nokogiri-backport 1.11.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/Gemfile +3 -0
- data/LICENSE-DEPENDENCIES.md +1682 -0
- data/LICENSE.md +9 -0
- data/README.md +272 -0
- data/bin/nokogiri +118 -0
- data/dependencies.yml +74 -0
- data/ext/java/nokogiri/EncodingHandler.java +124 -0
- data/ext/java/nokogiri/HtmlDocument.java +178 -0
- data/ext/java/nokogiri/HtmlElementDescription.java +148 -0
- data/ext/java/nokogiri/HtmlEntityLookup.java +79 -0
- data/ext/java/nokogiri/HtmlSaxParserContext.java +282 -0
- data/ext/java/nokogiri/HtmlSaxPushParser.java +222 -0
- data/ext/java/nokogiri/NokogiriService.java +597 -0
- data/ext/java/nokogiri/XmlAttr.java +162 -0
- data/ext/java/nokogiri/XmlAttributeDecl.java +129 -0
- data/ext/java/nokogiri/XmlCdata.java +82 -0
- data/ext/java/nokogiri/XmlComment.java +97 -0
- data/ext/java/nokogiri/XmlDocument.java +633 -0
- data/ext/java/nokogiri/XmlDocumentFragment.java +185 -0
- data/ext/java/nokogiri/XmlDtd.java +481 -0
- data/ext/java/nokogiri/XmlElement.java +68 -0
- data/ext/java/nokogiri/XmlElementContent.java +382 -0
- data/ext/java/nokogiri/XmlElementDecl.java +147 -0
- data/ext/java/nokogiri/XmlEntityDecl.java +157 -0
- data/ext/java/nokogiri/XmlEntityReference.java +101 -0
- data/ext/java/nokogiri/XmlNamespace.java +199 -0
- data/ext/java/nokogiri/XmlNode.java +1684 -0
- data/ext/java/nokogiri/XmlNodeSet.java +434 -0
- data/ext/java/nokogiri/XmlProcessingInstruction.java +100 -0
- data/ext/java/nokogiri/XmlReader.java +531 -0
- data/ext/java/nokogiri/XmlRelaxng.java +151 -0
- data/ext/java/nokogiri/XmlSaxParserContext.java +374 -0
- data/ext/java/nokogiri/XmlSaxPushParser.java +286 -0
- data/ext/java/nokogiri/XmlSchema.java +388 -0
- data/ext/java/nokogiri/XmlSyntaxError.java +138 -0
- data/ext/java/nokogiri/XmlText.java +110 -0
- data/ext/java/nokogiri/XmlXpathContext.java +301 -0
- data/ext/java/nokogiri/XsltStylesheet.java +347 -0
- data/ext/java/nokogiri/internals/ClosedStreamException.java +10 -0
- data/ext/java/nokogiri/internals/HtmlDomParserContext.java +252 -0
- data/ext/java/nokogiri/internals/IgnoreSchemaErrorsErrorHandler.java +20 -0
- data/ext/java/nokogiri/internals/NokogiriBlockingQueueInputStream.java +151 -0
- data/ext/java/nokogiri/internals/NokogiriDomParser.java +116 -0
- data/ext/java/nokogiri/internals/NokogiriEntityResolver.java +121 -0
- data/ext/java/nokogiri/internals/NokogiriErrorHandler.java +69 -0
- data/ext/java/nokogiri/internals/NokogiriHandler.java +327 -0
- data/ext/java/nokogiri/internals/NokogiriHelpers.java +734 -0
- data/ext/java/nokogiri/internals/NokogiriNamespaceCache.java +217 -0
- data/ext/java/nokogiri/internals/NokogiriNamespaceContext.java +127 -0
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler.java +100 -0
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler4NekoHtml.java +121 -0
- data/ext/java/nokogiri/internals/NokogiriStrictErrorHandler.java +78 -0
- data/ext/java/nokogiri/internals/NokogiriXPathFunction.java +180 -0
- data/ext/java/nokogiri/internals/NokogiriXPathFunctionResolver.java +72 -0
- data/ext/java/nokogiri/internals/NokogiriXPathVariableResolver.java +60 -0
- data/ext/java/nokogiri/internals/NokogiriXsltErrorListener.java +87 -0
- data/ext/java/nokogiri/internals/ParserContext.java +259 -0
- data/ext/java/nokogiri/internals/ReaderNode.java +488 -0
- data/ext/java/nokogiri/internals/SaveContextVisitor.java +778 -0
- data/ext/java/nokogiri/internals/SchemaErrorHandler.java +73 -0
- data/ext/java/nokogiri/internals/XalanDTMManagerPatch.java +168 -0
- data/ext/java/nokogiri/internals/XmlDeclHandler.java +42 -0
- data/ext/java/nokogiri/internals/XmlDomParserContext.java +274 -0
- data/ext/java/nokogiri/internals/XmlSaxParser.java +65 -0
- data/ext/java/nokogiri/internals/c14n/AttrCompare.java +119 -0
- data/ext/java/nokogiri/internals/c14n/C14nHelper.java +159 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalFilter.java +37 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizationException.java +93 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer.java +252 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer11.java +639 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer11_OmitComments.java +38 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer11_WithComments.java +38 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315.java +367 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315Excl.java +295 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315ExclOmitComments.java +40 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315ExclWithComments.java +44 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315OmitComments.java +44 -0
- data/ext/java/nokogiri/internals/c14n/Canonicalizer20010315WithComments.java +43 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizerBase.java +630 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizerPhysical.java +173 -0
- data/ext/java/nokogiri/internals/c14n/CanonicalizerSpi.java +76 -0
- data/ext/java/nokogiri/internals/c14n/Constants.java +42 -0
- data/ext/java/nokogiri/internals/c14n/ElementProxy.java +293 -0
- data/ext/java/nokogiri/internals/c14n/HelperNodeList.java +93 -0
- data/ext/java/nokogiri/internals/c14n/IgnoreAllErrorHandler.java +79 -0
- data/ext/java/nokogiri/internals/c14n/InclusiveNamespaces.java +166 -0
- data/ext/java/nokogiri/internals/c14n/InvalidCanonicalizerException.java +76 -0
- data/ext/java/nokogiri/internals/c14n/NameSpaceSymbTable.java +402 -0
- data/ext/java/nokogiri/internals/c14n/NodeFilter.java +51 -0
- data/ext/java/nokogiri/internals/c14n/UtfHelpper.java +179 -0
- data/ext/java/nokogiri/internals/c14n/XMLUtils.java +507 -0
- data/ext/java/nokogiri/internals/dom2dtm/DOM2DTM.java +1745 -0
- data/ext/java/nokogiri/internals/dom2dtm/DOM2DTMdefaultNamespaceDeclarationNode.java +685 -0
- data/ext/nokogiri/depend +477 -0
- data/ext/nokogiri/extconf.rb +836 -0
- data/ext/nokogiri/html_document.c +171 -0
- data/ext/nokogiri/html_document.h +10 -0
- data/ext/nokogiri/html_element_description.c +279 -0
- data/ext/nokogiri/html_element_description.h +10 -0
- data/ext/nokogiri/html_entity_lookup.c +32 -0
- data/ext/nokogiri/html_entity_lookup.h +8 -0
- data/ext/nokogiri/html_sax_parser_context.c +116 -0
- data/ext/nokogiri/html_sax_parser_context.h +11 -0
- data/ext/nokogiri/html_sax_push_parser.c +87 -0
- data/ext/nokogiri/html_sax_push_parser.h +9 -0
- data/ext/nokogiri/nokogiri.c +135 -0
- data/ext/nokogiri/nokogiri.h +130 -0
- data/ext/nokogiri/xml_attr.c +103 -0
- data/ext/nokogiri/xml_attr.h +9 -0
- data/ext/nokogiri/xml_attribute_decl.c +70 -0
- data/ext/nokogiri/xml_attribute_decl.h +9 -0
- data/ext/nokogiri/xml_cdata.c +62 -0
- data/ext/nokogiri/xml_cdata.h +9 -0
- data/ext/nokogiri/xml_comment.c +69 -0
- data/ext/nokogiri/xml_comment.h +9 -0
- data/ext/nokogiri/xml_document.c +622 -0
- data/ext/nokogiri/xml_document.h +23 -0
- data/ext/nokogiri/xml_document_fragment.c +48 -0
- data/ext/nokogiri/xml_document_fragment.h +10 -0
- data/ext/nokogiri/xml_dtd.c +202 -0
- data/ext/nokogiri/xml_dtd.h +10 -0
- data/ext/nokogiri/xml_element_content.c +123 -0
- data/ext/nokogiri/xml_element_content.h +10 -0
- data/ext/nokogiri/xml_element_decl.c +69 -0
- data/ext/nokogiri/xml_element_decl.h +9 -0
- data/ext/nokogiri/xml_encoding_handler.c +79 -0
- data/ext/nokogiri/xml_encoding_handler.h +8 -0
- data/ext/nokogiri/xml_entity_decl.c +110 -0
- data/ext/nokogiri/xml_entity_decl.h +10 -0
- data/ext/nokogiri/xml_entity_reference.c +52 -0
- data/ext/nokogiri/xml_entity_reference.h +9 -0
- data/ext/nokogiri/xml_io.c +63 -0
- data/ext/nokogiri/xml_io.h +11 -0
- data/ext/nokogiri/xml_libxml2_hacks.c +112 -0
- data/ext/nokogiri/xml_libxml2_hacks.h +12 -0
- data/ext/nokogiri/xml_namespace.c +111 -0
- data/ext/nokogiri/xml_namespace.h +14 -0
- data/ext/nokogiri/xml_node.c +1773 -0
- data/ext/nokogiri/xml_node.h +13 -0
- data/ext/nokogiri/xml_node_set.c +486 -0
- data/ext/nokogiri/xml_node_set.h +12 -0
- data/ext/nokogiri/xml_processing_instruction.c +56 -0
- data/ext/nokogiri/xml_processing_instruction.h +9 -0
- data/ext/nokogiri/xml_reader.c +657 -0
- data/ext/nokogiri/xml_reader.h +10 -0
- data/ext/nokogiri/xml_relax_ng.c +179 -0
- data/ext/nokogiri/xml_relax_ng.h +9 -0
- data/ext/nokogiri/xml_sax_parser.c +305 -0
- data/ext/nokogiri/xml_sax_parser.h +39 -0
- data/ext/nokogiri/xml_sax_parser_context.c +262 -0
- data/ext/nokogiri/xml_sax_parser_context.h +10 -0
- data/ext/nokogiri/xml_sax_push_parser.c +159 -0
- data/ext/nokogiri/xml_sax_push_parser.h +9 -0
- data/ext/nokogiri/xml_schema.c +276 -0
- data/ext/nokogiri/xml_schema.h +9 -0
- data/ext/nokogiri/xml_syntax_error.c +64 -0
- data/ext/nokogiri/xml_syntax_error.h +13 -0
- data/ext/nokogiri/xml_text.c +52 -0
- data/ext/nokogiri/xml_text.h +9 -0
- data/ext/nokogiri/xml_xpath_context.c +374 -0
- data/ext/nokogiri/xml_xpath_context.h +10 -0
- data/ext/nokogiri/xslt_stylesheet.c +263 -0
- data/ext/nokogiri/xslt_stylesheet.h +14 -0
- data/lib/isorelax.jar +0 -0
- data/lib/jing.jar +0 -0
- data/lib/nekodtd.jar +0 -0
- data/lib/nekohtml.jar +0 -0
- data/lib/nokogiri/css/node.rb +53 -0
- data/lib/nokogiri/css/parser.rb +751 -0
- data/lib/nokogiri/css/parser.y +272 -0
- data/lib/nokogiri/css/parser_extras.rb +94 -0
- data/lib/nokogiri/css/syntax_error.rb +8 -0
- data/lib/nokogiri/css/tokenizer.rb +154 -0
- data/lib/nokogiri/css/tokenizer.rex +55 -0
- data/lib/nokogiri/css/xpath_visitor.rb +260 -0
- data/lib/nokogiri/css.rb +28 -0
- data/lib/nokogiri/decorators/slop.rb +43 -0
- data/lib/nokogiri/html/builder.rb +36 -0
- data/lib/nokogiri/html/document.rb +322 -0
- data/lib/nokogiri/html/document_fragment.rb +50 -0
- data/lib/nokogiri/html/element_description.rb +24 -0
- data/lib/nokogiri/html/element_description_defaults.rb +672 -0
- data/lib/nokogiri/html/entity_lookup.rb +14 -0
- data/lib/nokogiri/html/sax/parser.rb +63 -0
- data/lib/nokogiri/html/sax/parser_context.rb +17 -0
- data/lib/nokogiri/html/sax/push_parser.rb +37 -0
- data/lib/nokogiri/html.rb +38 -0
- data/lib/nokogiri/jruby/dependencies.rb +20 -0
- data/lib/nokogiri/syntax_error.rb +5 -0
- data/lib/nokogiri/version/constant.rb +5 -0
- data/lib/nokogiri/version/info.rb +182 -0
- data/lib/nokogiri/version.rb +3 -0
- data/lib/nokogiri/xml/attr.rb +15 -0
- data/lib/nokogiri/xml/attribute_decl.rb +19 -0
- data/lib/nokogiri/xml/builder.rb +447 -0
- data/lib/nokogiri/xml/cdata.rb +12 -0
- data/lib/nokogiri/xml/character_data.rb +8 -0
- data/lib/nokogiri/xml/document.rb +290 -0
- data/lib/nokogiri/xml/document_fragment.rb +159 -0
- data/lib/nokogiri/xml/dtd.rb +33 -0
- data/lib/nokogiri/xml/element_content.rb +37 -0
- data/lib/nokogiri/xml/element_decl.rb +14 -0
- data/lib/nokogiri/xml/entity_decl.rb +20 -0
- data/lib/nokogiri/xml/entity_reference.rb +19 -0
- data/lib/nokogiri/xml/namespace.rb +14 -0
- data/lib/nokogiri/xml/node/save_options.rb +62 -0
- data/lib/nokogiri/xml/node.rb +1240 -0
- data/lib/nokogiri/xml/node_set.rb +372 -0
- data/lib/nokogiri/xml/notation.rb +7 -0
- data/lib/nokogiri/xml/parse_options.rb +127 -0
- data/lib/nokogiri/xml/pp/character_data.rb +19 -0
- data/lib/nokogiri/xml/pp/node.rb +57 -0
- data/lib/nokogiri/xml/pp.rb +3 -0
- data/lib/nokogiri/xml/processing_instruction.rb +9 -0
- data/lib/nokogiri/xml/reader.rb +116 -0
- data/lib/nokogiri/xml/relax_ng.rb +37 -0
- data/lib/nokogiri/xml/sax/document.rb +172 -0
- data/lib/nokogiri/xml/sax/parser.rb +123 -0
- data/lib/nokogiri/xml/sax/parser_context.rb +17 -0
- data/lib/nokogiri/xml/sax/push_parser.rb +61 -0
- data/lib/nokogiri/xml/sax.rb +5 -0
- data/lib/nokogiri/xml/schema.rb +72 -0
- data/lib/nokogiri/xml/searchable.rb +239 -0
- data/lib/nokogiri/xml/syntax_error.rb +71 -0
- data/lib/nokogiri/xml/text.rb +10 -0
- data/lib/nokogiri/xml/xpath/syntax_error.rb +12 -0
- data/lib/nokogiri/xml/xpath.rb +11 -0
- data/lib/nokogiri/xml/xpath_context.rb +17 -0
- data/lib/nokogiri/xml.rb +76 -0
- data/lib/nokogiri/xslt/stylesheet.rb +26 -0
- data/lib/nokogiri/xslt.rb +57 -0
- data/lib/nokogiri.rb +144 -0
- data/lib/serializer.jar +0 -0
- data/lib/xalan.jar +0 -0
- data/lib/xercesImpl.jar +0 -0
- data/lib/xml-apis.jar +0 -0
- data/lib/xsd/xmlparser/nokogiri.rb +103 -0
- metadata +531 -0
@@ -0,0 +1,69 @@
|
|
1
|
+
/**
|
2
|
+
* (The MIT License)
|
3
|
+
*
|
4
|
+
* Copyright (c) 2008 - 2011:
|
5
|
+
*
|
6
|
+
* * {Aaron Patterson}[http://tenderlovemaking.com]
|
7
|
+
* * {Mike Dalessio}[http://mike.daless.io]
|
8
|
+
* * {Charles Nutter}[http://blog.headius.com]
|
9
|
+
* * {Sergio Arbeo}[http://www.serabe.com]
|
10
|
+
* * {Patrick Mahoney}[http://polycrystal.org]
|
11
|
+
* * {Yoko Harada}[http://yokolet.blogspot.com]
|
12
|
+
*
|
13
|
+
* Permission is hereby granted, free of charge, to any person obtaining
|
14
|
+
* a copy of this software and associated documentation files (the
|
15
|
+
* 'Software'), to deal in the Software without restriction, including
|
16
|
+
* without limitation the rights to use, copy, modify, merge, publish,
|
17
|
+
* distribute, sublicense, and/or sell copies of the Software, and to
|
18
|
+
* permit persons to whom the Software is furnished to do so, subject to
|
19
|
+
* the following conditions:
|
20
|
+
*
|
21
|
+
* The above copyright notice and this permission notice shall be
|
22
|
+
* included in all copies or substantial portions of the Software.
|
23
|
+
*
|
24
|
+
* THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
25
|
+
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
26
|
+
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
27
|
+
* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
28
|
+
* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
|
29
|
+
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
|
30
|
+
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
31
|
+
*/
|
32
|
+
|
33
|
+
package nokogiri.internals;
|
34
|
+
|
35
|
+
import java.util.ArrayList;
|
36
|
+
import java.util.List;
|
37
|
+
|
38
|
+
import org.apache.xerces.xni.parser.XMLErrorHandler;
|
39
|
+
import org.xml.sax.ErrorHandler;
|
40
|
+
|
41
|
+
/**
|
42
|
+
* Super class of error handlers.
|
43
|
+
*
|
44
|
+
* XMLErrorHandler is used by nokogiri.internals.HtmlDomParserContext since NekoHtml
|
45
|
+
* uses this type of the error handler.
|
46
|
+
*
|
47
|
+
* @author sergio
|
48
|
+
* @author Yoko Harada <yokolet@gmail.com>
|
49
|
+
*/
|
50
|
+
public abstract class NokogiriErrorHandler implements ErrorHandler, XMLErrorHandler {
|
51
|
+
protected final List<Exception> errors;
|
52
|
+
protected boolean noerror;
|
53
|
+
protected boolean nowarning;
|
54
|
+
|
55
|
+
public NokogiriErrorHandler(boolean noerror, boolean nowarning) {
|
56
|
+
this.errors = new ArrayList<Exception>(4);
|
57
|
+
this.noerror = noerror;
|
58
|
+
this.nowarning = nowarning;
|
59
|
+
}
|
60
|
+
|
61
|
+
List<Exception> getErrors() { return errors; }
|
62
|
+
|
63
|
+
public void addError(Exception ex) { errors.add(ex); }
|
64
|
+
|
65
|
+
protected boolean usesNekoHtml(String domain) {
|
66
|
+
return "http://cyberneko.org/html".equals(domain);
|
67
|
+
}
|
68
|
+
|
69
|
+
}
|
@@ -0,0 +1,327 @@
|
|
1
|
+
/**
|
2
|
+
* (The MIT License)
|
3
|
+
*
|
4
|
+
* Copyright (c) 2008 - 2011:
|
5
|
+
*
|
6
|
+
* * {Aaron Patterson}[http://tenderlovemaking.com]
|
7
|
+
* * {Mike Dalessio}[http://mike.daless.io]
|
8
|
+
* * {Charles Nutter}[http://blog.headius.com]
|
9
|
+
* * {Sergio Arbeo}[http://www.serabe.com]
|
10
|
+
* * {Patrick Mahoney}[http://polycrystal.org]
|
11
|
+
* * {Yoko Harada}[http://yokolet.blogspot.com]
|
12
|
+
*
|
13
|
+
* Permission is hereby granted, free of charge, to any person obtaining
|
14
|
+
* a copy of this software and associated documentation files (the
|
15
|
+
* 'Software'), to deal in the Software without restriction, including
|
16
|
+
* without limitation the rights to use, copy, modify, merge, publish,
|
17
|
+
* distribute, sublicense, and/or sell copies of the Software, and to
|
18
|
+
* permit persons to whom the Software is furnished to do so, subject to
|
19
|
+
* the following conditions:
|
20
|
+
*
|
21
|
+
* The above copyright notice and this permission notice shall be
|
22
|
+
* included in all copies or substantial portions of the Software.
|
23
|
+
*
|
24
|
+
* THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
25
|
+
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
26
|
+
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
27
|
+
* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
28
|
+
* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
|
29
|
+
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
|
30
|
+
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
31
|
+
*/
|
32
|
+
|
33
|
+
package nokogiri.internals;
|
34
|
+
|
35
|
+
import static nokogiri.internals.NokogiriHelpers.getLocalPart;
|
36
|
+
import static nokogiri.internals.NokogiriHelpers.getPrefix;
|
37
|
+
import static nokogiri.internals.NokogiriHelpers.isNamespace;
|
38
|
+
import static nokogiri.internals.NokogiriHelpers.stringOrNil;
|
39
|
+
|
40
|
+
import java.util.Arrays;
|
41
|
+
import java.util.HashSet;
|
42
|
+
import java.util.LinkedList;
|
43
|
+
import java.util.Set;
|
44
|
+
|
45
|
+
import org.jruby.Ruby;
|
46
|
+
import org.jruby.RubyArray;
|
47
|
+
import org.jruby.RubyClass;
|
48
|
+
import org.jruby.exceptions.RaiseException;
|
49
|
+
import org.jruby.runtime.Helpers;
|
50
|
+
import org.jruby.runtime.ThreadContext;
|
51
|
+
import org.jruby.runtime.builtin.IRubyObject;
|
52
|
+
import org.xml.sax.Attributes;
|
53
|
+
import org.xml.sax.Locator;
|
54
|
+
import org.xml.sax.SAXException;
|
55
|
+
import org.xml.sax.SAXParseException;
|
56
|
+
import org.xml.sax.ext.DefaultHandler2;
|
57
|
+
|
58
|
+
import nokogiri.XmlSyntaxError;
|
59
|
+
|
60
|
+
/**
|
61
|
+
* A handler for SAX parsing.
|
62
|
+
*
|
63
|
+
* @author sergio
|
64
|
+
* @author Yoko Harada <yokolet@gmail.com>
|
65
|
+
*/
|
66
|
+
public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler {
|
67
|
+
|
68
|
+
StringBuilder charactersBuilder;
|
69
|
+
private final Ruby runtime;
|
70
|
+
private final RubyClass attrClass;
|
71
|
+
private final IRubyObject object;
|
72
|
+
|
73
|
+
/**
|
74
|
+
* Stores parse errors with the most-recent error last.
|
75
|
+
*
|
76
|
+
* TODO: should these be stored in the document 'errors' array?
|
77
|
+
* Currently only string messages are stored there.
|
78
|
+
*/
|
79
|
+
private final LinkedList<RaiseException> errors = new LinkedList<RaiseException>();
|
80
|
+
|
81
|
+
private Locator locator;
|
82
|
+
private boolean needEmptyAttrCheck;
|
83
|
+
|
84
|
+
public NokogiriHandler(Ruby runtime, IRubyObject object) {
|
85
|
+
assert object != null;
|
86
|
+
this.runtime = runtime;
|
87
|
+
this.attrClass = (RubyClass) runtime.getClassFromPath("Nokogiri::XML::SAX::Parser::Attribute");
|
88
|
+
this.object = object;
|
89
|
+
charactersBuilder = new StringBuilder();
|
90
|
+
String objectName = object.getMetaClass().getName();
|
91
|
+
if ("Nokogiri::HTML::SAX::Parser".equals(objectName)) needEmptyAttrCheck = true;
|
92
|
+
}
|
93
|
+
|
94
|
+
@Override
|
95
|
+
public void skippedEntity(String skippedEntity) {
|
96
|
+
call("error", runtime.newString("Entity '" + skippedEntity + "' not defined\n"));
|
97
|
+
}
|
98
|
+
|
99
|
+
@Override
|
100
|
+
public void setDocumentLocator(Locator locator) {
|
101
|
+
this.locator = locator;
|
102
|
+
}
|
103
|
+
|
104
|
+
@Override
|
105
|
+
public void startDocument() {
|
106
|
+
call("start_document");
|
107
|
+
}
|
108
|
+
|
109
|
+
@Override
|
110
|
+
public void xmlDecl(String version, String encoding, String standalone) {
|
111
|
+
call("xmldecl", stringOrNil(runtime, version), stringOrNil(runtime, encoding), stringOrNil(runtime, standalone));
|
112
|
+
}
|
113
|
+
|
114
|
+
@Override
|
115
|
+
public void endDocument() {
|
116
|
+
populateCharacters();
|
117
|
+
call("end_document");
|
118
|
+
}
|
119
|
+
|
120
|
+
@Override
|
121
|
+
public void processingInstruction(String target, String data) {
|
122
|
+
call("processing_instruction", runtime.newString(target), runtime.newString(data));
|
123
|
+
}
|
124
|
+
|
125
|
+
/*
|
126
|
+
* This calls "start_element_namespace".
|
127
|
+
*
|
128
|
+
* Attributes that define namespaces are passed in a separate
|
129
|
+
* array of <code>[:prefix, :uri]</code> arrays and are not
|
130
|
+
* passed with the other attributes.
|
131
|
+
*/
|
132
|
+
@Override
|
133
|
+
public void startElement(String uri, String localName, String qName, Attributes attrs) throws SAXException {
|
134
|
+
final Ruby runtime = this.runtime;
|
135
|
+
final ThreadContext context = runtime.getCurrentContext();
|
136
|
+
|
137
|
+
// for attributes other than namespace attrs
|
138
|
+
RubyArray rubyAttr = RubyArray.newArray(runtime);
|
139
|
+
// for namespace defining attributes
|
140
|
+
RubyArray rubyNSAttr = RubyArray.newArray(runtime);
|
141
|
+
|
142
|
+
boolean fromFragmentHandler = false; // isFromFragmentHandler();
|
143
|
+
|
144
|
+
for (int i = 0; i < attrs.getLength(); i++) {
|
145
|
+
String u = attrs.getURI(i);
|
146
|
+
String qn = attrs.getQName(i);
|
147
|
+
String ln = attrs.getLocalName(i);
|
148
|
+
String val = attrs.getValue(i);
|
149
|
+
String pre;
|
150
|
+
|
151
|
+
pre = getPrefix(qn);
|
152
|
+
if (ln == null || ln.isEmpty()) ln = getLocalPart(qn);
|
153
|
+
|
154
|
+
if (isNamespace(qn) && !fromFragmentHandler) {
|
155
|
+
// I haven't figured the reason out yet, but, in somewhere,
|
156
|
+
// namespace is converted to array in array and cause
|
157
|
+
// TypeError at line 45 in fragment_handler.rb
|
158
|
+
if (ln.equals("xmlns")) ln = null;
|
159
|
+
rubyNSAttr.append( runtime.newArray( stringOrNil(runtime, ln), runtime.newString(val) ) );
|
160
|
+
} else {
|
161
|
+
IRubyObject[] args = null;
|
162
|
+
if (needEmptyAttrCheck) {
|
163
|
+
if (isEmptyAttr(ln)) {
|
164
|
+
args = new IRubyObject[] {
|
165
|
+
stringOrNil(runtime, ln),
|
166
|
+
stringOrNil(runtime, pre),
|
167
|
+
stringOrNil(runtime, u)
|
168
|
+
};
|
169
|
+
}
|
170
|
+
}
|
171
|
+
if (args == null) {
|
172
|
+
args = new IRubyObject[] {
|
173
|
+
stringOrNil(runtime, ln),
|
174
|
+
stringOrNil(runtime, pre),
|
175
|
+
stringOrNil(runtime, u),
|
176
|
+
stringOrNil(runtime, val)
|
177
|
+
};
|
178
|
+
}
|
179
|
+
|
180
|
+
rubyAttr.append( Helpers.invoke(context, attrClass, "new", args) );
|
181
|
+
}
|
182
|
+
}
|
183
|
+
|
184
|
+
if (localName == null || localName.isEmpty()) localName = getLocalPart(qName);
|
185
|
+
populateCharacters();
|
186
|
+
call("start_element_namespace",
|
187
|
+
stringOrNil(runtime, localName),
|
188
|
+
rubyAttr,
|
189
|
+
stringOrNil(runtime, getPrefix(qName)),
|
190
|
+
stringOrNil(runtime, uri),
|
191
|
+
rubyNSAttr);
|
192
|
+
}
|
193
|
+
|
194
|
+
static final Set<String> EMPTY_ATTRS;
|
195
|
+
static {
|
196
|
+
final String[] emptyAttrs = {
|
197
|
+
"checked", "compact", "declare", "defer", "disabled", "ismap", "multiple",
|
198
|
+
"noresize", "nohref", "noshade", "nowrap", "readonly", "selected"
|
199
|
+
};
|
200
|
+
EMPTY_ATTRS = new HashSet<String>(Arrays.asList(emptyAttrs));
|
201
|
+
}
|
202
|
+
|
203
|
+
private static boolean isEmptyAttr(String name) {
|
204
|
+
return EMPTY_ATTRS.contains(name);
|
205
|
+
}
|
206
|
+
|
207
|
+
public final Integer getLine() { // -1 if none is available
|
208
|
+
final int line = locator.getLineNumber();
|
209
|
+
return line == -1 ? null : line;
|
210
|
+
}
|
211
|
+
|
212
|
+
public final Integer getColumn() { // -1 if none is available
|
213
|
+
final int column = locator.getColumnNumber();
|
214
|
+
return column == -1 ? null : column - 1;
|
215
|
+
}
|
216
|
+
|
217
|
+
@Override
|
218
|
+
public void endElement(String uri, String localName, String qName) {
|
219
|
+
populateCharacters();
|
220
|
+
call("end_element_namespace",
|
221
|
+
stringOrNil(runtime, localName),
|
222
|
+
stringOrNil(runtime, getPrefix(qName)),
|
223
|
+
stringOrNil(runtime, uri));
|
224
|
+
}
|
225
|
+
|
226
|
+
@Override
|
227
|
+
public void characters(char[] ch, int start, int length) {
|
228
|
+
charactersBuilder.append(ch, start, length);
|
229
|
+
}
|
230
|
+
|
231
|
+
@Override
|
232
|
+
public void comment(char[] ch, int start, int length) {
|
233
|
+
populateCharacters();
|
234
|
+
call("comment", runtime.newString(new String(ch, start, length)));
|
235
|
+
}
|
236
|
+
|
237
|
+
@Override
|
238
|
+
public void startCDATA() {
|
239
|
+
populateCharacters();
|
240
|
+
}
|
241
|
+
|
242
|
+
@Override
|
243
|
+
public void endCDATA() {
|
244
|
+
call("cdata_block", runtime.newString(charactersBuilder.toString()));
|
245
|
+
charactersBuilder.setLength(0);
|
246
|
+
}
|
247
|
+
|
248
|
+
void handleError(SAXParseException ex) {
|
249
|
+
try {
|
250
|
+
final String msg = ex.getMessage();
|
251
|
+
call("error", runtime.newString(msg == null ? "" : msg));
|
252
|
+
addError(XmlSyntaxError.createError(runtime, ex).toThrowable());
|
253
|
+
} catch( RaiseException e) {
|
254
|
+
addError(e);
|
255
|
+
throw e;
|
256
|
+
}
|
257
|
+
}
|
258
|
+
|
259
|
+
@Override
|
260
|
+
public void error(SAXParseException ex) {
|
261
|
+
handleError(ex);
|
262
|
+
}
|
263
|
+
|
264
|
+
@Override
|
265
|
+
public void fatalError(SAXParseException ex) {
|
266
|
+
handleError(ex);
|
267
|
+
}
|
268
|
+
|
269
|
+
@Override
|
270
|
+
public void warning(SAXParseException ex) {
|
271
|
+
final String msg = ex.getMessage();
|
272
|
+
call("warning", runtime.newString(msg == null ? "" : msg));
|
273
|
+
}
|
274
|
+
|
275
|
+
protected synchronized void addError(RaiseException e) {
|
276
|
+
errors.add(e);
|
277
|
+
}
|
278
|
+
|
279
|
+
public synchronized int getErrorCount() {
|
280
|
+
return errors.size();
|
281
|
+
}
|
282
|
+
|
283
|
+
public synchronized RaiseException getLastError() {
|
284
|
+
return errors.getLast();
|
285
|
+
}
|
286
|
+
|
287
|
+
private void call(String methodName) {
|
288
|
+
ThreadContext context = runtime.getCurrentContext();
|
289
|
+
Helpers.invoke(context, document(context), methodName);
|
290
|
+
}
|
291
|
+
|
292
|
+
private void call(String methodName, IRubyObject argument) {
|
293
|
+
ThreadContext context = runtime.getCurrentContext();
|
294
|
+
Helpers.invoke(context, document(context), methodName, argument);
|
295
|
+
}
|
296
|
+
|
297
|
+
private void call(String methodName, IRubyObject arg1, IRubyObject arg2) {
|
298
|
+
ThreadContext context = runtime.getCurrentContext();
|
299
|
+
Helpers.invoke(context, document(context), methodName, arg1, arg2);
|
300
|
+
}
|
301
|
+
|
302
|
+
private void call(String methodName, IRubyObject arg1, IRubyObject arg2, IRubyObject arg3) {
|
303
|
+
ThreadContext context = runtime.getCurrentContext();
|
304
|
+
Helpers.invoke(context, document(context), methodName, arg1, arg2, arg3);
|
305
|
+
}
|
306
|
+
|
307
|
+
private void call(String methodName,
|
308
|
+
IRubyObject arg0,
|
309
|
+
IRubyObject arg1,
|
310
|
+
IRubyObject arg2,
|
311
|
+
IRubyObject arg3,
|
312
|
+
IRubyObject arg4) {
|
313
|
+
ThreadContext context = runtime.getCurrentContext();
|
314
|
+
Helpers.invoke(context, document(context), methodName, arg0, arg1, arg2, arg3, arg4);
|
315
|
+
}
|
316
|
+
|
317
|
+
private IRubyObject document(ThreadContext context) {
|
318
|
+
return object.getInstanceVariables().getInstanceVariable("@document");
|
319
|
+
}
|
320
|
+
|
321
|
+
protected void populateCharacters() {
|
322
|
+
if (charactersBuilder.length() > 0) {
|
323
|
+
call("characters", runtime.newString(charactersBuilder.toString()));
|
324
|
+
charactersBuilder.setLength(0);
|
325
|
+
}
|
326
|
+
}
|
327
|
+
}
|