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 +1,222 @@
|
|
1
|
+
/**
|
2
|
+
* (The MIT License)
|
3
|
+
*
|
4
|
+
* Copyright (c) 2008 - 2012:
|
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;
|
34
|
+
|
35
|
+
import static nokogiri.XmlSaxPushParser.terminateExecution;
|
36
|
+
import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
|
37
|
+
import static org.jruby.runtime.Helpers.invoke;
|
38
|
+
|
39
|
+
import java.io.ByteArrayInputStream;
|
40
|
+
import java.io.InputStream;
|
41
|
+
import java.io.IOException;
|
42
|
+
import java.util.concurrent.Callable;
|
43
|
+
import java.util.concurrent.ExecutionException;
|
44
|
+
import java.util.concurrent.ExecutorService;
|
45
|
+
import java.util.concurrent.Executors;
|
46
|
+
import java.util.concurrent.Future;
|
47
|
+
import java.util.concurrent.FutureTask;
|
48
|
+
import java.util.concurrent.ThreadFactory;
|
49
|
+
|
50
|
+
import nokogiri.internals.*;
|
51
|
+
|
52
|
+
import org.jruby.Ruby;
|
53
|
+
import org.jruby.RubyClass;
|
54
|
+
import org.jruby.RubyObject;
|
55
|
+
import org.jruby.anno.JRubyClass;
|
56
|
+
import org.jruby.anno.JRubyMethod;
|
57
|
+
import org.jruby.exceptions.RaiseException;
|
58
|
+
import org.jruby.runtime.ThreadContext;
|
59
|
+
import org.jruby.runtime.builtin.IRubyObject;
|
60
|
+
|
61
|
+
/**
|
62
|
+
* Class for Nokogiri::HTML::SAX::PushParser
|
63
|
+
*
|
64
|
+
* @author
|
65
|
+
* @author Piotr Szmielew <p.szmielew@ava.waw.pl> - based on Nokogiri::XML::SAX::PushParser
|
66
|
+
*/
|
67
|
+
@JRubyClass(name="Nokogiri::HTML::SAX::PushParser")
|
68
|
+
public class HtmlSaxPushParser extends RubyObject {
|
69
|
+
ParserContext.Options options;
|
70
|
+
IRubyObject saxParser;
|
71
|
+
|
72
|
+
NokogiriBlockingQueueInputStream stream;
|
73
|
+
|
74
|
+
private ParserTask parserTask = null;
|
75
|
+
private FutureTask<HtmlSaxParserContext> futureTask = null;
|
76
|
+
private ExecutorService executor = null;
|
77
|
+
|
78
|
+
public HtmlSaxPushParser(Ruby ruby, RubyClass rubyClass) {
|
79
|
+
super(ruby, rubyClass);
|
80
|
+
}
|
81
|
+
|
82
|
+
@Override
|
83
|
+
public void finalize() {
|
84
|
+
try {
|
85
|
+
terminateImpl();
|
86
|
+
}
|
87
|
+
catch (Exception e) { /* ignored */ }
|
88
|
+
}
|
89
|
+
|
90
|
+
@JRubyMethod
|
91
|
+
public IRubyObject initialize_native(final ThreadContext context,
|
92
|
+
IRubyObject saxParser,
|
93
|
+
IRubyObject fileName,
|
94
|
+
IRubyObject encoding) {
|
95
|
+
// NOTE: Silently skips provided encoding
|
96
|
+
options = new ParserContext.Options(0);
|
97
|
+
this.saxParser = saxParser;
|
98
|
+
return this;
|
99
|
+
}
|
100
|
+
|
101
|
+
private transient IRubyObject parse_options;
|
102
|
+
|
103
|
+
private IRubyObject parse_options(final ThreadContext context) {
|
104
|
+
if (parse_options == null) {
|
105
|
+
parse_options = invoke(context, context.runtime.getClassFromPath("Nokogiri::XML::ParseOptions"), "new");
|
106
|
+
}
|
107
|
+
return parse_options;
|
108
|
+
}
|
109
|
+
|
110
|
+
@JRubyMethod(name="options")
|
111
|
+
public IRubyObject getOptions(ThreadContext context) {
|
112
|
+
return invoke(context, parse_options(context), "options");
|
113
|
+
}
|
114
|
+
|
115
|
+
@JRubyMethod(name="options=")
|
116
|
+
public IRubyObject setOptions(ThreadContext context, IRubyObject opts) {
|
117
|
+
invoke(context, parse_options(context), "options=", opts);
|
118
|
+
options = new ParserContext.Options(opts.convertToInteger().getLongValue());
|
119
|
+
return getOptions(context);
|
120
|
+
}
|
121
|
+
|
122
|
+
@JRubyMethod
|
123
|
+
public IRubyObject native_write(ThreadContext context, IRubyObject chunk, IRubyObject isLast) {
|
124
|
+
try {
|
125
|
+
initialize_task(context);
|
126
|
+
} catch (IOException e) {
|
127
|
+
throw context.getRuntime().newRuntimeError(e.getMessage());
|
128
|
+
}
|
129
|
+
final ByteArrayInputStream data = NokogiriHelpers.stringBytesToStream(chunk);
|
130
|
+
if (data == null) {
|
131
|
+
terminateTask(context.runtime);
|
132
|
+
throw XmlSyntaxError.createHTMLSyntaxError(context.runtime).toThrowable(); // Nokogiri::HTML::SyntaxError
|
133
|
+
}
|
134
|
+
|
135
|
+
int errorCount0 = parserTask.getErrorCount();
|
136
|
+
|
137
|
+
if (isLast.isTrue()) {
|
138
|
+
IRubyObject document = invoke(context, this, "document");
|
139
|
+
invoke(context, document, "end_document");
|
140
|
+
terminateTask(context.runtime);
|
141
|
+
} else {
|
142
|
+
try {
|
143
|
+
Future<Void> task = stream.addChunk(data);
|
144
|
+
task.get();
|
145
|
+
}
|
146
|
+
catch (ClosedStreamException ex) {
|
147
|
+
// this means the stream is closed, ignore this exception
|
148
|
+
}
|
149
|
+
catch (Exception e) {
|
150
|
+
throw context.runtime.newRuntimeError(e.getMessage());
|
151
|
+
}
|
152
|
+
|
153
|
+
}
|
154
|
+
|
155
|
+
if (!options.recover && parserTask.getErrorCount() > errorCount0) {
|
156
|
+
terminateTask(context.runtime);
|
157
|
+
throw parserTask.getLastError();
|
158
|
+
}
|
159
|
+
|
160
|
+
return this;
|
161
|
+
}
|
162
|
+
|
163
|
+
@SuppressWarnings("unchecked")
|
164
|
+
private void initialize_task(ThreadContext context) throws IOException {
|
165
|
+
if (futureTask == null || stream == null) {
|
166
|
+
stream = new NokogiriBlockingQueueInputStream();
|
167
|
+
|
168
|
+
assert saxParser != null : "saxParser null";
|
169
|
+
parserTask = new ParserTask(context, saxParser, stream);
|
170
|
+
futureTask = new FutureTask<HtmlSaxParserContext>((Callable) parserTask);
|
171
|
+
executor = Executors.newSingleThreadExecutor(new ThreadFactory() {
|
172
|
+
@Override
|
173
|
+
public Thread newThread(Runnable r) {
|
174
|
+
Thread t = new Thread(r);
|
175
|
+
t.setName("HtmlSaxPushParser");
|
176
|
+
t.setDaemon(true);
|
177
|
+
return t;
|
178
|
+
}
|
179
|
+
});
|
180
|
+
executor.submit(futureTask);
|
181
|
+
}
|
182
|
+
}
|
183
|
+
|
184
|
+
private void terminateTask(final Ruby runtime) {
|
185
|
+
if (executor == null) return;
|
186
|
+
|
187
|
+
try {
|
188
|
+
terminateImpl();
|
189
|
+
}
|
190
|
+
catch (InterruptedException e) {
|
191
|
+
throw runtime.newRuntimeError(e.toString());
|
192
|
+
}
|
193
|
+
catch (Exception e) {
|
194
|
+
throw runtime.newRuntimeError(e.toString());
|
195
|
+
}
|
196
|
+
}
|
197
|
+
|
198
|
+
private synchronized void terminateImpl() throws InterruptedException, ExecutionException {
|
199
|
+
terminateExecution(executor, stream, futureTask);
|
200
|
+
|
201
|
+
executor = null; stream = null; futureTask = null;
|
202
|
+
}
|
203
|
+
|
204
|
+
private static HtmlSaxParserContext parse(final Ruby runtime, final InputStream stream) {
|
205
|
+
RubyClass klazz = getNokogiriClass(runtime, "Nokogiri::HTML::SAX::ParserContext");
|
206
|
+
return HtmlSaxParserContext.parse_stream(runtime, klazz, stream);
|
207
|
+
}
|
208
|
+
|
209
|
+
static class ParserTask extends XmlSaxPushParser.ParserTask /* <HtmlSaxPushParser> */ {
|
210
|
+
|
211
|
+
private ParserTask(ThreadContext context, IRubyObject handler, InputStream stream) {
|
212
|
+
super(context, handler, parse(context.runtime, stream), stream);
|
213
|
+
}
|
214
|
+
|
215
|
+
@Override
|
216
|
+
public HtmlSaxParserContext call() throws Exception {
|
217
|
+
return (HtmlSaxParserContext) super.call();
|
218
|
+
}
|
219
|
+
|
220
|
+
}
|
221
|
+
|
222
|
+
}
|