 
376  // Value 1 << 2 was for a flag that has since been removed.
377 
383 
391 
394 
397 
404 
407 
410 
417 
421 typedef struct {
427  GumboVector /* GumboNode* */ children;
428 
429  // True if there was an explicit doctype token as opposed to it being omitted.
430  bool has_doctype;
431 
432  // Fields from the doctype token, copied verbatim.
433  const char* name;
434  const char* public_identifier;
435  const char* system_identifier;
436 
442 } GumboDocument;
443 
448 typedef struct {
453  const char* text;
454 
460 
466 } GumboText;
467 
472 typedef struct {
477  GumboVector /* GumboNode* */ children;
478 
481 
484 
492 
499 
502 
505 
510  GumboVector /* GumboAttribute* */ attributes;
511 } GumboElement;
512 
520 
522  GumboNode* parent;
523 
526 
533 
535  union {
536  GumboDocument document; // For GUMBO_NODE_DOCUMENT.
537  GumboElement element; // For GUMBO_NODE_ELEMENT.
538  GumboText text; // For everything else.
539  } v;
540 };
541 
548 // TODO(jdtang): Add checks throughout the codebase for out-of-memory condition.
549 typedef void* (*GumboAllocatorFunction)(void* userdata, size_t size);
550 
555 typedef void (*GumboDeallocatorFunction)(void* userdata, void* ptr);
556 
563 typedef struct GumboInternalOptions {
566 
569 
574  void* userdata;
575 
580  int tab_stop;
581 
587 
596 
611 
619 } GumboOptions;
620 
623 
625 typedef struct GumboInternalOutput {
630  GumboNode* document;
631 
636  GumboNode* root;
637 
645  GumboVector /* GumboError */ errors;
646 } GumboOutput;
647 
655 GumboOutput* gumbo_parse(const char* buffer);
656 
662  const GumboOptions* options, const char* buffer, size_t buffer_length);
663 
665 void gumbo_destroy_output(const GumboOptions* options, GumboOutput* output);
666 
667 #ifdef __cplusplus
668 }
669 #endif
670 
671 #endif // GUMBO_GUMBO_H_
Definition: gumbo.h:625
GumboNamespaceEnum tag_namespace
Definition: gumbo.h:483
GumboSourcePosition value_start
Definition: gumbo.h:274
Definition: gumbo.h:563
Definition: gumbo.h:299
Definition: gumbo.h:421
Definition: gumbo.h:472
GumboSourcePosition start_pos
Definition: gumbo.h:501
GumboSourcePosition name_end
Definition: gumbo.h:271
const char * value
Definition: gumbo.h:251
GumboVector children
Definition: gumbo.h:477
Definition: gumbo.h:382
void(* GumboDeallocatorFunction)(void *userdata, void *ptr)
Definition: gumbo.h:555
GumboStringPiece original_value
Definition: gumbo.h:261
Definition: gumbo.h:415
GumboSourcePosition end_pos
Definition: gumbo.h:504
Definition: gumbo.h:396
GumboOutput * gumbo_parse(const char *buffer)
Definition: gumbo.h:448
unsigned int length
Definition: gumbo.h:129
GumboParseFlags parse_flags
Definition: gumbo.h:532
bool stop_on_first_error
Definition: gumbo.h:586
const GumboVector kGumboEmptyVector
GumboSourcePosition value_end
Definition: gumbo.h:277
GumboNodeType type
Definition: gumbo.h:519
GumboTag
Definition: gumbo.h:156
void ** data
Definition: gumbo.h:126
GumboTag tag
Definition: gumbo.h:480
Definition: gumbo.h:122
GumboStringPiece original_tag
Definition: gumbo.h:491
Definition: gumbo.h:224
GumboAllocatorFunction allocator
Definition: gumbo.h:565
GumboStringPiece original_end_tag
Definition: gumbo.h:498
Definition: gumbo.h:297
const char * text
Definition: gumbo.h:453
Definition: gumbo.h:393
Definition: gumbo.h:67
Definition: gumbo.h:309
GumboQuirksModeEnum doc_type_quirks_mode
Definition: gumbo.h:441
void *(* GumboAllocatorFunction)(void *userdata, size_t size)
Definition: gumbo.h:549
const GumboOptions kGumboDefaultOptions
void gumbo_destroy_output(const GumboOptions *options, GumboOutput *output)
Definition: gumbo.h:517
unsigned int capacity
Definition: gumbo.h:132
int max_errors
Definition: gumbo.h:595
GumboStringPiece original_name
Definition: gumbo.h:243
GumboNode * root
Definition: gumbo.h:636
int gumbo_vector_index_of(GumboVector *vector, const void *element)
size_t index_within_parent
Definition: gumbo.h:525
bool gumbo_string_equals_ignore_case(const GumboStringPiece *str1, const GumboStringPiece *str2)
void * userdata
Definition: gumbo.h:574
GumboStringPiece original_text
Definition: gumbo.h:459
const GumboStringPiece kGumboEmptyString
int tab_stop
Definition: gumbo.h:580
GumboParseFlags
Definition: gumbo.h:348
GumboTag fragment_context
Definition: gumbo.h:610
GumboOutput * gumbo_parse_with_options(const GumboOptions *options, const char *buffer, size_t buffer_length)
bool gumbo_string_equals(const GumboStringPiece *str1, const GumboStringPiece *str2)
const GumboSourcePosition kGumboEmptySourcePosition
GumboAttribute * gumbo_get_attribute(const GumboVector *attrs, const char *name)
GumboDeallocatorFunction deallocator
Definition: gumbo.h:568
Definition: gumbo.h:353
GumboNode * document
Definition: gumbo.h:630
GumboNodeType
Definition: gumbo.h:291
const char * data
Definition: gumbo.h:90
Definition: gumbo.h:293
Definition: gumbo.h:295
size_t length
Definition: gumbo.h:93
const char * gumbo_normalized_tagname(GumboTag tag)
const char * name
Definition: gumbo.h:237
union GumboInternalNode::@0 v
GumboVector children
Definition: gumbo.h:427
GumboAttributeNamespaceEnum attr_namespace
Definition: gumbo.h:231
GumboSourcePosition start_pos
Definition: gumbo.h:465
Definition: gumbo.h:303
Definition: gumbo.h:301
void gumbo_tag_from_original_text(GumboStringPiece *text)
Definition: gumbo.h:361
GumboNamespaceEnum
Definition: gumbo.h:334
const char * gumbo_normalize_svg_tagname(const GumboStringPiece *tagname)
Definition: gumbo.h:88
GumboSourcePosition name_start
Definition: gumbo.h:264
GumboVector attributes
Definition: gumbo.h:510
GumboQuirksModeEnum
Definition: gumbo.h:321
GumboNamespaceEnum fragment_namespace
Definition: gumbo.h:618
GumboVector errors
Definition: gumbo.h:645
Definition: gumbo.h:374
GumboTag gumbo_tag_enum(const char *tagname)
GumboAttributeNamespaceEnum
Definition: gumbo.h:212
GumboNode * parent
Definition: gumbo.h:522