Nokogiri_precompiled_aarch64_dedshit 1.14.5

Sign up to get free protection for your applications and to get access to all the features.
Files changed (263) hide show
  1. checksums.yaml +7 -0
  2. data/Gemfile +44 -0
  3. data/LICENSE-DEPENDENCIES.md +2224 -0
  4. data/LICENSE.md +9 -0
  5. data/README.md +287 -0
  6. data/bin/nokogiri +131 -0
  7. data/dependencies.yml +41 -0
  8. data/ext/java/nokogiri/Html4Document.java +157 -0
  9. data/ext/java/nokogiri/Html4ElementDescription.java +133 -0
  10. data/ext/java/nokogiri/Html4EntityLookup.java +63 -0
  11. data/ext/java/nokogiri/Html4SaxParserContext.java +289 -0
  12. data/ext/java/nokogiri/Html4SaxPushParser.java +213 -0
  13. data/ext/java/nokogiri/NokogiriService.java +613 -0
  14. data/ext/java/nokogiri/XmlAttr.java +154 -0
  15. data/ext/java/nokogiri/XmlAttributeDecl.java +119 -0
  16. data/ext/java/nokogiri/XmlCdata.java +60 -0
  17. data/ext/java/nokogiri/XmlComment.java +77 -0
  18. data/ext/java/nokogiri/XmlDocument.java +705 -0
  19. data/ext/java/nokogiri/XmlDocumentFragment.java +163 -0
  20. data/ext/java/nokogiri/XmlDtd.java +516 -0
  21. data/ext/java/nokogiri/XmlElement.java +44 -0
  22. data/ext/java/nokogiri/XmlElementContent.java +412 -0
  23. data/ext/java/nokogiri/XmlElementDecl.java +148 -0
  24. data/ext/java/nokogiri/XmlEntityDecl.java +151 -0
  25. data/ext/java/nokogiri/XmlEntityReference.java +79 -0
  26. data/ext/java/nokogiri/XmlNamespace.java +193 -0
  27. data/ext/java/nokogiri/XmlNode.java +1938 -0
  28. data/ext/java/nokogiri/XmlNodeSet.java +463 -0
  29. data/ext/java/nokogiri/XmlProcessingInstruction.java +79 -0
  30. data/ext/java/nokogiri/XmlReader.java +615 -0
  31. data/ext/java/nokogiri/XmlRelaxng.java +133 -0
  32. data/ext/java/nokogiri/XmlSaxParserContext.java +329 -0
  33. data/ext/java/nokogiri/XmlSaxPushParser.java +288 -0
  34. data/ext/java/nokogiri/XmlSchema.java +423 -0
  35. data/ext/java/nokogiri/XmlSyntaxError.java +137 -0
  36. data/ext/java/nokogiri/XmlText.java +90 -0
  37. data/ext/java/nokogiri/XmlXpathContext.java +305 -0
  38. data/ext/java/nokogiri/XsltStylesheet.java +368 -0
  39. data/ext/java/nokogiri/internals/ClosedStreamException.java +13 -0
  40. data/ext/java/nokogiri/internals/HtmlDomParserContext.java +252 -0
  41. data/ext/java/nokogiri/internals/IgnoreSchemaErrorsErrorHandler.java +27 -0
  42. data/ext/java/nokogiri/internals/NokogiriBlockingQueueInputStream.java +178 -0
  43. data/ext/java/nokogiri/internals/NokogiriDomParser.java +99 -0
  44. data/ext/java/nokogiri/internals/NokogiriEntityResolver.java +140 -0
  45. data/ext/java/nokogiri/internals/NokogiriErrorHandler.java +65 -0
  46. data/ext/java/nokogiri/internals/NokogiriHandler.java +339 -0
  47. data/ext/java/nokogiri/internals/NokogiriHelpers.java +817 -0
  48. data/ext/java/nokogiri/internals/NokogiriNamespaceCache.java +228 -0
  49. data/ext/java/nokogiri/internals/NokogiriNamespaceContext.java +110 -0
  50. data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler.java +86 -0
  51. data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler4NekoHtml.java +107 -0
  52. data/ext/java/nokogiri/internals/NokogiriStrictErrorHandler.java +62 -0
  53. data/ext/java/nokogiri/internals/NokogiriXPathFunction.java +165 -0
  54. data/ext/java/nokogiri/internals/NokogiriXPathFunctionResolver.java +50 -0
  55. data/ext/java/nokogiri/internals/NokogiriXPathVariableResolver.java +37 -0
  56. data/ext/java/nokogiri/internals/NokogiriXsltErrorListener.java +70 -0
  57. data/ext/java/nokogiri/internals/ParserContext.java +262 -0
  58. data/ext/java/nokogiri/internals/ReaderNode.java +564 -0
  59. data/ext/java/nokogiri/internals/SaveContextVisitor.java +865 -0
  60. data/ext/java/nokogiri/internals/SchemaErrorHandler.java +50 -0
  61. data/ext/java/nokogiri/internals/XalanDTMManagerPatch.java +174 -0
  62. data/ext/java/nokogiri/internals/XmlDeclHandler.java +11 -0
  63. data/ext/java/nokogiri/internals/XmlDomParserContext.java +265 -0
  64. data/ext/java/nokogiri/internals/XmlSaxParser.java +40 -0
  65. data/ext/java/nokogiri/internals/c14n/AttrCompare.java +122 -0
  66. data/ext/java/nokogiri/internals/c14n/C14nHelper.java +178 -0
  67. data/ext/java/nokogiri/internals/c14n/CanonicalFilter.java +43 -0
  68. data/ext/java/nokogiri/internals/c14n/CanonicalizationException.java +106 -0
  69. data/ext/java/nokogiri/internals/c14n/Canonicalizer.java +278 -0
  70. data/ext/java/nokogiri/internals/c14n/Canonicalizer11.java +664 -0
  71. data/ext/java/nokogiri/internals/c14n/Canonicalizer11_OmitComments.java +45 -0
  72. data/ext/java/nokogiri/internals/c14n/Canonicalizer11_WithComments.java +45 -0
  73. data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315.java +388 -0
  74. data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315Excl.java +308 -0
  75. data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315ExclOmitComments.java +47 -0
  76. data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315ExclWithComments.java +51 -0
  77. data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315OmitComments.java +51 -0
  78. data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315WithComments.java +50 -0
  79. data/ext/java/nokogiri/internals/c14n/CanonicalizerBase.java +660 -0
  80. data/ext/java/nokogiri/internals/c14n/CanonicalizerPhysical.java +194 -0
  81. data/ext/java/nokogiri/internals/c14n/CanonicalizerSpi.java +77 -0
  82. data/ext/java/nokogiri/internals/c14n/Constants.java +45 -0
  83. data/ext/java/nokogiri/internals/c14n/ElementProxy.java +325 -0
  84. data/ext/java/nokogiri/internals/c14n/HelperNodeList.java +106 -0
  85. data/ext/java/nokogiri/internals/c14n/IgnoreAllErrorHandler.java +86 -0
  86. data/ext/java/nokogiri/internals/c14n/InclusiveNamespaces.java +181 -0
  87. data/ext/java/nokogiri/internals/c14n/InvalidCanonicalizerException.java +87 -0
  88. data/ext/java/nokogiri/internals/c14n/NameSpaceSymbTable.java +452 -0
  89. data/ext/java/nokogiri/internals/c14n/NodeFilter.java +52 -0
  90. data/ext/java/nokogiri/internals/c14n/UtfHelpper.java +190 -0
  91. data/ext/java/nokogiri/internals/c14n/XMLUtils.java +540 -0
  92. data/ext/java/nokogiri/internals/dom2dtm/DOM2DTM.java +1712 -0
  93. data/ext/java/nokogiri/internals/dom2dtm/DOM2DTMdefaultNamespaceDeclarationNode.java +737 -0
  94. data/ext/nokogiri/depend +38 -0
  95. data/ext/nokogiri/extconf.rb +1086 -0
  96. data/ext/nokogiri/gumbo.c +594 -0
  97. data/ext/nokogiri/html4_document.c +167 -0
  98. data/ext/nokogiri/html4_element_description.c +294 -0
  99. data/ext/nokogiri/html4_entity_lookup.c +37 -0
  100. data/ext/nokogiri/html4_sax_parser_context.c +116 -0
  101. data/ext/nokogiri/html4_sax_push_parser.c +95 -0
  102. data/ext/nokogiri/libxml2_backwards_compat.c +121 -0
  103. data/ext/nokogiri/nokogiri.c +265 -0
  104. data/ext/nokogiri/nokogiri.h +235 -0
  105. data/ext/nokogiri/test_global_handlers.c +42 -0
  106. data/ext/nokogiri/xml_attr.c +103 -0
  107. data/ext/nokogiri/xml_attribute_decl.c +70 -0
  108. data/ext/nokogiri/xml_cdata.c +57 -0
  109. data/ext/nokogiri/xml_comment.c +62 -0
  110. data/ext/nokogiri/xml_document.c +689 -0
  111. data/ext/nokogiri/xml_document_fragment.c +44 -0
  112. data/ext/nokogiri/xml_dtd.c +210 -0
  113. data/ext/nokogiri/xml_element_content.c +128 -0
  114. data/ext/nokogiri/xml_element_decl.c +69 -0
  115. data/ext/nokogiri/xml_encoding_handler.c +104 -0
  116. data/ext/nokogiri/xml_entity_decl.c +112 -0
  117. data/ext/nokogiri/xml_entity_reference.c +50 -0
  118. data/ext/nokogiri/xml_namespace.c +186 -0
  119. data/ext/nokogiri/xml_node.c +2426 -0
  120. data/ext/nokogiri/xml_node_set.c +496 -0
  121. data/ext/nokogiri/xml_processing_instruction.c +54 -0
  122. data/ext/nokogiri/xml_reader.c +794 -0
  123. data/ext/nokogiri/xml_relax_ng.c +164 -0
  124. data/ext/nokogiri/xml_sax_parser.c +316 -0
  125. data/ext/nokogiri/xml_sax_parser_context.c +283 -0
  126. data/ext/nokogiri/xml_sax_push_parser.c +166 -0
  127. data/ext/nokogiri/xml_schema.c +260 -0
  128. data/ext/nokogiri/xml_syntax_error.c +85 -0
  129. data/ext/nokogiri/xml_text.c +48 -0
  130. data/ext/nokogiri/xml_xpath_context.c +415 -0
  131. data/ext/nokogiri/xslt_stylesheet.c +363 -0
  132. data/gumbo-parser/CHANGES.md +63 -0
  133. data/gumbo-parser/Makefile +111 -0
  134. data/gumbo-parser/THANKS +27 -0
  135. data/gumbo-parser/src/Makefile +34 -0
  136. data/gumbo-parser/src/README.md +41 -0
  137. data/gumbo-parser/src/ascii.c +75 -0
  138. data/gumbo-parser/src/ascii.h +115 -0
  139. data/gumbo-parser/src/attribute.c +42 -0
  140. data/gumbo-parser/src/attribute.h +17 -0
  141. data/gumbo-parser/src/char_ref.c +22225 -0
  142. data/gumbo-parser/src/char_ref.h +29 -0
  143. data/gumbo-parser/src/char_ref.rl +2154 -0
  144. data/gumbo-parser/src/error.c +626 -0
  145. data/gumbo-parser/src/error.h +148 -0
  146. data/gumbo-parser/src/foreign_attrs.c +104 -0
  147. data/gumbo-parser/src/foreign_attrs.gperf +27 -0
  148. data/gumbo-parser/src/insertion_mode.h +33 -0
  149. data/gumbo-parser/src/macros.h +91 -0
  150. data/gumbo-parser/src/nokogiri_gumbo.h +944 -0
  151. data/gumbo-parser/src/parser.c +4878 -0
  152. data/gumbo-parser/src/parser.h +41 -0
  153. data/gumbo-parser/src/replacement.h +33 -0
  154. data/gumbo-parser/src/string_buffer.c +103 -0
  155. data/gumbo-parser/src/string_buffer.h +68 -0
  156. data/gumbo-parser/src/string_piece.c +48 -0
  157. data/gumbo-parser/src/svg_attrs.c +174 -0
  158. data/gumbo-parser/src/svg_attrs.gperf +77 -0
  159. data/gumbo-parser/src/svg_tags.c +137 -0
  160. data/gumbo-parser/src/svg_tags.gperf +55 -0
  161. data/gumbo-parser/src/tag.c +223 -0
  162. data/gumbo-parser/src/tag_lookup.c +382 -0
  163. data/gumbo-parser/src/tag_lookup.gperf +170 -0
  164. data/gumbo-parser/src/tag_lookup.h +13 -0
  165. data/gumbo-parser/src/token_buffer.c +79 -0
  166. data/gumbo-parser/src/token_buffer.h +71 -0
  167. data/gumbo-parser/src/token_type.h +17 -0
  168. data/gumbo-parser/src/tokenizer.c +3463 -0
  169. data/gumbo-parser/src/tokenizer.h +112 -0
  170. data/gumbo-parser/src/tokenizer_states.h +339 -0
  171. data/gumbo-parser/src/utf8.c +245 -0
  172. data/gumbo-parser/src/utf8.h +164 -0
  173. data/gumbo-parser/src/util.c +66 -0
  174. data/gumbo-parser/src/util.h +34 -0
  175. data/gumbo-parser/src/vector.c +111 -0
  176. data/gumbo-parser/src/vector.h +45 -0
  177. data/lib/nokogiri/class_resolver.rb +67 -0
  178. data/lib/nokogiri/css/node.rb +54 -0
  179. data/lib/nokogiri/css/parser.rb +770 -0
  180. data/lib/nokogiri/css/parser.y +277 -0
  181. data/lib/nokogiri/css/parser_extras.rb +96 -0
  182. data/lib/nokogiri/css/syntax_error.rb +9 -0
  183. data/lib/nokogiri/css/tokenizer.rb +155 -0
  184. data/lib/nokogiri/css/tokenizer.rex +56 -0
  185. data/lib/nokogiri/css/xpath_visitor.rb +359 -0
  186. data/lib/nokogiri/css.rb +66 -0
  187. data/lib/nokogiri/decorators/slop.rb +44 -0
  188. data/lib/nokogiri/encoding_handler.rb +57 -0
  189. data/lib/nokogiri/extension.rb +32 -0
  190. data/lib/nokogiri/gumbo.rb +15 -0
  191. data/lib/nokogiri/html.rb +48 -0
  192. data/lib/nokogiri/html4/builder.rb +37 -0
  193. data/lib/nokogiri/html4/document.rb +214 -0
  194. data/lib/nokogiri/html4/document_fragment.rb +54 -0
  195. data/lib/nokogiri/html4/element_description.rb +25 -0
  196. data/lib/nokogiri/html4/element_description_defaults.rb +572 -0
  197. data/lib/nokogiri/html4/encoding_reader.rb +121 -0
  198. data/lib/nokogiri/html4/entity_lookup.rb +15 -0
  199. data/lib/nokogiri/html4/sax/parser.rb +63 -0
  200. data/lib/nokogiri/html4/sax/parser_context.rb +20 -0
  201. data/lib/nokogiri/html4/sax/push_parser.rb +37 -0
  202. data/lib/nokogiri/html4.rb +47 -0
  203. data/lib/nokogiri/html5/document.rb +168 -0
  204. data/lib/nokogiri/html5/document_fragment.rb +90 -0
  205. data/lib/nokogiri/html5/node.rb +98 -0
  206. data/lib/nokogiri/html5.rb +389 -0
  207. data/lib/nokogiri/jruby/dependencies.rb +3 -0
  208. data/lib/nokogiri/jruby/isorelax/isorelax/20030108/isorelax-20030108.jar +0 -0
  209. data/lib/nokogiri/jruby/net/sf/saxon/Saxon-HE/9.6.0-4/Saxon-HE-9.6.0-4.jar +0 -0
  210. data/lib/nokogiri/jruby/net/sourceforge/htmlunit/neko-htmlunit/2.63.0/neko-htmlunit-2.63.0.jar +0 -0
  211. data/lib/nokogiri/jruby/nokogiri_jars.rb +43 -0
  212. data/lib/nokogiri/jruby/nu/validator/jing/20200702VNU/jing-20200702VNU.jar +0 -0
  213. data/lib/nokogiri/jruby/org/nokogiri/nekodtd/0.1.11.noko2/nekodtd-0.1.11.noko2.jar +0 -0
  214. data/lib/nokogiri/jruby/xalan/serializer/2.7.3/serializer-2.7.3.jar +0 -0
  215. data/lib/nokogiri/jruby/xalan/xalan/2.7.3/xalan-2.7.3.jar +0 -0
  216. data/lib/nokogiri/jruby/xerces/xercesImpl/2.12.2/xercesImpl-2.12.2.jar +0 -0
  217. data/lib/nokogiri/jruby/xml-apis/xml-apis/1.4.01/xml-apis-1.4.01.jar +0 -0
  218. data/lib/nokogiri/syntax_error.rb +6 -0
  219. data/lib/nokogiri/version/constant.rb +6 -0
  220. data/lib/nokogiri/version/info.rb +223 -0
  221. data/lib/nokogiri/version.rb +4 -0
  222. data/lib/nokogiri/xml/attr.rb +66 -0
  223. data/lib/nokogiri/xml/attribute_decl.rb +20 -0
  224. data/lib/nokogiri/xml/builder.rb +487 -0
  225. data/lib/nokogiri/xml/cdata.rb +13 -0
  226. data/lib/nokogiri/xml/character_data.rb +9 -0
  227. data/lib/nokogiri/xml/document.rb +471 -0
  228. data/lib/nokogiri/xml/document_fragment.rb +205 -0
  229. data/lib/nokogiri/xml/dtd.rb +34 -0
  230. data/lib/nokogiri/xml/element_content.rb +38 -0
  231. data/lib/nokogiri/xml/element_decl.rb +15 -0
  232. data/lib/nokogiri/xml/entity_decl.rb +21 -0
  233. data/lib/nokogiri/xml/entity_reference.rb +20 -0
  234. data/lib/nokogiri/xml/namespace.rb +58 -0
  235. data/lib/nokogiri/xml/node/save_options.rb +68 -0
  236. data/lib/nokogiri/xml/node.rb +1563 -0
  237. data/lib/nokogiri/xml/node_set.rb +447 -0
  238. data/lib/nokogiri/xml/notation.rb +19 -0
  239. data/lib/nokogiri/xml/parse_options.rb +213 -0
  240. data/lib/nokogiri/xml/pp/character_data.rb +21 -0
  241. data/lib/nokogiri/xml/pp/node.rb +57 -0
  242. data/lib/nokogiri/xml/pp.rb +4 -0
  243. data/lib/nokogiri/xml/processing_instruction.rb +11 -0
  244. data/lib/nokogiri/xml/reader.rb +105 -0
  245. data/lib/nokogiri/xml/relax_ng.rb +38 -0
  246. data/lib/nokogiri/xml/sax/document.rb +167 -0
  247. data/lib/nokogiri/xml/sax/parser.rb +125 -0
  248. data/lib/nokogiri/xml/sax/parser_context.rb +21 -0
  249. data/lib/nokogiri/xml/sax/push_parser.rb +61 -0
  250. data/lib/nokogiri/xml/sax.rb +6 -0
  251. data/lib/nokogiri/xml/schema.rb +73 -0
  252. data/lib/nokogiri/xml/searchable.rb +270 -0
  253. data/lib/nokogiri/xml/syntax_error.rb +72 -0
  254. data/lib/nokogiri/xml/text.rb +11 -0
  255. data/lib/nokogiri/xml/xpath/syntax_error.rb +13 -0
  256. data/lib/nokogiri/xml/xpath.rb +21 -0
  257. data/lib/nokogiri/xml/xpath_context.rb +16 -0
  258. data/lib/nokogiri/xml.rb +76 -0
  259. data/lib/nokogiri/xslt/stylesheet.rb +27 -0
  260. data/lib/nokogiri/xslt.rb +65 -0
  261. data/lib/nokogiri.rb +120 -0
  262. data/lib/xsd/xmlparser/nokogiri.rb +106 -0
  263. metadata +391 -0
@@ -0,0 +1,57 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ # :nodoc: all
6
+ module PP
7
+ module Node
8
+ COLLECTIONS = [:attribute_nodes, :children]
9
+
10
+ def inspect
11
+ attributes = inspect_attributes.reject do |x|
12
+ attribute = send(x)
13
+ !attribute || (attribute.respond_to?(:empty?) && attribute.empty?)
14
+ rescue NoMethodError
15
+ true
16
+ end.map do |attribute|
17
+ "#{attribute.to_s.sub(/_\w+/, "s")}=#{send(attribute).inspect}"
18
+ end.join(" ")
19
+ "#<#{self.class.name}:#{format("0x%x", object_id)} #{attributes}>"
20
+ end
21
+
22
+ def pretty_print(pp)
23
+ nice_name = self.class.name.split("::").last
24
+ pp.group(2, "#(#{nice_name}:#{format("0x%x", object_id)} {", "})") do
25
+ pp.breakable
26
+ attrs = inspect_attributes.filter_map do |t|
27
+ [t, send(t)] if respond_to?(t)
28
+ end.find_all do |x|
29
+ if x.last
30
+ if COLLECTIONS.include?(x.first)
31
+ !x.last.empty?
32
+ else
33
+ true
34
+ end
35
+ end
36
+ end
37
+
38
+ pp.seplist(attrs) do |v|
39
+ if COLLECTIONS.include?(v.first)
40
+ pp.group(2, "#{v.first.to_s.sub(/_\w+$/, "s")} = [", "]") do
41
+ pp.breakable
42
+ pp.seplist(v.last) do |item|
43
+ pp.pp(item)
44
+ end
45
+ end
46
+ else
47
+ pp.text("#{v.first} = ")
48
+ pp.pp(v.last)
49
+ end
50
+ end
51
+ pp.breakable
52
+ end
53
+ end
54
+ end
55
+ end
56
+ end
57
+ end
@@ -0,0 +1,4 @@
1
+ # frozen_string_literal: true
2
+
3
+ require_relative "pp/node"
4
+ require_relative "pp/character_data"
@@ -0,0 +1,11 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ class ProcessingInstruction < Node
6
+ def initialize(document, name, content)
7
+ super(document, name)
8
+ end
9
+ end
10
+ end
11
+ end
@@ -0,0 +1,105 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ ###
6
+ # Nokogiri::XML::Reader parses an XML document similar to the way a cursor
7
+ # would move. The Reader is given an XML document, and yields nodes
8
+ # to an each block.
9
+ #
10
+ # Here is an example of usage:
11
+ #
12
+ # reader = Nokogiri::XML::Reader(<<-eoxml)
13
+ # <x xmlns:tenderlove='http://tenderlovemaking.com/'>
14
+ # <tenderlove:foo awesome='true'>snuggles!</tenderlove:foo>
15
+ # </x>
16
+ # eoxml
17
+ #
18
+ # reader.each do |node|
19
+ #
20
+ # # node is an instance of Nokogiri::XML::Reader
21
+ # puts node.name
22
+ #
23
+ # end
24
+ #
25
+ # Note that Nokogiri::XML::Reader#each can only be called once!! Once
26
+ # the cursor moves through the entire document, you must parse the
27
+ # document again. So make sure that you capture any information you
28
+ # need during the first iteration.
29
+ #
30
+ # The Reader parser is good for when you need the speed of a SAX parser,
31
+ # but do not want to write a Document handler.
32
+ class Reader
33
+ include Enumerable
34
+
35
+ TYPE_NONE = 0
36
+ # Element node type
37
+ TYPE_ELEMENT = 1
38
+ # Attribute node type
39
+ TYPE_ATTRIBUTE = 2
40
+ # Text node type
41
+ TYPE_TEXT = 3
42
+ # CDATA node type
43
+ TYPE_CDATA = 4
44
+ # Entity Reference node type
45
+ TYPE_ENTITY_REFERENCE = 5
46
+ # Entity node type
47
+ TYPE_ENTITY = 6
48
+ # PI node type
49
+ TYPE_PROCESSING_INSTRUCTION = 7
50
+ # Comment node type
51
+ TYPE_COMMENT = 8
52
+ # Document node type
53
+ TYPE_DOCUMENT = 9
54
+ # Document Type node type
55
+ TYPE_DOCUMENT_TYPE = 10
56
+ # Document Fragment node type
57
+ TYPE_DOCUMENT_FRAGMENT = 11
58
+ # Notation node type
59
+ TYPE_NOTATION = 12
60
+ # Whitespace node type
61
+ TYPE_WHITESPACE = 13
62
+ # Significant Whitespace node type
63
+ TYPE_SIGNIFICANT_WHITESPACE = 14
64
+ # Element end node type
65
+ TYPE_END_ELEMENT = 15
66
+ # Entity end node type
67
+ TYPE_END_ENTITY = 16
68
+ # XML Declaration node type
69
+ TYPE_XML_DECLARATION = 17
70
+
71
+ # A list of errors encountered while parsing
72
+ attr_accessor :errors
73
+
74
+ # The XML source
75
+ attr_reader :source
76
+
77
+ alias_method :self_closing?, :empty_element?
78
+
79
+ def initialize(source, url = nil, encoding = nil) # :nodoc:
80
+ @source = source
81
+ @errors = []
82
+ @encoding = encoding
83
+ end
84
+ private :initialize
85
+
86
+ # Get the attributes and namespaces of the current node as a Hash.
87
+ #
88
+ # This is the union of Reader#attribute_hash and Reader#namespaces
89
+ #
90
+ # [Returns]
91
+ # (Hash<String, String>) Attribute names and values, and namespace prefixes and hrefs.
92
+ def attributes
93
+ attribute_hash.merge(namespaces)
94
+ end
95
+
96
+ ###
97
+ # Move the cursor through the document yielding the cursor to the block
98
+ def each
99
+ while (cursor = read)
100
+ yield cursor
101
+ end
102
+ end
103
+ end
104
+ end
105
+ end
@@ -0,0 +1,38 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ class << self
6
+ ###
7
+ # Create a new Nokogiri::XML::RelaxNG document from +string_or_io+.
8
+ # See Nokogiri::XML::RelaxNG for an example.
9
+ def RelaxNG(string_or_io, options = ParseOptions::DEFAULT_SCHEMA)
10
+ RelaxNG.new(string_or_io, options)
11
+ end
12
+ end
13
+
14
+ ###
15
+ # Nokogiri::XML::RelaxNG is used for validating XML against a
16
+ # RelaxNG schema.
17
+ #
18
+ # == Synopsis
19
+ #
20
+ # Validate an XML document against a RelaxNG schema. Loop over the errors
21
+ # that are returned and print them out:
22
+ #
23
+ # schema = Nokogiri::XML::RelaxNG(File.open(ADDRESS_SCHEMA_FILE))
24
+ # doc = Nokogiri::XML(File.open(ADDRESS_XML_FILE))
25
+ #
26
+ # schema.validate(doc).each do |error|
27
+ # puts error.message
28
+ # end
29
+ #
30
+ # The list of errors are Nokogiri::XML::SyntaxError objects.
31
+ #
32
+ # NOTE: RelaxNG input is always treated as TRUSTED documents, meaning that they will cause the
33
+ # underlying parsing libraries to access network resources. This is counter to Nokogiri's
34
+ # "untrusted by default" security policy, but is a limitation of the underlying libraries.
35
+ class RelaxNG < Nokogiri::XML::Schema
36
+ end
37
+ end
38
+ end
@@ -0,0 +1,167 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ ###
6
+ # SAX Parsers are event driven parsers. Nokogiri provides two different event based parsers when
7
+ # dealing with XML. If you want to do SAX style parsing using HTML, check out
8
+ # Nokogiri::HTML4::SAX.
9
+ #
10
+ # The basic way a SAX style parser works is by creating a parser, telling the parser about the
11
+ # events we're interested in, then giving the parser some XML to process. The parser will notify
12
+ # you when it encounters events you said you would like to know about.
13
+ #
14
+ # To register for events, you simply subclass Nokogiri::XML::SAX::Document, and implement the
15
+ # methods for which you would like notification.
16
+ #
17
+ # For example, if I want to be notified when a document ends, and when an element starts, I
18
+ # would write a class like this:
19
+ #
20
+ # class MyDocument < Nokogiri::XML::SAX::Document
21
+ # def end_document
22
+ # puts "the document has ended"
23
+ # end
24
+ #
25
+ # def start_element name, attributes = []
26
+ # puts "#{name} started"
27
+ # end
28
+ # end
29
+ #
30
+ # Then I would instantiate a SAX parser with this document, and feed the parser some XML
31
+ #
32
+ # # Create a new parser
33
+ # parser = Nokogiri::XML::SAX::Parser.new(MyDocument.new)
34
+ #
35
+ # # Feed the parser some XML
36
+ # parser.parse(File.open(ARGV[0]))
37
+ #
38
+ # Now my document handler will be called when each node starts, and when then document ends. To
39
+ # see what kinds of events are available, take a look at Nokogiri::XML::SAX::Document.
40
+ #
41
+ # Two SAX parsers for XML are available, a parser that reads from a string or IO object as it
42
+ # feels necessary, and a parser that lets you spoon feed it XML. If you want to let Nokogiri
43
+ # deal with reading your XML, use the Nokogiri::XML::SAX::Parser. If you want to have fine grain
44
+ # control over the XML input, use the Nokogiri::XML::SAX::PushParser.
45
+ module SAX
46
+ ###
47
+ # This class is used for registering types of events you are interested in handling. All of
48
+ # the methods on this class are available as possible events while parsing an XML document. To
49
+ # register for any particular event, just subclass this class and implement the methods you
50
+ # are interested in knowing about.
51
+ #
52
+ # To only be notified about start and end element events, write a class like this:
53
+ #
54
+ # class MyDocument < Nokogiri::XML::SAX::Document
55
+ # def start_element name, attrs = []
56
+ # puts "#{name} started!"
57
+ # end
58
+ #
59
+ # def end_element name
60
+ # puts "#{name} ended"
61
+ # end
62
+ # end
63
+ #
64
+ # You can use this event handler for any SAX style parser included with Nokogiri. See
65
+ # Nokogiri::XML::SAX, and Nokogiri::HTML4::SAX.
66
+ class Document
67
+ ###
68
+ # Called when an XML declaration is parsed
69
+ def xmldecl(version, encoding, standalone)
70
+ end
71
+
72
+ ###
73
+ # Called when document starts parsing
74
+ def start_document
75
+ end
76
+
77
+ ###
78
+ # Called when document ends parsing
79
+ def end_document
80
+ end
81
+
82
+ ###
83
+ # Called at the beginning of an element
84
+ # * +name+ is the name of the tag
85
+ # * +attrs+ are an assoc list of namespaces and attributes, e.g.:
86
+ # [ ["xmlns:foo", "http://sample.net"], ["size", "large"] ]
87
+ def start_element(name, attrs = [])
88
+ end
89
+
90
+ ###
91
+ # Called at the end of an element
92
+ # +name+ is the tag name
93
+ def end_element(name)
94
+ end
95
+
96
+ ###
97
+ # Called at the beginning of an element
98
+ # +name+ is the element name
99
+ # +attrs+ is a list of attributes
100
+ # +prefix+ is the namespace prefix for the element
101
+ # +uri+ is the associated namespace URI
102
+ # +ns+ is a hash of namespace prefix:urls associated with the element
103
+ def start_element_namespace(name, attrs = [], prefix = nil, uri = nil, ns = [])
104
+ ###
105
+ # Deal with SAX v1 interface
106
+ name = [prefix, name].compact.join(":")
107
+ attributes = ns.map do |ns_prefix, ns_uri|
108
+ [["xmlns", ns_prefix].compact.join(":"), ns_uri]
109
+ end + attrs.map do |attr|
110
+ [[attr.prefix, attr.localname].compact.join(":"), attr.value]
111
+ end
112
+ start_element(name, attributes)
113
+ end
114
+
115
+ ###
116
+ # Called at the end of an element
117
+ # +name+ is the element's name
118
+ # +prefix+ is the namespace prefix associated with the element
119
+ # +uri+ is the associated namespace URI
120
+ def end_element_namespace(name, prefix = nil, uri = nil)
121
+ ###
122
+ # Deal with SAX v1 interface
123
+ end_element([prefix, name].compact.join(":"))
124
+ end
125
+
126
+ ###
127
+ # Characters read between a tag. This method might be called multiple
128
+ # times given one contiguous string of characters.
129
+ #
130
+ # +string+ contains the character data
131
+ def characters(string)
132
+ end
133
+
134
+ ###
135
+ # Called when comments are encountered
136
+ # +string+ contains the comment data
137
+ def comment(string)
138
+ end
139
+
140
+ ###
141
+ # Called on document warnings
142
+ # +string+ contains the warning
143
+ def warning(string)
144
+ end
145
+
146
+ ###
147
+ # Called on document errors
148
+ # +string+ contains the error
149
+ def error(string)
150
+ end
151
+
152
+ ###
153
+ # Called when cdata blocks are found
154
+ # +string+ contains the cdata content
155
+ def cdata_block(string)
156
+ end
157
+
158
+ ###
159
+ # Called when processing instructions are found
160
+ # +name+ is the target of the instruction
161
+ # +content+ is the value of the instruction
162
+ def processing_instruction(name, content)
163
+ end
164
+ end
165
+ end
166
+ end
167
+ end
@@ -0,0 +1,125 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ module SAX
6
+ ###
7
+ # This parser is a SAX style parser that reads it's input as it
8
+ # deems necessary. The parser takes a Nokogiri::XML::SAX::Document,
9
+ # an optional encoding, then given an XML input, sends messages to
10
+ # the Nokogiri::XML::SAX::Document.
11
+ #
12
+ # Here is an example of using this parser:
13
+ #
14
+ # # Create a subclass of Nokogiri::XML::SAX::Document and implement
15
+ # # the events we care about:
16
+ # class MyDoc < Nokogiri::XML::SAX::Document
17
+ # def start_element name, attrs = []
18
+ # puts "starting: #{name}"
19
+ # end
20
+ #
21
+ # def end_element name
22
+ # puts "ending: #{name}"
23
+ # end
24
+ # end
25
+ #
26
+ # # Create our parser
27
+ # parser = Nokogiri::XML::SAX::Parser.new(MyDoc.new)
28
+ #
29
+ # # Send some XML to the parser
30
+ # parser.parse(File.open(ARGV[0]))
31
+ #
32
+ # For more information about SAX parsers, see Nokogiri::XML::SAX. Also
33
+ # see Nokogiri::XML::SAX::Document for the available events.
34
+ class Parser
35
+ class Attribute < Struct.new(:localname, :prefix, :uri, :value)
36
+ end
37
+
38
+ # Encodinds this parser supports
39
+ ENCODINGS = {
40
+ "NONE" => 0, # No char encoding detected
41
+ "UTF-8" => 1, # UTF-8
42
+ "UTF16LE" => 2, # UTF-16 little endian
43
+ "UTF16BE" => 3, # UTF-16 big endian
44
+ "UCS4LE" => 4, # UCS-4 little endian
45
+ "UCS4BE" => 5, # UCS-4 big endian
46
+ "EBCDIC" => 6, # EBCDIC uh!
47
+ "UCS4-2143" => 7, # UCS-4 unusual ordering
48
+ "UCS4-3412" => 8, # UCS-4 unusual ordering
49
+ "UCS2" => 9, # UCS-2
50
+ "ISO-8859-1" => 10, # ISO-8859-1 ISO Latin 1
51
+ "ISO-8859-2" => 11, # ISO-8859-2 ISO Latin 2
52
+ "ISO-8859-3" => 12, # ISO-8859-3
53
+ "ISO-8859-4" => 13, # ISO-8859-4
54
+ "ISO-8859-5" => 14, # ISO-8859-5
55
+ "ISO-8859-6" => 15, # ISO-8859-6
56
+ "ISO-8859-7" => 16, # ISO-8859-7
57
+ "ISO-8859-8" => 17, # ISO-8859-8
58
+ "ISO-8859-9" => 18, # ISO-8859-9
59
+ "ISO-2022-JP" => 19, # ISO-2022-JP
60
+ "SHIFT-JIS" => 20, # Shift_JIS
61
+ "EUC-JP" => 21, # EUC-JP
62
+ "ASCII" => 22, # pure ASCII
63
+ }
64
+
65
+ # The Nokogiri::XML::SAX::Document where events will be sent.
66
+ attr_accessor :document
67
+
68
+ # The encoding beings used for this document.
69
+ attr_accessor :encoding
70
+
71
+ # Create a new Parser with +doc+ and +encoding+
72
+ def initialize(doc = Nokogiri::XML::SAX::Document.new, encoding = "UTF-8")
73
+ @encoding = check_encoding(encoding)
74
+ @document = doc
75
+ @warned = false
76
+ end
77
+
78
+ ###
79
+ # Parse given +thing+ which may be a string containing xml, or an
80
+ # IO object.
81
+ def parse(thing, &block)
82
+ if thing.respond_to?(:read) && thing.respond_to?(:close)
83
+ parse_io(thing, &block)
84
+ else
85
+ parse_memory(thing, &block)
86
+ end
87
+ end
88
+
89
+ ###
90
+ # Parse given +io+
91
+ def parse_io(io, encoding = @encoding)
92
+ ctx = ParserContext.io(io, ENCODINGS[check_encoding(encoding)])
93
+ yield ctx if block_given?
94
+ ctx.parse_with(self)
95
+ end
96
+
97
+ ###
98
+ # Parse a file with +filename+
99
+ def parse_file(filename)
100
+ raise ArgumentError unless filename
101
+ raise Errno::ENOENT unless File.exist?(filename)
102
+ raise Errno::EISDIR if File.directory?(filename)
103
+
104
+ ctx = ParserContext.file(filename)
105
+ yield ctx if block_given?
106
+ ctx.parse_with(self)
107
+ end
108
+
109
+ def parse_memory(data)
110
+ ctx = ParserContext.memory(data)
111
+ yield ctx if block_given?
112
+ ctx.parse_with(self)
113
+ end
114
+
115
+ private
116
+
117
+ def check_encoding(encoding)
118
+ encoding.upcase.tap do |enc|
119
+ raise ArgumentError, "'#{enc}' is not a valid encoding" unless ENCODINGS[enc]
120
+ end
121
+ end
122
+ end
123
+ end
124
+ end
125
+ end
@@ -0,0 +1,21 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ module SAX
6
+ ###
7
+ # Context for XML SAX parsers. This class is usually not instantiated
8
+ # by the user. Instead, you should be looking at
9
+ # Nokogiri::XML::SAX::Parser
10
+ class ParserContext
11
+ def self.new(thing, encoding = "UTF-8")
12
+ if [:read, :close].all? { |x| thing.respond_to?(x) }
13
+ io(thing, Parser::ENCODINGS[encoding])
14
+ else
15
+ memory(thing)
16
+ end
17
+ end
18
+ end
19
+ end
20
+ end
21
+ end
@@ -0,0 +1,61 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ module SAX
6
+ ###
7
+ # PushParser can parse a document that is fed to it manually. It
8
+ # must be given a SAX::Document object which will be called with
9
+ # SAX events as the document is being parsed.
10
+ #
11
+ # Calling PushParser#<< writes XML to the parser, calling any SAX
12
+ # callbacks it can.
13
+ #
14
+ # PushParser#finish tells the parser that the document is finished
15
+ # and calls the end_document SAX method.
16
+ #
17
+ # Example:
18
+ #
19
+ # parser = PushParser.new(Class.new(XML::SAX::Document) {
20
+ # def start_document
21
+ # puts "start document called"
22
+ # end
23
+ # }.new)
24
+ # parser << "<div>hello<"
25
+ # parser << "/div>"
26
+ # parser.finish
27
+ class PushParser
28
+ # The Nokogiri::XML::SAX::Document on which the PushParser will be
29
+ # operating
30
+ attr_accessor :document
31
+
32
+ ###
33
+ # Create a new PushParser with +doc+ as the SAX Document, providing
34
+ # an optional +file_name+ and +encoding+
35
+ def initialize(doc = XML::SAX::Document.new, file_name = nil, encoding = "UTF-8")
36
+ @document = doc
37
+ @encoding = encoding
38
+ @sax_parser = XML::SAX::Parser.new(doc)
39
+
40
+ ## Create our push parser context
41
+ initialize_native(@sax_parser, file_name)
42
+ end
43
+
44
+ ###
45
+ # Write a +chunk+ of XML to the PushParser. Any callback methods
46
+ # that can be called will be called immediately.
47
+ def write(chunk, last_chunk = false)
48
+ native_write(chunk, last_chunk)
49
+ end
50
+ alias_method :<<, :write
51
+
52
+ ###
53
+ # Finish the parsing. This method is only necessary for
54
+ # Nokogiri::XML::SAX::Document#end_document to be called.
55
+ def finish
56
+ write("", true)
57
+ end
58
+ end
59
+ end
60
+ end
61
+ end
@@ -0,0 +1,6 @@
1
+ # frozen_string_literal: true
2
+
3
+ require_relative "sax/document"
4
+ require_relative "sax/parser_context"
5
+ require_relative "sax/parser"
6
+ require_relative "sax/push_parser"
@@ -0,0 +1,73 @@
1
+ # frozen_string_literal: true
2
+
3
+ module Nokogiri
4
+ module XML
5
+ class << self
6
+ ###
7
+ # Create a new Nokogiri::XML::Schema object using a +string_or_io+
8
+ # object.
9
+ def Schema(string_or_io, options = ParseOptions::DEFAULT_SCHEMA)
10
+ Schema.new(string_or_io, options)
11
+ end
12
+ end
13
+
14
+ ###
15
+ # Nokogiri::XML::Schema is used for validating XML against a schema
16
+ # (usually from an xsd file).
17
+ #
18
+ # == Synopsis
19
+ #
20
+ # Validate an XML document against a Schema. Loop over the errors that
21
+ # are returned and print them out:
22
+ #
23
+ # xsd = Nokogiri::XML::Schema(File.read(PO_SCHEMA_FILE))
24
+ # doc = Nokogiri::XML(File.read(PO_XML_FILE))
25
+ #
26
+ # xsd.validate(doc).each do |error|
27
+ # puts error.message
28
+ # end
29
+ #
30
+ # The list of errors are Nokogiri::XML::SyntaxError objects.
31
+ #
32
+ # NOTE: As of v1.11.0, Schema treats inputs as UNTRUSTED by default, and so external entities
33
+ # are not resolved from the network (`http://` or `ftp://`). Previously, parsing treated
34
+ # documents as "trusted" by default which was counter to Nokogiri's "untrusted by default"
35
+ # security policy. If a document is trusted, then the caller may turn off the NONET option via
36
+ # the ParseOptions to re-enable external entity resolution over a network connection.
37
+ class Schema
38
+ # Errors while parsing the schema file
39
+ attr_accessor :errors
40
+ # The Nokogiri::XML::ParseOptions used to parse the schema
41
+ attr_accessor :parse_options
42
+
43
+ ###
44
+ # Create a new Nokogiri::XML::Schema object using a +string_or_io+
45
+ # object.
46
+ def self.new(string_or_io, options = ParseOptions::DEFAULT_SCHEMA)
47
+ from_document(Nokogiri::XML(string_or_io), options)
48
+ end
49
+
50
+ ###
51
+ # Validate +thing+ against this schema. +thing+ can be a
52
+ # Nokogiri::XML::Document object, or a filename. An Array of
53
+ # Nokogiri::XML::SyntaxError objects found while validating the
54
+ # +thing+ is returned.
55
+ def validate(thing)
56
+ if thing.is_a?(Nokogiri::XML::Document)
57
+ validate_document(thing)
58
+ elsif File.file?(thing)
59
+ validate_file(thing)
60
+ else
61
+ raise ArgumentError, "Must provide Nokogiri::Xml::Document or the name of an existing file"
62
+ end
63
+ end
64
+
65
+ ###
66
+ # Returns true if +thing+ is a valid Nokogiri::XML::Document or
67
+ # file.
68
+ def valid?(thing)
69
+ validate(thing).empty?
70
+ end
71
+ end
72
+ end
73
+ end