Nokogiri_precompiled_aarch64_dedshit 1.14.5
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/Gemfile +44 -0
- data/LICENSE-DEPENDENCIES.md +2224 -0
- data/LICENSE.md +9 -0
- data/README.md +287 -0
- data/bin/nokogiri +131 -0
- data/dependencies.yml +41 -0
- data/ext/java/nokogiri/Html4Document.java +157 -0
- data/ext/java/nokogiri/Html4ElementDescription.java +133 -0
- data/ext/java/nokogiri/Html4EntityLookup.java +63 -0
- data/ext/java/nokogiri/Html4SaxParserContext.java +289 -0
- data/ext/java/nokogiri/Html4SaxPushParser.java +213 -0
- data/ext/java/nokogiri/NokogiriService.java +613 -0
- data/ext/java/nokogiri/XmlAttr.java +154 -0
- data/ext/java/nokogiri/XmlAttributeDecl.java +119 -0
- data/ext/java/nokogiri/XmlCdata.java +60 -0
- data/ext/java/nokogiri/XmlComment.java +77 -0
- data/ext/java/nokogiri/XmlDocument.java +705 -0
- data/ext/java/nokogiri/XmlDocumentFragment.java +163 -0
- data/ext/java/nokogiri/XmlDtd.java +516 -0
- data/ext/java/nokogiri/XmlElement.java +44 -0
- data/ext/java/nokogiri/XmlElementContent.java +412 -0
- data/ext/java/nokogiri/XmlElementDecl.java +148 -0
- data/ext/java/nokogiri/XmlEntityDecl.java +151 -0
- data/ext/java/nokogiri/XmlEntityReference.java +79 -0
- data/ext/java/nokogiri/XmlNamespace.java +193 -0
- data/ext/java/nokogiri/XmlNode.java +1938 -0
- data/ext/java/nokogiri/XmlNodeSet.java +463 -0
- data/ext/java/nokogiri/XmlProcessingInstruction.java +79 -0
- data/ext/java/nokogiri/XmlReader.java +615 -0
- data/ext/java/nokogiri/XmlRelaxng.java +133 -0
- data/ext/java/nokogiri/XmlSaxParserContext.java +329 -0
- data/ext/java/nokogiri/XmlSaxPushParser.java +288 -0
- data/ext/java/nokogiri/XmlSchema.java +423 -0
- data/ext/java/nokogiri/XmlSyntaxError.java +137 -0
- data/ext/java/nokogiri/XmlText.java +90 -0
- data/ext/java/nokogiri/XmlXpathContext.java +305 -0
- data/ext/java/nokogiri/XsltStylesheet.java +368 -0
- data/ext/java/nokogiri/internals/ClosedStreamException.java +13 -0
- data/ext/java/nokogiri/internals/HtmlDomParserContext.java +252 -0
- data/ext/java/nokogiri/internals/IgnoreSchemaErrorsErrorHandler.java +27 -0
- data/ext/java/nokogiri/internals/NokogiriBlockingQueueInputStream.java +178 -0
- data/ext/java/nokogiri/internals/NokogiriDomParser.java +99 -0
- data/ext/java/nokogiri/internals/NokogiriEntityResolver.java +140 -0
- data/ext/java/nokogiri/internals/NokogiriErrorHandler.java +65 -0
- data/ext/java/nokogiri/internals/NokogiriHandler.java +339 -0
- data/ext/java/nokogiri/internals/NokogiriHelpers.java +817 -0
- data/ext/java/nokogiri/internals/NokogiriNamespaceCache.java +228 -0
- data/ext/java/nokogiri/internals/NokogiriNamespaceContext.java +110 -0
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler.java +86 -0
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler4NekoHtml.java +107 -0
- data/ext/java/nokogiri/internals/NokogiriStrictErrorHandler.java +62 -0
- data/ext/java/nokogiri/internals/NokogiriXPathFunction.java +165 -0
- data/ext/java/nokogiri/internals/NokogiriXPathFunctionResolver.java +50 -0
- data/ext/java/nokogiri/internals/NokogiriXPathVariableResolver.java +37 -0
- data/ext/java/nokogiri/internals/NokogiriXsltErrorListener.java +70 -0
- data/ext/java/nokogiri/internals/ParserContext.java +262 -0
- data/ext/java/nokogiri/internals/ReaderNode.java +564 -0
- data/ext/java/nokogiri/internals/SaveContextVisitor.java +865 -0
- data/ext/java/nokogiri/internals/SchemaErrorHandler.java +50 -0
- data/ext/java/nokogiri/internals/XalanDTMManagerPatch.java +174 -0
- data/ext/java/nokogiri/internals/XmlDeclHandler.java +11 -0
- data/ext/java/nokogiri/internals/XmlDomParserContext.java +265 -0
- data/ext/java/nokogiri/internals/XmlSaxParser.java +40 -0
- data/ext/java/nokogiri/internals/c14n/AttrCompare.java +122 -0
- data/ext/java/nokogiri/internals/c14n/C14nHelper.java +178 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalFilter.java +43 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizationException.java +106 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer.java +278 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer11.java +664 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer11_OmitComments.java +45 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer11_WithComments.java +45 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315.java +388 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315Excl.java +308 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315ExclOmitComments.java +47 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315ExclWithComments.java +51 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315OmitComments.java +51 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315WithComments.java +50 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizerBase.java +660 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizerPhysical.java +194 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizerSpi.java +77 -0
- data/ext/java/nokogiri/internals/c14n/Constants.java +45 -0
- data/ext/java/nokogiri/internals/c14n/ElementProxy.java +325 -0
- data/ext/java/nokogiri/internals/c14n/HelperNodeList.java +106 -0
- data/ext/java/nokogiri/internals/c14n/IgnoreAllErrorHandler.java +86 -0
- data/ext/java/nokogiri/internals/c14n/InclusiveNamespaces.java +181 -0
- data/ext/java/nokogiri/internals/c14n/InvalidCanonicalizerException.java +87 -0
- data/ext/java/nokogiri/internals/c14n/NameSpaceSymbTable.java +452 -0
- data/ext/java/nokogiri/internals/c14n/NodeFilter.java +52 -0
- data/ext/java/nokogiri/internals/c14n/UtfHelpper.java +190 -0
- data/ext/java/nokogiri/internals/c14n/XMLUtils.java +540 -0
- data/ext/java/nokogiri/internals/dom2dtm/DOM2DTM.java +1712 -0
- data/ext/java/nokogiri/internals/dom2dtm/DOM2DTMdefaultNamespaceDeclarationNode.java +737 -0
- data/ext/nokogiri/depend +38 -0
- data/ext/nokogiri/extconf.rb +1086 -0
- data/ext/nokogiri/gumbo.c +594 -0
- data/ext/nokogiri/html4_document.c +167 -0
- data/ext/nokogiri/html4_element_description.c +294 -0
- data/ext/nokogiri/html4_entity_lookup.c +37 -0
- data/ext/nokogiri/html4_sax_parser_context.c +116 -0
- data/ext/nokogiri/html4_sax_push_parser.c +95 -0
- data/ext/nokogiri/libxml2_backwards_compat.c +121 -0
- data/ext/nokogiri/nokogiri.c +265 -0
- data/ext/nokogiri/nokogiri.h +235 -0
- data/ext/nokogiri/test_global_handlers.c +42 -0
- data/ext/nokogiri/xml_attr.c +103 -0
- data/ext/nokogiri/xml_attribute_decl.c +70 -0
- data/ext/nokogiri/xml_cdata.c +57 -0
- data/ext/nokogiri/xml_comment.c +62 -0
- data/ext/nokogiri/xml_document.c +689 -0
- data/ext/nokogiri/xml_document_fragment.c +44 -0
- data/ext/nokogiri/xml_dtd.c +210 -0
- data/ext/nokogiri/xml_element_content.c +128 -0
- data/ext/nokogiri/xml_element_decl.c +69 -0
- data/ext/nokogiri/xml_encoding_handler.c +104 -0
- data/ext/nokogiri/xml_entity_decl.c +112 -0
- data/ext/nokogiri/xml_entity_reference.c +50 -0
- data/ext/nokogiri/xml_namespace.c +186 -0
- data/ext/nokogiri/xml_node.c +2426 -0
- data/ext/nokogiri/xml_node_set.c +496 -0
- data/ext/nokogiri/xml_processing_instruction.c +54 -0
- data/ext/nokogiri/xml_reader.c +794 -0
- data/ext/nokogiri/xml_relax_ng.c +164 -0
- data/ext/nokogiri/xml_sax_parser.c +316 -0
- data/ext/nokogiri/xml_sax_parser_context.c +283 -0
- data/ext/nokogiri/xml_sax_push_parser.c +166 -0
- data/ext/nokogiri/xml_schema.c +260 -0
- data/ext/nokogiri/xml_syntax_error.c +85 -0
- data/ext/nokogiri/xml_text.c +48 -0
- data/ext/nokogiri/xml_xpath_context.c +415 -0
- data/ext/nokogiri/xslt_stylesheet.c +363 -0
- data/gumbo-parser/CHANGES.md +63 -0
- data/gumbo-parser/Makefile +111 -0
- data/gumbo-parser/THANKS +27 -0
- data/gumbo-parser/src/Makefile +34 -0
- data/gumbo-parser/src/README.md +41 -0
- data/gumbo-parser/src/ascii.c +75 -0
- data/gumbo-parser/src/ascii.h +115 -0
- data/gumbo-parser/src/attribute.c +42 -0
- data/gumbo-parser/src/attribute.h +17 -0
- data/gumbo-parser/src/char_ref.c +22225 -0
- data/gumbo-parser/src/char_ref.h +29 -0
- data/gumbo-parser/src/char_ref.rl +2154 -0
- data/gumbo-parser/src/error.c +626 -0
- data/gumbo-parser/src/error.h +148 -0
- data/gumbo-parser/src/foreign_attrs.c +104 -0
- data/gumbo-parser/src/foreign_attrs.gperf +27 -0
- data/gumbo-parser/src/insertion_mode.h +33 -0
- data/gumbo-parser/src/macros.h +91 -0
- data/gumbo-parser/src/nokogiri_gumbo.h +944 -0
- data/gumbo-parser/src/parser.c +4878 -0
- data/gumbo-parser/src/parser.h +41 -0
- data/gumbo-parser/src/replacement.h +33 -0
- data/gumbo-parser/src/string_buffer.c +103 -0
- data/gumbo-parser/src/string_buffer.h +68 -0
- data/gumbo-parser/src/string_piece.c +48 -0
- data/gumbo-parser/src/svg_attrs.c +174 -0
- data/gumbo-parser/src/svg_attrs.gperf +77 -0
- data/gumbo-parser/src/svg_tags.c +137 -0
- data/gumbo-parser/src/svg_tags.gperf +55 -0
- data/gumbo-parser/src/tag.c +223 -0
- data/gumbo-parser/src/tag_lookup.c +382 -0
- data/gumbo-parser/src/tag_lookup.gperf +170 -0
- data/gumbo-parser/src/tag_lookup.h +13 -0
- data/gumbo-parser/src/token_buffer.c +79 -0
- data/gumbo-parser/src/token_buffer.h +71 -0
- data/gumbo-parser/src/token_type.h +17 -0
- data/gumbo-parser/src/tokenizer.c +3463 -0
- data/gumbo-parser/src/tokenizer.h +112 -0
- data/gumbo-parser/src/tokenizer_states.h +339 -0
- data/gumbo-parser/src/utf8.c +245 -0
- data/gumbo-parser/src/utf8.h +164 -0
- data/gumbo-parser/src/util.c +66 -0
- data/gumbo-parser/src/util.h +34 -0
- data/gumbo-parser/src/vector.c +111 -0
- data/gumbo-parser/src/vector.h +45 -0
- data/lib/nokogiri/class_resolver.rb +67 -0
- data/lib/nokogiri/css/node.rb +54 -0
- data/lib/nokogiri/css/parser.rb +770 -0
- data/lib/nokogiri/css/parser.y +277 -0
- data/lib/nokogiri/css/parser_extras.rb +96 -0
- data/lib/nokogiri/css/syntax_error.rb +9 -0
- data/lib/nokogiri/css/tokenizer.rb +155 -0
- data/lib/nokogiri/css/tokenizer.rex +56 -0
- data/lib/nokogiri/css/xpath_visitor.rb +359 -0
- data/lib/nokogiri/css.rb +66 -0
- data/lib/nokogiri/decorators/slop.rb +44 -0
- data/lib/nokogiri/encoding_handler.rb +57 -0
- data/lib/nokogiri/extension.rb +32 -0
- data/lib/nokogiri/gumbo.rb +15 -0
- data/lib/nokogiri/html.rb +48 -0
- data/lib/nokogiri/html4/builder.rb +37 -0
- data/lib/nokogiri/html4/document.rb +214 -0
- data/lib/nokogiri/html4/document_fragment.rb +54 -0
- data/lib/nokogiri/html4/element_description.rb +25 -0
- data/lib/nokogiri/html4/element_description_defaults.rb +572 -0
- data/lib/nokogiri/html4/encoding_reader.rb +121 -0
- data/lib/nokogiri/html4/entity_lookup.rb +15 -0
- data/lib/nokogiri/html4/sax/parser.rb +63 -0
- data/lib/nokogiri/html4/sax/parser_context.rb +20 -0
- data/lib/nokogiri/html4/sax/push_parser.rb +37 -0
- data/lib/nokogiri/html4.rb +47 -0
- data/lib/nokogiri/html5/document.rb +168 -0
- data/lib/nokogiri/html5/document_fragment.rb +90 -0
- data/lib/nokogiri/html5/node.rb +98 -0
- data/lib/nokogiri/html5.rb +389 -0
- data/lib/nokogiri/jruby/dependencies.rb +3 -0
- data/lib/nokogiri/jruby/isorelax/isorelax/20030108/isorelax-20030108.jar +0 -0
- data/lib/nokogiri/jruby/net/sf/saxon/Saxon-HE/9.6.0-4/Saxon-HE-9.6.0-4.jar +0 -0
- data/lib/nokogiri/jruby/net/sourceforge/htmlunit/neko-htmlunit/2.63.0/neko-htmlunit-2.63.0.jar +0 -0
- data/lib/nokogiri/jruby/nokogiri_jars.rb +43 -0
- data/lib/nokogiri/jruby/nu/validator/jing/20200702VNU/jing-20200702VNU.jar +0 -0
- data/lib/nokogiri/jruby/org/nokogiri/nekodtd/0.1.11.noko2/nekodtd-0.1.11.noko2.jar +0 -0
- data/lib/nokogiri/jruby/xalan/serializer/2.7.3/serializer-2.7.3.jar +0 -0
- data/lib/nokogiri/jruby/xalan/xalan/2.7.3/xalan-2.7.3.jar +0 -0
- data/lib/nokogiri/jruby/xerces/xercesImpl/2.12.2/xercesImpl-2.12.2.jar +0 -0
- data/lib/nokogiri/jruby/xml-apis/xml-apis/1.4.01/xml-apis-1.4.01.jar +0 -0
- data/lib/nokogiri/syntax_error.rb +6 -0
- data/lib/nokogiri/version/constant.rb +6 -0
- data/lib/nokogiri/version/info.rb +223 -0
- data/lib/nokogiri/version.rb +4 -0
- data/lib/nokogiri/xml/attr.rb +66 -0
- data/lib/nokogiri/xml/attribute_decl.rb +20 -0
- data/lib/nokogiri/xml/builder.rb +487 -0
- data/lib/nokogiri/xml/cdata.rb +13 -0
- data/lib/nokogiri/xml/character_data.rb +9 -0
- data/lib/nokogiri/xml/document.rb +471 -0
- data/lib/nokogiri/xml/document_fragment.rb +205 -0
- data/lib/nokogiri/xml/dtd.rb +34 -0
- data/lib/nokogiri/xml/element_content.rb +38 -0
- data/lib/nokogiri/xml/element_decl.rb +15 -0
- data/lib/nokogiri/xml/entity_decl.rb +21 -0
- data/lib/nokogiri/xml/entity_reference.rb +20 -0
- data/lib/nokogiri/xml/namespace.rb +58 -0
- data/lib/nokogiri/xml/node/save_options.rb +68 -0
- data/lib/nokogiri/xml/node.rb +1563 -0
- data/lib/nokogiri/xml/node_set.rb +447 -0
- data/lib/nokogiri/xml/notation.rb +19 -0
- data/lib/nokogiri/xml/parse_options.rb +213 -0
- data/lib/nokogiri/xml/pp/character_data.rb +21 -0
- data/lib/nokogiri/xml/pp/node.rb +57 -0
- data/lib/nokogiri/xml/pp.rb +4 -0
- data/lib/nokogiri/xml/processing_instruction.rb +11 -0
- data/lib/nokogiri/xml/reader.rb +105 -0
- data/lib/nokogiri/xml/relax_ng.rb +38 -0
- data/lib/nokogiri/xml/sax/document.rb +167 -0
- data/lib/nokogiri/xml/sax/parser.rb +125 -0
- data/lib/nokogiri/xml/sax/parser_context.rb +21 -0
- data/lib/nokogiri/xml/sax/push_parser.rb +61 -0
- data/lib/nokogiri/xml/sax.rb +6 -0
- data/lib/nokogiri/xml/schema.rb +73 -0
- data/lib/nokogiri/xml/searchable.rb +270 -0
- data/lib/nokogiri/xml/syntax_error.rb +72 -0
- data/lib/nokogiri/xml/text.rb +11 -0
- data/lib/nokogiri/xml/xpath/syntax_error.rb +13 -0
- data/lib/nokogiri/xml/xpath.rb +21 -0
- data/lib/nokogiri/xml/xpath_context.rb +16 -0
- data/lib/nokogiri/xml.rb +76 -0
- data/lib/nokogiri/xslt/stylesheet.rb +27 -0
- data/lib/nokogiri/xslt.rb +65 -0
- data/lib/nokogiri.rb +120 -0
- data/lib/xsd/xmlparser/nokogiri.rb +106 -0
- metadata +391 -0
@@ -0,0 +1,112 @@
|
|
1
|
+
#ifndef GUMBO_TOKENIZER_H_
|
2
|
+
#define GUMBO_TOKENIZER_H_
|
3
|
+
|
4
|
+
// This contains an implementation of a tokenizer for HTML5. It consumes a
|
5
|
+
// buffer of UTF-8 characters, and then emits a stream of tokens.
|
6
|
+
|
7
|
+
#include <stdbool.h>
|
8
|
+
#include <stddef.h>
|
9
|
+
|
10
|
+
#include "nokogiri_gumbo.h"
|
11
|
+
#include "token_type.h"
|
12
|
+
#include "tokenizer_states.h"
|
13
|
+
|
14
|
+
#ifdef __cplusplus
|
15
|
+
extern "C" {
|
16
|
+
#endif
|
17
|
+
|
18
|
+
struct GumboInternalParser;
|
19
|
+
|
20
|
+
// Struct containing all information pertaining to doctype tokens.
|
21
|
+
typedef struct GumboInternalTokenDocType {
|
22
|
+
const char* name;
|
23
|
+
const char* public_identifier;
|
24
|
+
const char* system_identifier;
|
25
|
+
bool force_quirks;
|
26
|
+
// There's no way to tell a 0-length public or system ID apart from the
|
27
|
+
// absence of a public or system ID, but they're handled different by the
|
28
|
+
// spec, so we need bool flags for them.
|
29
|
+
bool has_public_identifier;
|
30
|
+
bool has_system_identifier;
|
31
|
+
} GumboTokenDocType;
|
32
|
+
|
33
|
+
// Struct containing all information pertaining to start tag tokens.
|
34
|
+
typedef struct GumboInternalTokenStartTag {
|
35
|
+
GumboTag tag;
|
36
|
+
// NULL unless tag is GUMBO_TAG_UNKNOWN
|
37
|
+
char *name;
|
38
|
+
GumboVector /* GumboAttribute */ attributes;
|
39
|
+
bool is_self_closing;
|
40
|
+
} GumboTokenStartTag;
|
41
|
+
|
42
|
+
// Struct containing all information pertaining to end tag tokens.
|
43
|
+
typedef struct GumboInternalTokenEndTag {
|
44
|
+
GumboTag tag;
|
45
|
+
// NULL unless tag is GUMBO_TAG_UNKNOWN
|
46
|
+
char *name;
|
47
|
+
} GumboTokenEndTag;
|
48
|
+
|
49
|
+
// A data structure representing a single token in the input stream. This
|
50
|
+
// contains an enum for the type, the source position, a GumboStringPiece
|
51
|
+
// pointing to the original text, and then a union for any parsed data.
|
52
|
+
typedef struct GumboInternalToken {
|
53
|
+
GumboTokenType type;
|
54
|
+
GumboSourcePosition position;
|
55
|
+
GumboStringPiece original_text;
|
56
|
+
union {
|
57
|
+
GumboTokenDocType doc_type;
|
58
|
+
GumboTokenStartTag start_tag;
|
59
|
+
GumboTokenEndTag end_tag;
|
60
|
+
const char* text; // For comments.
|
61
|
+
int character; // For character, whitespace, null, and EOF tokens.
|
62
|
+
} v;
|
63
|
+
} GumboToken;
|
64
|
+
|
65
|
+
// Initializes the tokenizer state within the GumboParser object, setting up a
|
66
|
+
// parse of the specified text.
|
67
|
+
void gumbo_tokenizer_state_init (
|
68
|
+
struct GumboInternalParser* parser,
|
69
|
+
const char* text,
|
70
|
+
size_t text_length
|
71
|
+
);
|
72
|
+
|
73
|
+
// Destroys the tokenizer state within the GumboParser object, freeing any
|
74
|
+
// dynamically-allocated structures within it.
|
75
|
+
void gumbo_tokenizer_state_destroy(struct GumboInternalParser* parser);
|
76
|
+
|
77
|
+
// Sets the tokenizer state to the specified value. This is needed by some
|
78
|
+
// parser states, which alter the state of the tokenizer in response to tags
|
79
|
+
// seen.
|
80
|
+
void gumbo_tokenizer_set_state (
|
81
|
+
struct GumboInternalParser* parser,
|
82
|
+
GumboTokenizerEnum state
|
83
|
+
);
|
84
|
+
|
85
|
+
// Flags whether the adjusted current node is a foreign content element. This
|
86
|
+
// is necessary for the markup declaration open state, where the tokenizer
|
87
|
+
// must be aware of the state of the parser to properly tokenize bad comment
|
88
|
+
// tags.
|
89
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#markup-declaration-open-state
|
90
|
+
void gumbo_tokenizer_set_is_adjusted_current_node_foreign (
|
91
|
+
struct GumboInternalParser* parser,
|
92
|
+
bool is_foreign
|
93
|
+
);
|
94
|
+
|
95
|
+
// Lexes a single token from the specified buffer, filling the output with the
|
96
|
+
// parsed GumboToken data structure.
|
97
|
+
void gumbo_lex(struct GumboInternalParser* parser, GumboToken* output);
|
98
|
+
|
99
|
+
// Frees the internally-allocated pointers within a GumboToken. Note that this
|
100
|
+
// doesn't free the token itself, since oftentimes it will be allocated on the
|
101
|
+
// stack.
|
102
|
+
//
|
103
|
+
// Note that if you are handing over ownership of the internal strings to some
|
104
|
+
// other data structure - for example, a parse tree - these do not need to be
|
105
|
+
// freed.
|
106
|
+
void gumbo_token_destroy(GumboToken* token);
|
107
|
+
|
108
|
+
#ifdef __cplusplus
|
109
|
+
}
|
110
|
+
#endif
|
111
|
+
|
112
|
+
#endif // GUMBO_TOKENIZER_H_
|
@@ -0,0 +1,339 @@
|
|
1
|
+
#ifndef GUMBO_TOKENIZER_STATES_H_
|
2
|
+
#define GUMBO_TOKENIZER_STATES_H_
|
3
|
+
|
4
|
+
// This contains the list of states used in the tokenizer. Although at first
|
5
|
+
// glance it seems like these could be kept internal to the tokenizer, several
|
6
|
+
// of the actions in the parser require that it reach into the tokenizer and
|
7
|
+
// reset the tokenizer state. For that to work, it needs to have the
|
8
|
+
// definitions of individual states available.
|
9
|
+
//
|
10
|
+
// This may also be useful for providing more detailed error messages for parse
|
11
|
+
// errors, as we can match up states and inputs in a table without having to
|
12
|
+
// clutter the tokenizer code with lots of precise error messages.
|
13
|
+
|
14
|
+
// The ordering of this enum is also used to build the dispatch table for the
|
15
|
+
// tokenizer state machine, so if it is changed, be sure to update that too.
|
16
|
+
typedef enum {
|
17
|
+
// 12.2.5.1 Data state
|
18
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#data-state
|
19
|
+
GUMBO_LEX_DATA,
|
20
|
+
|
21
|
+
// 12.2.5.2 RCDATA state
|
22
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rcdata-state
|
23
|
+
GUMBO_LEX_RCDATA,
|
24
|
+
|
25
|
+
// 12.2.5.3 RAWTEXT state
|
26
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rawtext-state<Paste>
|
27
|
+
GUMBO_LEX_RAWTEXT,
|
28
|
+
|
29
|
+
// 12.2.5.4 Script data state
|
30
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-state
|
31
|
+
GUMBO_LEX_SCRIPT_DATA,
|
32
|
+
|
33
|
+
// 12.2.5.5 PLAINTEXT state
|
34
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#plaintext-state
|
35
|
+
GUMBO_LEX_PLAINTEXT,
|
36
|
+
|
37
|
+
// 12.2.5.6 Tag open state
|
38
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#tag-open-state
|
39
|
+
GUMBO_LEX_TAG_OPEN,
|
40
|
+
|
41
|
+
// 12.2.5.7 End tag open state
|
42
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#end-tag-open-state
|
43
|
+
GUMBO_LEX_END_TAG_OPEN,
|
44
|
+
|
45
|
+
// 12.2.5.8 Tag name state
|
46
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#tag-name-state
|
47
|
+
GUMBO_LEX_TAG_NAME,
|
48
|
+
|
49
|
+
// 12.2.5.9 RCDATA less-than sign state
|
50
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rcdata-less-than-sign-state
|
51
|
+
GUMBO_LEX_RCDATA_LT,
|
52
|
+
|
53
|
+
// 12.2.5.10 RCDATA end tag open state
|
54
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rcdata-end-tag-open-state
|
55
|
+
GUMBO_LEX_RCDATA_END_TAG_OPEN,
|
56
|
+
|
57
|
+
// 12.2.5.11 RCDATA end tag name state
|
58
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rcdata-end-tag-name-state
|
59
|
+
GUMBO_LEX_RCDATA_END_TAG_NAME,
|
60
|
+
|
61
|
+
// 12.2.5.12 RAWTEXT less-than sign state
|
62
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rawtext-less-than-sign-state
|
63
|
+
GUMBO_LEX_RAWTEXT_LT,
|
64
|
+
|
65
|
+
// 12.2.5.13 RAWTEXT end tag open state
|
66
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rawtext-end-tag-open-state
|
67
|
+
GUMBO_LEX_RAWTEXT_END_TAG_OPEN,
|
68
|
+
|
69
|
+
// 12.2.5.14 RAWTEXT end tag name state
|
70
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#rawtext-end-tag-name-state
|
71
|
+
GUMBO_LEX_RAWTEXT_END_TAG_NAME,
|
72
|
+
|
73
|
+
// 12.2.5.15 Script data less-than sign state
|
74
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-less-than-sign-state
|
75
|
+
GUMBO_LEX_SCRIPT_DATA_LT,
|
76
|
+
|
77
|
+
// 12.2.5.16 Script data end tag open state
|
78
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-end-tag-open-state
|
79
|
+
GUMBO_LEX_SCRIPT_DATA_END_TAG_OPEN,
|
80
|
+
|
81
|
+
// 12.2.5.17 Script data end tag name state
|
82
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-end-tag-name-state
|
83
|
+
GUMBO_LEX_SCRIPT_DATA_END_TAG_NAME,
|
84
|
+
|
85
|
+
// 12.2.5.18 Script data escape start state
|
86
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escape-start-state
|
87
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED_START,
|
88
|
+
|
89
|
+
// 12.2.5.19 Script data escape start dash state
|
90
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escape-start-dash-state
|
91
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED_START_DASH,
|
92
|
+
|
93
|
+
// 12.2.5.20 Script data escaped state
|
94
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escaped-state
|
95
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED,
|
96
|
+
|
97
|
+
// 12.2.5.21 Script data escaped dash state
|
98
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escaped-dash-state
|
99
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED_DASH,
|
100
|
+
|
101
|
+
// 12.2.5.22 Script data escaped dash dash state
|
102
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escaped-dash-dash-state
|
103
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED_DASH_DASH,
|
104
|
+
|
105
|
+
// 12.2.5.23 Script data escaped less than sign state
|
106
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escaped-less-than-sign-state
|
107
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED_LT,
|
108
|
+
|
109
|
+
// 12.2.5.24 Script data escaped end tag open state
|
110
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escaped-end-tag-open-state
|
111
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED_END_TAG_OPEN,
|
112
|
+
|
113
|
+
// 12.2.5.25 Script data escaped end tag name state
|
114
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-escaped-end-tag-name-state
|
115
|
+
GUMBO_LEX_SCRIPT_DATA_ESCAPED_END_TAG_NAME,
|
116
|
+
|
117
|
+
// 12.2.5.26 Script data double escape start state
|
118
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-double-escape-start-state
|
119
|
+
GUMBO_LEX_SCRIPT_DATA_DOUBLE_ESCAPED_START,
|
120
|
+
|
121
|
+
// 12.2.5.27 Script data double escaped state
|
122
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-double-escaped-state
|
123
|
+
GUMBO_LEX_SCRIPT_DATA_DOUBLE_ESCAPED,
|
124
|
+
|
125
|
+
// 12.2.5.28 Script data double escaped dash state
|
126
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-double-escaped-dash-state
|
127
|
+
GUMBO_LEX_SCRIPT_DATA_DOUBLE_ESCAPED_DASH,
|
128
|
+
|
129
|
+
// 12.2.5.29 Script data double escaped dash dash state
|
130
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-double-escaped-dash-dash-state
|
131
|
+
GUMBO_LEX_SCRIPT_DATA_DOUBLE_ESCAPED_DASH_DASH,
|
132
|
+
|
133
|
+
// 12.2.5.30 Script data double escaped less-than sign state
|
134
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-double-escaped-less-than-sign-state
|
135
|
+
GUMBO_LEX_SCRIPT_DATA_DOUBLE_ESCAPED_LT,
|
136
|
+
|
137
|
+
// 12.2.5.31 Script data double escape end state (XXX: spec bug with the
|
138
|
+
// name?)
|
139
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#script-data-double-escape-end-state
|
140
|
+
GUMBO_LEX_SCRIPT_DATA_DOUBLE_ESCAPED_END,
|
141
|
+
|
142
|
+
// 12.2.5.32 Before attribute name state
|
143
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#before-attribute-name-state
|
144
|
+
GUMBO_LEX_BEFORE_ATTR_NAME,
|
145
|
+
|
146
|
+
// 12.2.5.33 Attributet name state
|
147
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#attribute-name-state
|
148
|
+
GUMBO_LEX_ATTR_NAME,
|
149
|
+
|
150
|
+
// 12.2.5.34 After attribute name state
|
151
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#after-attribute-name-state
|
152
|
+
GUMBO_LEX_AFTER_ATTR_NAME,
|
153
|
+
|
154
|
+
// 12.2.5.35 Before attribute value state
|
155
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#before-attribute-value-state
|
156
|
+
GUMBO_LEX_BEFORE_ATTR_VALUE,
|
157
|
+
|
158
|
+
// 12.2.5.36 Attribute value (double-quoted) state
|
159
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#attribute-value-(double-quoted)-state
|
160
|
+
GUMBO_LEX_ATTR_VALUE_DOUBLE_QUOTED,
|
161
|
+
|
162
|
+
// 12.2.5.37 Attribute value (single-quoted) state
|
163
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#attribute-value-(single-quoted)-state
|
164
|
+
GUMBO_LEX_ATTR_VALUE_SINGLE_QUOTED,
|
165
|
+
|
166
|
+
// 12.2.5.38 Attribute value (unquoted) state
|
167
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#attribute-value-(unquoted)-state
|
168
|
+
GUMBO_LEX_ATTR_VALUE_UNQUOTED,
|
169
|
+
|
170
|
+
// 12.2.5.39 After attribute value (quoted) state
|
171
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#after-attribute-value-(quoted)-state
|
172
|
+
GUMBO_LEX_AFTER_ATTR_VALUE_QUOTED,
|
173
|
+
|
174
|
+
// 12.2.5.40 Self-closing start tag state
|
175
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#self-closing-start-tag-state
|
176
|
+
GUMBO_LEX_SELF_CLOSING_START_TAG,
|
177
|
+
|
178
|
+
// 12.2.5.41 Bogus comment state
|
179
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#bogus-comment-state
|
180
|
+
GUMBO_LEX_BOGUS_COMMENT,
|
181
|
+
|
182
|
+
// 12.2.5.42 Markup declaration open state
|
183
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#markup-declaration-open-state
|
184
|
+
GUMBO_LEX_MARKUP_DECLARATION_OPEN,
|
185
|
+
|
186
|
+
// 12.2.5.43 Comment start state
|
187
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-start-state
|
188
|
+
GUMBO_LEX_COMMENT_START,
|
189
|
+
|
190
|
+
// 12.2.5.44 Comment start dash state
|
191
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-start-dash-state
|
192
|
+
GUMBO_LEX_COMMENT_START_DASH,
|
193
|
+
|
194
|
+
// 12.2.5.45 Comment state
|
195
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-state
|
196
|
+
GUMBO_LEX_COMMENT,
|
197
|
+
|
198
|
+
// 12.2.5.46 Comment less-than sign state
|
199
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-less-than-sign-state
|
200
|
+
GUMBO_LEX_COMMENT_LT,
|
201
|
+
|
202
|
+
// 12.2.5.47 Comment less-than sign bang state
|
203
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-less-than-sign-bang-state
|
204
|
+
GUMBO_LEX_COMMENT_LT_BANG,
|
205
|
+
|
206
|
+
// 12.2.5.48 Comment less-than sign bang dash state
|
207
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-less-than-sign-bang-dash-state
|
208
|
+
GUMBO_LEX_COMMENT_LT_BANG_DASH,
|
209
|
+
|
210
|
+
// 12.2.5.49 Comment less-than sign bang dash dash state
|
211
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-less-than-sign-bang-dash-dash-state
|
212
|
+
GUMBO_LEX_COMMENT_LT_BANG_DASH_DASH,
|
213
|
+
|
214
|
+
// 12.2.5.50 Comment end dash state
|
215
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-end-dash-state
|
216
|
+
GUMBO_LEX_COMMENT_END_DASH,
|
217
|
+
|
218
|
+
// 12.2.5.51 Comment end state
|
219
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-end-state
|
220
|
+
GUMBO_LEX_COMMENT_END,
|
221
|
+
|
222
|
+
// 12.2.5.52 Comment end bang state
|
223
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#comment-end-bang-state
|
224
|
+
GUMBO_LEX_COMMENT_END_BANG,
|
225
|
+
|
226
|
+
// 12.2.5.53 DOCTYPE state
|
227
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#doctype-state
|
228
|
+
GUMBO_LEX_DOCTYPE,
|
229
|
+
|
230
|
+
// 12.2.5.54 Before DOCTYPE name state
|
231
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#before-doctype-name-state
|
232
|
+
GUMBO_LEX_BEFORE_DOCTYPE_NAME,
|
233
|
+
|
234
|
+
// 12.2.5.55 DOCTYPE name state
|
235
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#doctype-name-state
|
236
|
+
GUMBO_LEX_DOCTYPE_NAME,
|
237
|
+
|
238
|
+
// 12.2.5.56 After DOCTYPE name state
|
239
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#after-doctype-name-state
|
240
|
+
GUMBO_LEX_AFTER_DOCTYPE_NAME,
|
241
|
+
|
242
|
+
// 12.2.5.57 After DOCTYPE public keyword state
|
243
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#after-doctype-public-keyword-state
|
244
|
+
GUMBO_LEX_AFTER_DOCTYPE_PUBLIC_KEYWORD,
|
245
|
+
|
246
|
+
// 12.2.5.58 Before DOCTYPE public identifier state
|
247
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#before-doctype-public-identifier-state
|
248
|
+
GUMBO_LEX_BEFORE_DOCTYPE_PUBLIC_ID,
|
249
|
+
|
250
|
+
// 12.2.5.59 DOCTYPE public identifier (double-quoted) state
|
251
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#doctype-public-identifier-(double-quoted)-state
|
252
|
+
GUMBO_LEX_DOCTYPE_PUBLIC_ID_DOUBLE_QUOTED,
|
253
|
+
|
254
|
+
// 12.2.5.60 DOCTYPE public identifier (single-quoted) state
|
255
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#doctype-public-identifier-(single-quoted)-state
|
256
|
+
GUMBO_LEX_DOCTYPE_PUBLIC_ID_SINGLE_QUOTED,
|
257
|
+
|
258
|
+
// 12.2.5.61 After DOCTYPE public identifier state
|
259
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#after-doctype-public-identifier-state
|
260
|
+
GUMBO_LEX_AFTER_DOCTYPE_PUBLIC_ID,
|
261
|
+
|
262
|
+
// 12.2.5.62 Between DOCTYPE public and system identifiers state
|
263
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#between-doctype-public-and-system-identifiers-state
|
264
|
+
GUMBO_LEX_BETWEEN_DOCTYPE_PUBLIC_SYSTEM_ID,
|
265
|
+
|
266
|
+
// 12.2.5.63 After DOCTYPE system keyword state
|
267
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#after-doctype-system-keyword-state
|
268
|
+
GUMBO_LEX_AFTER_DOCTYPE_SYSTEM_KEYWORD,
|
269
|
+
|
270
|
+
// 12.2.5.64 Before DOCTYPE system identifier state
|
271
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#before-doctype-system-identifier-state
|
272
|
+
GUMBO_LEX_BEFORE_DOCTYPE_SYSTEM_ID,
|
273
|
+
|
274
|
+
// 12.2.5.65 DOCTYPE system identifier (double-quoted) state
|
275
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#doctype-system-identifier-(double-quoted)-state
|
276
|
+
GUMBO_LEX_DOCTYPE_SYSTEM_ID_DOUBLE_QUOTED,
|
277
|
+
|
278
|
+
// 12.2.5.66 DOCTYPE system identifier (single-quoted) state
|
279
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#doctype-system-identifier-(single-quoted)-state
|
280
|
+
GUMBO_LEX_DOCTYPE_SYSTEM_ID_SINGLE_QUOTED,
|
281
|
+
|
282
|
+
// 12.2.5.67 After DOCTYPE system identifier state
|
283
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#after-doctype-system-identifier-state
|
284
|
+
GUMBO_LEX_AFTER_DOCTYPE_SYSTEM_ID,
|
285
|
+
|
286
|
+
// 12.2.5.68 Bogus DOCTYPE state
|
287
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#bogus-doctype-state
|
288
|
+
GUMBO_LEX_BOGUS_DOCTYPE,
|
289
|
+
|
290
|
+
// 12.2.5.69 CDATA section state
|
291
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#cdata-section-state
|
292
|
+
GUMBO_LEX_CDATA_SECTION,
|
293
|
+
|
294
|
+
// 12.2.5.70 CDATA section bracket state
|
295
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#cdata-section-bracket-state
|
296
|
+
GUMBO_LEX_CDATA_SECTION_BRACKET,
|
297
|
+
|
298
|
+
// 12.2.5.71 CDATA section end state
|
299
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#cdata-section-end-state
|
300
|
+
GUMBO_LEX_CDATA_SECTION_END,
|
301
|
+
|
302
|
+
// 12.2.5.72 Character reference state
|
303
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#character-reference-state
|
304
|
+
GUMBO_LEX_CHARACTER_REFERENCE,
|
305
|
+
|
306
|
+
// 12.2.5.73 Named character reference state
|
307
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#named-character-reference-state
|
308
|
+
GUMBO_LEX_NAMED_CHARACTER_REFERENCE,
|
309
|
+
|
310
|
+
// 12.2.5.74 Ambiguous ampersand state
|
311
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#ambiguous-ampersand-state
|
312
|
+
GUMBO_LEX_AMBIGUOUS_AMPERSAND,
|
313
|
+
|
314
|
+
// 12.2.5.75 Numeric character reference state
|
315
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#numeric-character-reference-state
|
316
|
+
GUMBO_LEX_NUMERIC_CHARACTER_REFERENCE,
|
317
|
+
|
318
|
+
// 12.2.5.76 Hexadecimal character reference start state
|
319
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#hexadecimal-character-reference-start-state
|
320
|
+
GUMBO_LEX_HEXADECIMAL_CHARACTER_REFERENCE_START,
|
321
|
+
|
322
|
+
// 12.2.5.77 Decimal character reference start state
|
323
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#decimal-character-reference-start-state
|
324
|
+
GUMBO_LEX_DECIMAL_CHARACTER_REFERENCE_START,
|
325
|
+
|
326
|
+
// 12.2.5.78 Hexadecimal character reference state
|
327
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#hexadecimal-character-reference-state
|
328
|
+
GUMBO_LEX_HEXADECIMAL_CHARACTER_REFERENCE,
|
329
|
+
|
330
|
+
// 12.2.5.79 Decimal character reference state
|
331
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#decimal-character-reference-state
|
332
|
+
GUMBO_LEX_DECIMAL_CHARACTER_REFERENCE,
|
333
|
+
|
334
|
+
// 12.2.5.80 Numeric character reference end state
|
335
|
+
// https://html.spec.whatwg.org/multipage/parsing.html#numeric-character-reference-end-state
|
336
|
+
GUMBO_LEX_NUMERIC_CHARACTER_REFERENCE_END
|
337
|
+
} GumboTokenizerEnum;
|
338
|
+
|
339
|
+
#endif // GUMBO_TOKENIZER_STATES_H_
|