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 +1,50 @@
1
+ package nokogiri.internals;
2
+
3
+ import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
4
+ import nokogiri.NokogiriService;
5
+ import nokogiri.XmlSyntaxError;
6
+
7
+ import org.jruby.Ruby;
8
+ import org.jruby.RubyArray;
9
+ import org.xml.sax.ErrorHandler;
10
+ import org.xml.sax.SAXException;
11
+ import org.xml.sax.SAXParseException;
12
+
13
+ /**
14
+ * Error handler for Relax and W3C XML Schema.
15
+ *
16
+ * @author sergio
17
+ * @author Yoko Harada <yokolet@gmail.com>
18
+ */
19
+ public class SchemaErrorHandler implements ErrorHandler
20
+ {
21
+
22
+ private final Ruby runtime;
23
+ final RubyArray<?> errors;
24
+
25
+ public
26
+ SchemaErrorHandler(Ruby ruby, RubyArray<?> array)
27
+ {
28
+ this.runtime = ruby;
29
+ this.errors = array;
30
+ }
31
+
32
+ public void
33
+ warning(SAXParseException ex) throws SAXException
34
+ {
35
+ errors.append(XmlSyntaxError.createWarning(runtime, ex));
36
+ }
37
+
38
+ public void
39
+ error(SAXParseException ex) throws SAXException
40
+ {
41
+ errors.append(XmlSyntaxError.createError(runtime, ex));
42
+ }
43
+
44
+ public void
45
+ fatalError(SAXParseException ex) throws SAXException
46
+ {
47
+ throw ex;
48
+ }
49
+
50
+ }
@@ -0,0 +1,174 @@
1
+ /*
2
+ * Copyright (c) 2017 [Karol Bucek](http://kares.org/)
3
+ *
4
+ * Licensed to the Apache Software Foundation (ASF) under one
5
+ * or more contributor license agreements. See the NOTICE file
6
+ * distributed with this work for additional information
7
+ * regarding copyright ownership. The ASF licenses this file
8
+ * to you under the Apache License, Version 2.0 (the "License");
9
+ * you may not use this file except in compliance with the License.
10
+ * You may obtain a copy of the License at
11
+ *
12
+ * http://www.apache.org/licenses/LICENSE-2.0
13
+ *
14
+ * Unless required by applicable law or agreed to in writing, software
15
+ * distributed under the License is distributed on an "AS IS" BASIS,
16
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
17
+ * See the License for the specific language governing permissions and
18
+ * limitations under the License.
19
+ */
20
+ package nokogiri.internals;
21
+
22
+ import javax.xml.transform.dom.DOMSource;
23
+
24
+ import org.apache.xml.dtm.DTM;
25
+ import nokogiri.internals.dom2dtm.DOM2DTM;
26
+ import nokogiri.internals.dom2dtm.DOM2DTMdefaultNamespaceDeclarationNode;
27
+ import org.apache.xml.dtm.DTMWSFilter;
28
+ import org.apache.xml.res.XMLErrorResources;
29
+ import org.apache.xml.res.XMLMessages;
30
+ import org.w3c.dom.Node;
31
+
32
+ /**
33
+ * @author kares
34
+ */
35
+ public final class XalanDTMManagerPatch extends org.apache.xml.dtm.ref.DTMManagerDefault
36
+ {
37
+
38
+ /**
39
+ * Given a W3C DOM node, try and return a DTM handle.
40
+ * Note: calling this may be non-optimal, and there is no guarantee that
41
+ * the node will be found in any particular DTM.
42
+ *
43
+ * @param node Non-null reference to a DOM node.
44
+ *
45
+ * @return a valid DTM handle.
46
+ */
47
+ @Override
48
+ public /* synchronized */ int
49
+ getDTMHandleFromNode(org.w3c.dom.Node node)
50
+ {
51
+ //if (node == null) // "node must be non-null for getDTMHandleFromNode!");
52
+ // throw new IllegalArgumentException(XMLMessages.createXMLMessage(XMLErrorResources.ER_NODE_NON_NULL, null));
53
+ assert node != null;
54
+
55
+ if (node instanceof org.apache.xml.dtm.ref.DTMNodeProxy) {
56
+ return ((org.apache.xml.dtm.ref.DTMNodeProxy) node).getDTMNodeNumber();
57
+ }
58
+
59
+ // Find the DOM2DTMs wrapped around this Document (if any)
60
+ // and check whether they contain the Node in question.
61
+ //
62
+ // NOTE that since a DOM2DTM may represent a subtree rather
63
+ // than a full document, we have to be prepared to check more
64
+ // than one -- and there is no guarantee that we will find
65
+ // one that contains ancestors or siblings of the node we're
66
+ // seeking.
67
+ //
68
+ // %REVIEW% We could search for the one which contains this
69
+ // node at the deepest level, and thus covers the widest
70
+ // subtree, but that's going to entail additional work
71
+ // checking more DTMs... and getHandleOfNode is not a
72
+ // cheap operation in most implementations.
73
+ //
74
+ // TODO: %REVIEW% If overflow addressing, we may recheck a DTM
75
+ // already examined. Ouch. But with the increased number of DTMs,
76
+ // scanning back to check this is painful.
77
+ // POSSIBLE SOLUTIONS:
78
+ // Generate a list of _unique_ DTM objects?
79
+ // Have each DTM cache last DOM node search?
80
+ for (int i = 0; i < m_dtms.length; i++) {
81
+ DTM thisDTM = m_dtms[i];
82
+ if (thisDTM instanceof DOM2DTM) {
83
+ int handle = ((DOM2DTM) thisDTM).getHandleOfNode(node);
84
+ if (handle != DTM.NULL) {
85
+ return handle;
86
+ }
87
+ }
88
+ }
89
+
90
+ // Not found; generate a new DTM.
91
+ //
92
+ // %REVIEW% Is this really desirable, or should we return null
93
+ // and make folks explicitly instantiate from a DOMSource? The
94
+ // latter is more work but gives the caller the opportunity to
95
+ // explicitly add the DTM to a DTMManager... and thus to know when
96
+ // it can be discarded again, which is something we need to pay much
97
+ // more attention to. (Especially since only DTMs which are assigned
98
+ // to a manager can use the overflow addressing scheme.)
99
+ //
100
+ // %BUG% If the source node was a DOM2DTM$defaultNamespaceDeclarationNode
101
+ // and the DTM wasn't registered with this DTMManager, we will create
102
+ // a new DTM and _still_ not be able to find the node (since it will
103
+ // be resynthesized). Another reason to push hard on making all DTMs
104
+ // be managed DTMs.
105
+
106
+ // Since the real root of our tree may be a DocumentFragment, we need to
107
+ // use getParent to find the root, instead of getOwnerDocument. Otherwise
108
+ // DOM2DTM#getHandleOfNode will be very unhappy.
109
+ Node root = node;
110
+ int rootType = root.getNodeType();
111
+ Node p = (rootType == Node.ATTRIBUTE_NODE) ? ((org.w3c.dom.Attr) root).getOwnerElement() : root.getParentNode();
112
+ for (; p != null; p = p.getParentNode()) { root = p; }
113
+
114
+ // DOM2DTM dtm = (DOM2DTM) getDTM(new DOMSource(root), false, null);
115
+ DOM2DTM dtm = getDTM(new DOMSource(root), false, null/*, true, true*/);
116
+
117
+ int handle;
118
+
119
+ if (node instanceof org.apache.xml.dtm.ref.dom2dtm.DOM2DTMdefaultNamespaceDeclarationNode
120
+ || node instanceof DOM2DTMdefaultNamespaceDeclarationNode) {
121
+ // Can't return the same node since it's unique to a specific DTM,
122
+ // but can return the equivalent node -- find the corresponding
123
+ // Document Element, then ask it for the xml: namespace decl.
124
+ handle = dtm.getHandleOfNode(((org.w3c.dom.Attr) node).getOwnerElement());
125
+ handle = dtm.getAttributeNode(handle, node.getNamespaceURI(), node.getLocalName());
126
+ } else {
127
+ handle = dtm.getHandleOfNode(node);
128
+
129
+ rootType = root.getNodeType();
130
+ // Is Node actually within the same document? If not, don't search!
131
+ // This would be easier if m_root was always the Document node, but
132
+ // we decided to allow wrapping a DTM around a subtree.
133
+ if ((root == node) ||
134
+ (rootType == Node.DOCUMENT_NODE && root == node.getOwnerDocument()) ||
135
+ (rootType != Node.DOCUMENT_NODE && root.getOwnerDocument() == node.getOwnerDocument())
136
+ ) {
137
+ // If node _is_ in m_root's tree, find its handle
138
+ //
139
+ // %OPT% This check may be improved significantly when DOM
140
+ // Level 3 nodeKey and relative-order tests become
141
+ // available!
142
+ for (Node cursor = node; cursor != null;
143
+ cursor = (cursor.getNodeType() != Node.ATTRIBUTE_NODE)
144
+ ? cursor.getParentNode()
145
+ : ((org.w3c.dom.Attr)cursor).getOwnerElement()) {
146
+ if (cursor == root) {
147
+ // We know this node; find its handle.
148
+ return (dtm).getHandleFromNode(node);
149
+ }
150
+ } // for ancestors of node
151
+ } // if node and m_root in same Document
152
+ }
153
+
154
+ if (DTM.NULL == handle) {
155
+ throw new RuntimeException(XMLMessages.createXMLMessage(XMLErrorResources.ER_COULD_NOT_RESOLVE_NODE,
156
+ null)); //"Could not resolve the node to a handle!");
157
+ }
158
+
159
+ return handle;
160
+ }
161
+
162
+ private DOM2DTM
163
+ getDTM(DOMSource source, boolean unique, DTMWSFilter whiteSpaceFilter/*, boolean incremental, boolean doIndexing*/)
164
+ {
165
+ int dtmPos = getFirstFreeDTMID();
166
+ int documentID = dtmPos << IDENT_DTM_NODE_BITS;
167
+
168
+ DOM2DTM dtm = new DOM2DTM(this, source, documentID, whiteSpaceFilter, m_xsf, true);
169
+
170
+ addDTM(dtm, dtmPos, 0);
171
+ return dtm;
172
+ }
173
+
174
+ }
@@ -0,0 +1,11 @@
1
+ package nokogiri.internals;
2
+
3
+ /**
4
+ * Interface for receiving xmlDecl information.
5
+ *
6
+ * @author Patrick Mahoney <pat@polycrystal.org>
7
+ */
8
+ public interface XmlDeclHandler
9
+ {
10
+ public void xmlDecl(String version, String encoding, String standalone);
11
+ }
@@ -0,0 +1,265 @@
1
+ package nokogiri.internals;
2
+
3
+ import nokogiri.XmlDocument;
4
+ import nokogiri.XmlDtd;
5
+ import nokogiri.XmlSyntaxError;
6
+ import org.apache.xerces.parsers.DOMParser;
7
+ import org.jruby.*;
8
+ import org.jruby.exceptions.RaiseException;
9
+ import org.jruby.runtime.Helpers;
10
+ import org.jruby.runtime.ThreadContext;
11
+ import org.jruby.runtime.builtin.IRubyObject;
12
+ import org.w3c.dom.Document;
13
+ import org.w3c.dom.Node;
14
+ import org.w3c.dom.NodeList;
15
+ import org.xml.sax.SAXException;
16
+
17
+ import java.io.IOException;
18
+ import java.util.ArrayList;
19
+ import java.util.List;
20
+
21
+ import static nokogiri.internals.NokogiriHelpers.isBlank;
22
+
23
+ /**
24
+ * Parser class for XML DOM processing. This class actually parses XML document
25
+ * and creates DOM tree in Java side. However, DOM tree in Ruby side is not since
26
+ * we delay creating objects for performance.
27
+ *
28
+ * @author sergio
29
+ * @author Yoko Harada <yokolet@gmail.com>
30
+ */
31
+ public class XmlDomParserContext extends ParserContext
32
+ {
33
+ private static final long serialVersionUID = 1L;
34
+
35
+ protected static final String FEATURE_LOAD_EXTERNAL_DTD =
36
+ "http://apache.org/xml/features/nonvalidating/load-external-dtd";
37
+ protected static final String FEATURE_LOAD_DTD_GRAMMAR =
38
+ "http://apache.org/xml/features/nonvalidating/load-dtd-grammar";
39
+ protected static final String FEATURE_INCLUDE_IGNORABLE_WHITESPACE =
40
+ "http://apache.org/xml/features/dom/include-ignorable-whitespace";
41
+ protected static final String CONTINUE_AFTER_FATAL_ERROR =
42
+ "http://apache.org/xml/features/continue-after-fatal-error";
43
+ protected static final String FEATURE_NOT_EXPAND_ENTITY =
44
+ "http://apache.org/xml/features/dom/create-entity-ref-nodes";
45
+ protected static final String FEATURE_VALIDATION = "http://xml.org/sax/features/validation";
46
+ private static final String SECURITY_MANAGER = "http://apache.org/xml/properties/security-manager";
47
+
48
+ protected ParserContext.Options options;
49
+ protected DOMParser parser;
50
+ protected NokogiriErrorHandler errorHandler;
51
+ protected IRubyObject ruby_encoding;
52
+
53
+ public
54
+ XmlDomParserContext(Ruby runtime, IRubyObject options)
55
+ {
56
+ this(runtime, runtime.getNil(), options);
57
+ }
58
+
59
+ public
60
+ XmlDomParserContext(Ruby runtime, IRubyObject encoding, IRubyObject options)
61
+ {
62
+ super(runtime);
63
+ this.options = new ParserContext.Options(RubyFixnum.fix2long(options));
64
+ java_encoding = NokogiriHelpers.getValidEncodingOrNull(encoding);
65
+ ruby_encoding = encoding;
66
+ initErrorHandler(runtime);
67
+ initParser(runtime);
68
+ }
69
+
70
+ protected void
71
+ initErrorHandler(Ruby runtime)
72
+ {
73
+ if (options.recover) {
74
+ errorHandler = new NokogiriNonStrictErrorHandler(runtime, options.noError, options.noWarning);
75
+ } else {
76
+ errorHandler = new NokogiriStrictErrorHandler(runtime, options.noError, options.noWarning);
77
+ }
78
+ }
79
+
80
+ protected void
81
+ initParser(Ruby runtime)
82
+ {
83
+ if (options.xInclude) {
84
+ System.setProperty("org.apache.xerces.xni.parser.XMLParserConfiguration",
85
+ "org.apache.xerces.parsers.XIncludeParserConfiguration");
86
+ }
87
+
88
+ parser = new NokogiriDomParser(options);
89
+ parser.setErrorHandler(errorHandler);
90
+
91
+ // Fix for Issue#586. This limits entity expansion up to 100000 and nodes up to 3000.
92
+ setProperty(SECURITY_MANAGER, new org.apache.xerces.util.SecurityManager());
93
+
94
+ if (options.noBlanks) {
95
+ setFeature(FEATURE_INCLUDE_IGNORABLE_WHITESPACE, false);
96
+ }
97
+
98
+ if (options.recover) {
99
+ setFeature(CONTINUE_AFTER_FATAL_ERROR, true);
100
+ }
101
+
102
+ if (options.dtdValid) {
103
+ setFeature(FEATURE_VALIDATION, true);
104
+ }
105
+
106
+ if (!options.noEnt) {
107
+ setFeature(FEATURE_NOT_EXPAND_ENTITY, true);
108
+ }
109
+ // If we turn off loading of external DTDs complete, we don't
110
+ // getthe publicID. Instead of turning off completely, we use
111
+ // an entity resolver that returns empty documents.
112
+ if (options.dtdLoad) {
113
+ setFeature(FEATURE_LOAD_EXTERNAL_DTD, true);
114
+ setFeature(FEATURE_LOAD_DTD_GRAMMAR, true);
115
+ }
116
+ parser.setEntityResolver(new NokogiriEntityResolver(runtime, errorHandler, options));
117
+ }
118
+
119
+ /**
120
+ * Convenience method that catches and ignores SAXException
121
+ * (unrecognized and unsupported exceptions).
122
+ */
123
+ protected void
124
+ setFeature(String feature, boolean value)
125
+ {
126
+ try {
127
+ parser.setFeature(feature, value);
128
+ } catch (SAXException e) {
129
+ // ignore
130
+ }
131
+ }
132
+
133
+ /**
134
+ * Convenience method that catches and ignores SAXException
135
+ * (unrecognized and unsupported exceptions).
136
+ */
137
+ protected void
138
+ setProperty(String property, Object value)
139
+ {
140
+ try {
141
+ parser.setProperty(property, value);
142
+ } catch (SAXException e) {
143
+ // ignore
144
+ }
145
+ }
146
+
147
+ public void
148
+ addErrorsIfNecessary(ThreadContext context, XmlDocument doc)
149
+ {
150
+ doc.setInstanceVariable("@errors", mapErrors(context, errorHandler));
151
+ }
152
+
153
+
154
+ public static RubyArray<?>
155
+ mapErrors(ThreadContext context, NokogiriErrorHandler errorHandler)
156
+ {
157
+ final Ruby runtime = context.runtime;
158
+ final List<RubyException> errors = errorHandler.getErrors();
159
+ final IRubyObject[] errorsAry = new IRubyObject[errors.size()];
160
+ for (int i = 0; i < errors.size(); i++) {
161
+ errorsAry[i] = errors.get(i);
162
+ }
163
+ return runtime.newArrayNoCopy(errorsAry);
164
+ }
165
+
166
+ public XmlDocument
167
+ getDocumentWithErrorsOrRaiseException(ThreadContext context, RubyClass klazz, Exception ex)
168
+ {
169
+ if (options.recover) {
170
+ XmlDocument xmlDocument = getInterruptedOrNewXmlDocument(context, klazz);
171
+ this.addErrorsIfNecessary(context, xmlDocument);
172
+ XmlSyntaxError xmlSyntaxError = XmlSyntaxError.createXMLSyntaxError(context.runtime);
173
+ xmlSyntaxError.setException(ex);
174
+ ((RubyArray) xmlDocument.getInstanceVariable("@errors")).append(xmlSyntaxError);
175
+ return xmlDocument;
176
+ } else {
177
+ XmlSyntaxError xmlSyntaxError = XmlSyntaxError.createXMLSyntaxError(context.runtime);
178
+ xmlSyntaxError.setException(ex);
179
+ throw xmlSyntaxError.toThrowable();
180
+ }
181
+ }
182
+
183
+ private XmlDocument
184
+ getInterruptedOrNewXmlDocument(ThreadContext context, RubyClass klass)
185
+ {
186
+ Document document = parser.getDocument();
187
+ XmlDocument xmlDocument = new XmlDocument(context.runtime, klass, document);
188
+ xmlDocument.setEncoding(ruby_encoding);
189
+ return xmlDocument;
190
+ }
191
+
192
+ /**
193
+ * This method is broken out so that HtmlDomParserContext can
194
+ * override it.
195
+ */
196
+ protected XmlDocument
197
+ wrapDocument(ThreadContext context, RubyClass klass, Document doc)
198
+ {
199
+ XmlDocument xmlDocument = new XmlDocument(context.runtime, klass, doc);
200
+ Helpers.invoke(context, xmlDocument, "initialize");
201
+ xmlDocument.setEncoding(ruby_encoding);
202
+
203
+ if (options.dtdLoad) {
204
+ IRubyObject dtd = XmlDtd.newFromExternalSubset(context.runtime, doc);
205
+ if (!dtd.isNil()) {
206
+ doc.setUserData(XmlDocument.DTD_EXTERNAL_SUBSET, (XmlDtd) dtd, null);
207
+ }
208
+ }
209
+ return xmlDocument;
210
+ }
211
+
212
+ /**
213
+ * Must call setInputSource() before this method.
214
+ */
215
+ public XmlDocument
216
+ parse(ThreadContext context, RubyClass klass, IRubyObject url)
217
+ {
218
+ XmlDocument xmlDoc;
219
+ try {
220
+ Document doc = do_parse();
221
+ xmlDoc = wrapDocument(context, klass, doc);
222
+ xmlDoc.setUrl(url);
223
+ addErrorsIfNecessary(context, xmlDoc);
224
+ return xmlDoc;
225
+ } catch (SAXException e) {
226
+ return getDocumentWithErrorsOrRaiseException(context, klass, e);
227
+ } catch (IOException e) {
228
+ return getDocumentWithErrorsOrRaiseException(context, klass, e);
229
+ }
230
+ }
231
+
232
+ protected Document
233
+ do_parse() throws SAXException, IOException
234
+ {
235
+ try {
236
+ parser.parse(getInputSource());
237
+ } catch (NullPointerException ex) {
238
+ // FIXME: this is really a hack to fix #838. Xerces will throw a NullPointerException
239
+ // if we tried to parse '<? ?>'. We should submit a patch to Xerces.
240
+ }
241
+ if (options.noBlanks) {
242
+ List<Node> emptyNodes = new ArrayList<Node>();
243
+ findEmptyTexts(parser.getDocument(), emptyNodes);
244
+ if (emptyNodes.size() > 0) {
245
+ for (Node node : emptyNodes) {
246
+ node.getParentNode().removeChild(node);
247
+ }
248
+ }
249
+ }
250
+ return parser.getDocument();
251
+ }
252
+
253
+ private static void
254
+ findEmptyTexts(Node node, List<Node> emptyNodes)
255
+ {
256
+ if (node.getNodeType() == Node.TEXT_NODE && isBlank(node.getTextContent())) {
257
+ emptyNodes.add(node);
258
+ } else {
259
+ NodeList children = node.getChildNodes();
260
+ for (int i = 0; i < children.getLength(); i++) {
261
+ findEmptyTexts(children.item(i), emptyNodes);
262
+ }
263
+ }
264
+ }
265
+ }
@@ -0,0 +1,40 @@
1
+ package nokogiri.internals;
2
+
3
+ import org.apache.xerces.parsers.SAXParser;
4
+ import org.apache.xerces.xni.Augmentations;
5
+ import org.apache.xerces.xni.XNIException;
6
+
7
+ /**
8
+ * Extends SAXParser in order to receive xmlDecl events and pass them
9
+ * on to a handler.
10
+ *
11
+ * @author Patrick Mahoney <pat@polycrystal.org>
12
+ */
13
+ public class XmlSaxParser extends SAXParser
14
+ {
15
+
16
+ protected XmlDeclHandler xmlDeclHandler = null;
17
+
18
+ public
19
+ XmlSaxParser()
20
+ {
21
+ super();
22
+ }
23
+
24
+ public void
25
+ setXmlDeclHandler(XmlDeclHandler xmlDeclHandler)
26
+ {
27
+ this.xmlDeclHandler = xmlDeclHandler;
28
+ }
29
+
30
+ @Override
31
+ public void
32
+ xmlDecl(String version, String encoding, String standalone,
33
+ Augmentations augs) throws XNIException
34
+ {
35
+ super.xmlDecl(version, encoding, standalone, augs);
36
+ if (xmlDeclHandler != null) {
37
+ xmlDeclHandler.xmlDecl(version, encoding, standalone);
38
+ }
39
+ }
40
+ }
@@ -0,0 +1,122 @@
1
+ /**
2
+ * Licensed to the Apache Software Foundation (ASF) under one
3
+ * or more contributor license agreements. See the NOTICE file
4
+ * distributed with this work for additional information
5
+ * regarding copyright ownership. The ASF licenses this file
6
+ * to you under the Apache License, Version 2.0 (the
7
+ * "License"); you may not use this file except in compliance
8
+ * with the License. You may obtain a copy of the License at
9
+ *
10
+ * http://www.apache.org/licenses/LICENSE-2.0
11
+ *
12
+ * Unless required by applicable law or agreed to in writing,
13
+ * software distributed under the License is distributed on an
14
+ * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
15
+ * KIND, either express or implied. See the License for the
16
+ * specific language governing permissions and limitations
17
+ * under the License.
18
+ */
19
+ package nokogiri.internals.c14n;
20
+
21
+ import java.io.Serializable;
22
+ import java.util.Comparator;
23
+
24
+
25
+ import org.w3c.dom.Attr;
26
+
27
+ /**
28
+ * Compares two attributes based on the C14n specification.
29
+ *
30
+ * <UL>
31
+ * <LI>Namespace nodes have a lesser document order position than attribute
32
+ * nodes.
33
+ * <LI> An element's namespace nodes are sorted lexicographically by
34
+ * local name (the default namespace node, if one exists, has no
35
+ * local name and is therefore lexicographically least).
36
+ * <LI> An element's attribute nodes are sorted lexicographically with
37
+ * namespace URI as the primary key and local name as the secondary
38
+ * key (an empty namespace URI is lexicographically least).
39
+ * </UL>
40
+ *
41
+ * @author Christian Geuer-Pollmann
42
+ */
43
+ public class AttrCompare implements Comparator<Attr>, Serializable
44
+ {
45
+
46
+ private static final long serialVersionUID = -7113259629930576230L;
47
+ private static final int ATTR0_BEFORE_ATTR1 = -1;
48
+ private static final int ATTR1_BEFORE_ATTR0 = 1;
49
+ private static final String XMLNS = Constants.NamespaceSpecNS;
50
+
51
+ /**
52
+ * Compares two attributes based on the C14n specification.
53
+ *
54
+ * <UL>
55
+ * <LI>Namespace nodes have a lesser document order position than
56
+ * attribute nodes.
57
+ * <LI> An element's namespace nodes are sorted lexicographically by
58
+ * local name (the default namespace node, if one exists, has no
59
+ * local name and is therefore lexicographically least).
60
+ * <LI> An element's attribute nodes are sorted lexicographically with
61
+ * namespace URI as the primary key and local name as the secondary
62
+ * key (an empty namespace URI is lexicographically least).
63
+ * </UL>
64
+ *
65
+ * @param attr0
66
+ * @param attr1
67
+ * @return returns a negative integer, zero, or a positive integer as
68
+ * obj0 is less than, equal to, or greater than obj1
69
+ *
70
+ */
71
+ public int
72
+ compare(Attr attr0, Attr attr1)
73
+ {
74
+ String namespaceURI0 = attr0.getNamespaceURI();
75
+ String namespaceURI1 = attr1.getNamespaceURI();
76
+
77
+ boolean isNamespaceAttr0 = XMLNS.equals(namespaceURI0);
78
+ boolean isNamespaceAttr1 = XMLNS.equals(namespaceURI1);
79
+
80
+ if (isNamespaceAttr0) {
81
+ if (isNamespaceAttr1) {
82
+ // both are namespaces
83
+ String localname0 = attr0.getLocalName();
84
+ String localname1 = attr1.getLocalName();
85
+
86
+ if ("xmlns".equals(localname0)) {
87
+ localname0 = "";
88
+ }
89
+
90
+ if ("xmlns".equals(localname1)) {
91
+ localname1 = "";
92
+ }
93
+
94
+ return localname0.compareTo(localname1);
95
+ }
96
+ // attr0 is a namespace, attr1 is not
97
+ return ATTR0_BEFORE_ATTR1;
98
+ } else if (isNamespaceAttr1) {
99
+ // attr1 is a namespace, attr0 is not
100
+ return ATTR1_BEFORE_ATTR0;
101
+ }
102
+
103
+ // none is a namespace
104
+ if (namespaceURI0 == null) {
105
+ if (namespaceURI1 == null) {
106
+ String name0 = attr0.getName();
107
+ String name1 = attr1.getName();
108
+ return name0.compareTo(name1);
109
+ }
110
+ return ATTR0_BEFORE_ATTR1;
111
+ } else if (namespaceURI1 == null) {
112
+ return ATTR1_BEFORE_ATTR0;
113
+ }
114
+
115
+ int a = namespaceURI0.compareTo(namespaceURI1);
116
+ if (a != 0) {
117
+ return a;
118
+ }
119
+
120
+ return (attr0.getLocalName()).compareTo(attr1.getLocalName());
121
+ }
122
+ }