nokogiri 1.5.0.beta.4 → 1.5.0
Sign up to get free protection for your applications and to get access to all the features.
Potentially problematic release.
This version of nokogiri might be problematic. Click here for more details.
- data/.gemtest +0 -0
- data/CHANGELOG.ja.rdoc +34 -0
- data/CHANGELOG.rdoc +40 -1
- data/Manifest.txt +11 -2
- data/README.rdoc +1 -1
- data/Rakefile +100 -104
- data/bin/nokogiri +1 -2
- data/ext/nokogiri/nokogiri.c +24 -1
- data/ext/nokogiri/xml_io.c +32 -7
- data/ext/nokogiri/xml_node.c +14 -13
- data/ext/nokogiri/xml_sax_parser.c +9 -4
- data/ext/nokogiri/xslt_stylesheet.c +7 -1
- data/lib/nokogiri.rb +3 -22
- data/lib/nokogiri/css.rb +4 -0
- data/lib/nokogiri/html/document.rb +10 -14
- data/lib/nokogiri/version.rb +76 -23
- data/lib/nokogiri/xml/builder.rb +7 -0
- data/lib/nokogiri/xml/document.rb +17 -1
- data/lib/nokogiri/xml/document_fragment.rb +14 -0
- data/lib/nokogiri/xml/node.rb +36 -28
- data/lib/nokogiri/xml/node/save_options.rb +17 -1
- data/lib/nokogiri/xml/node_set.rb +7 -0
- data/lib/nokogiri/xml/parse_options.rb +8 -0
- data/lib/nokogiri/xml/reader.rb +6 -6
- data/lib/nokogiri/xml/schema.rb +7 -1
- data/nokogiri_help_responses.md +40 -0
- data/tasks/cross_compile.rb +134 -159
- data/tasks/nokogiri.org.rb +18 -0
- data/tasks/test.rb +1 -1
- data/test/files/encoding.html +82 -0
- data/test/files/encoding.xhtml +84 -0
- data/test/files/metacharset.html +10 -0
- data/test/files/noencoding.html +47 -0
- data/test/helper.rb +2 -0
- data/test/html/test_document.rb +15 -0
- data/test/html/test_document_encoding.rb +13 -0
- data/test/test_memory_leak.rb +20 -0
- data/test/test_reader.rb +22 -0
- data/test/test_xslt_transforms.rb +6 -2
- data/test/xml/node/test_save_options.rb +10 -2
- data/test/xml/test_builder.rb +17 -0
- data/test/xml/test_document.rb +22 -0
- data/test/xml/test_node.rb +19 -1
- data/test/xml/test_node_reparenting.rb +16 -3
- data/test/xml/test_node_set.rb +34 -0
- data/test/xml/test_schema.rb +5 -0
- data/test/xslt/test_exception_handling.rb +37 -0
- metadata +185 -157
- data/deps.rip +0 -5
- data/ext/java/nokogiri/EncodingHandler.java +0 -124
- data/ext/java/nokogiri/HtmlDocument.java +0 -149
- data/ext/java/nokogiri/HtmlElementDescription.java +0 -145
- data/ext/java/nokogiri/HtmlEntityLookup.java +0 -79
- data/ext/java/nokogiri/HtmlSaxParserContext.java +0 -259
- data/ext/java/nokogiri/NokogiriService.java +0 -535
- data/ext/java/nokogiri/XmlAttr.java +0 -191
- data/ext/java/nokogiri/XmlAttributeDecl.java +0 -130
- data/ext/java/nokogiri/XmlCdata.java +0 -91
- data/ext/java/nokogiri/XmlComment.java +0 -86
- data/ext/java/nokogiri/XmlDocument.java +0 -529
- data/ext/java/nokogiri/XmlDocumentFragment.java +0 -217
- data/ext/java/nokogiri/XmlDtd.java +0 -467
- data/ext/java/nokogiri/XmlElement.java +0 -222
- data/ext/java/nokogiri/XmlElementContent.java +0 -382
- data/ext/java/nokogiri/XmlElementDecl.java +0 -148
- data/ext/java/nokogiri/XmlEntityDecl.java +0 -162
- data/ext/java/nokogiri/XmlEntityReference.java +0 -75
- data/ext/java/nokogiri/XmlNamespace.java +0 -128
- data/ext/java/nokogiri/XmlNode.java +0 -1399
- data/ext/java/nokogiri/XmlNodeSet.java +0 -311
- data/ext/java/nokogiri/XmlProcessingInstruction.java +0 -103
- data/ext/java/nokogiri/XmlReader.java +0 -411
- data/ext/java/nokogiri/XmlRelaxng.java +0 -144
- data/ext/java/nokogiri/XmlSaxParserContext.java +0 -367
- data/ext/java/nokogiri/XmlSaxPushParser.java +0 -184
- data/ext/java/nokogiri/XmlSchema.java +0 -319
- data/ext/java/nokogiri/XmlSyntaxError.java +0 -119
- data/ext/java/nokogiri/XmlText.java +0 -136
- data/ext/java/nokogiri/XmlXpathContext.java +0 -179
- data/ext/java/nokogiri/XsltStylesheet.java +0 -183
- data/ext/java/nokogiri/internals/HtmlDomParserContext.java +0 -206
- data/ext/java/nokogiri/internals/NokogiriDocumentCache.java +0 -73
- data/ext/java/nokogiri/internals/NokogiriErrorHandler.java +0 -86
- data/ext/java/nokogiri/internals/NokogiriHandler.java +0 -327
- data/ext/java/nokogiri/internals/NokogiriHelpers.java +0 -582
- data/ext/java/nokogiri/internals/NokogiriNamespaceCache.java +0 -171
- data/ext/java/nokogiri/internals/NokogiriNamespaceContext.java +0 -118
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler.java +0 -74
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler4NekoHtml.java +0 -121
- data/ext/java/nokogiri/internals/NokogiriStrictErrorHandler.java +0 -79
- data/ext/java/nokogiri/internals/NokogiriXPathFunction.java +0 -126
- data/ext/java/nokogiri/internals/NokogiriXPathFunctionResolver.java +0 -56
- data/ext/java/nokogiri/internals/ParserContext.java +0 -278
- data/ext/java/nokogiri/internals/PushInputStream.java +0 -411
- data/ext/java/nokogiri/internals/ReaderNode.java +0 -474
- data/ext/java/nokogiri/internals/SaveContext.java +0 -288
- data/ext/java/nokogiri/internals/SchemaErrorHandler.java +0 -76
- data/ext/java/nokogiri/internals/XmlDeclHandler.java +0 -42
- data/ext/java/nokogiri/internals/XmlDomParser.java +0 -77
- data/ext/java/nokogiri/internals/XmlDomParserContext.java +0 -238
- data/ext/java/nokogiri/internals/XmlSaxParser.java +0 -65
- data/ext/java/nokogiri/internals/XsltExtensionFunction.java +0 -72
- 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/xercesImpl.jar +0 -0
@@ -1,144 +0,0 @@
|
|
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;
|
34
|
-
|
35
|
-
import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
|
36
|
-
|
37
|
-
import java.io.ByteArrayInputStream;
|
38
|
-
import java.io.IOException;
|
39
|
-
import java.io.InputStream;
|
40
|
-
import java.io.StringWriter;
|
41
|
-
import java.io.UnsupportedEncodingException;
|
42
|
-
|
43
|
-
import javax.xml.transform.Source;
|
44
|
-
import javax.xml.transform.TransformerConfigurationException;
|
45
|
-
import javax.xml.transform.TransformerException;
|
46
|
-
import javax.xml.transform.TransformerFactory;
|
47
|
-
import javax.xml.transform.dom.DOMSource;
|
48
|
-
import javax.xml.transform.stream.StreamResult;
|
49
|
-
import javax.xml.transform.stream.StreamSource;
|
50
|
-
|
51
|
-
import org.iso_relax.verifier.Schema;
|
52
|
-
import org.iso_relax.verifier.Verifier;
|
53
|
-
import org.iso_relax.verifier.VerifierConfigurationException;
|
54
|
-
import org.iso_relax.verifier.VerifierFactory;
|
55
|
-
import org.jruby.Ruby;
|
56
|
-
import org.jruby.RubyClass;
|
57
|
-
import org.jruby.anno.JRubyClass;
|
58
|
-
import org.jruby.runtime.ThreadContext;
|
59
|
-
import org.w3c.dom.Document;
|
60
|
-
import org.xml.sax.ErrorHandler;
|
61
|
-
import org.xml.sax.SAXException;
|
62
|
-
|
63
|
-
/**
|
64
|
-
* Class for Nokogiri::XML::RelaxNG
|
65
|
-
*
|
66
|
-
* @author sergio
|
67
|
-
* @author Yoko Harada <yokolet@gmail.com>
|
68
|
-
*/
|
69
|
-
@JRubyClass(name="Nokogiri::XML::RelaxNG", parent="Nokogiri::XML::Schema")
|
70
|
-
public class XmlRelaxng extends XmlSchema {
|
71
|
-
private Verifier verifier;
|
72
|
-
|
73
|
-
public XmlRelaxng(Ruby ruby, RubyClass klazz) {
|
74
|
-
super(ruby, klazz);
|
75
|
-
}
|
76
|
-
|
77
|
-
private void setVerifier(Verifier verifier) {
|
78
|
-
this.verifier = verifier;
|
79
|
-
}
|
80
|
-
|
81
|
-
static XmlSchema createSchemaInstance(ThreadContext context, RubyClass klazz, Source source) {
|
82
|
-
Ruby runtime = context.getRuntime();
|
83
|
-
XmlRelaxng xmlRelaxng = (XmlRelaxng) NokogiriService.XML_RELAXNG_ALLOCATOR.allocate(runtime, klazz);
|
84
|
-
xmlRelaxng.setInstanceVariable("@errors", runtime.newEmptyArray());
|
85
|
-
|
86
|
-
try {
|
87
|
-
Schema schema = xmlRelaxng.getSchema(source, context);
|
88
|
-
xmlRelaxng.setVerifier(schema.newVerifier());
|
89
|
-
return xmlRelaxng;
|
90
|
-
} catch (VerifierConfigurationException ex) {
|
91
|
-
throw context.getRuntime().newRuntimeError("Could not parse document: " + ex.getMessage());
|
92
|
-
}
|
93
|
-
}
|
94
|
-
|
95
|
-
private Schema getSchema(Source source, ThreadContext context) {
|
96
|
-
InputStream is = null;
|
97
|
-
VerifierFactory factory = new com.thaiopensource.relaxng.jarv.VerifierFactoryImpl();
|
98
|
-
if (source instanceof StreamSource) {
|
99
|
-
StreamSource ss = (StreamSource)source;
|
100
|
-
is = ss.getInputStream();
|
101
|
-
} else { //if (this.source instanceof DOMSource)
|
102
|
-
DOMSource ds = (DOMSource)source;
|
103
|
-
StringWriter xmlAsWriter = new StringWriter();
|
104
|
-
StreamResult result = new StreamResult(xmlAsWriter);
|
105
|
-
try {
|
106
|
-
TransformerFactory.newInstance().newTransformer().transform(ds, result);
|
107
|
-
} catch (TransformerConfigurationException ex) {
|
108
|
-
throw context.getRuntime()
|
109
|
-
.newRuntimeError("Could not parse document: "+ex.getMessage());
|
110
|
-
} catch (TransformerException ex) {
|
111
|
-
throw context.getRuntime()
|
112
|
-
.newRuntimeError("Could not parse document: "+ex.getMessage());
|
113
|
-
}
|
114
|
-
try {
|
115
|
-
is = new ByteArrayInputStream(xmlAsWriter.toString().getBytes("UTF-8"));
|
116
|
-
} catch (UnsupportedEncodingException ex) {
|
117
|
-
throw context.getRuntime()
|
118
|
-
.newRuntimeError("Could not parse document: "+ex.getMessage());
|
119
|
-
}
|
120
|
-
}
|
121
|
-
|
122
|
-
try {
|
123
|
-
return factory.compileSchema(is);
|
124
|
-
} catch (VerifierConfigurationException ex) {
|
125
|
-
throw context.getRuntime()
|
126
|
-
.newRuntimeError("Could not parse document: "+ex.getMessage());
|
127
|
-
} catch (SAXException ex) {
|
128
|
-
throw context.getRuntime()
|
129
|
-
.newRuntimeError("Could not parse document: "+ex.getMessage());
|
130
|
-
} catch (IOException ex) {
|
131
|
-
throw context.getRuntime().newIOError(ex.getMessage());
|
132
|
-
}
|
133
|
-
}
|
134
|
-
|
135
|
-
@Override
|
136
|
-
protected void setErrorHandler(ErrorHandler errorHandler) {
|
137
|
-
verifier.setErrorHandler(errorHandler);
|
138
|
-
}
|
139
|
-
|
140
|
-
@Override
|
141
|
-
protected void validate(Document document) throws SAXException, IOException {
|
142
|
-
verifier.verify(document);
|
143
|
-
}
|
144
|
-
}
|
@@ -1,367 +0,0 @@
|
|
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;
|
34
|
-
|
35
|
-
import static nokogiri.internals.NokogiriHelpers.rubyStringToString;
|
36
|
-
import static org.jruby.javasupport.util.RuntimeHelpers.invoke;
|
37
|
-
|
38
|
-
import java.io.IOException;
|
39
|
-
import java.io.InputStream;
|
40
|
-
|
41
|
-
import nokogiri.internals.NokogiriHandler;
|
42
|
-
import nokogiri.internals.ParserContext;
|
43
|
-
import nokogiri.internals.XmlSaxParser;
|
44
|
-
|
45
|
-
import org.apache.xerces.parsers.AbstractSAXParser;
|
46
|
-
import org.jruby.Ruby;
|
47
|
-
import org.jruby.RubyClass;
|
48
|
-
import org.jruby.RubyFixnum;
|
49
|
-
import org.jruby.RubyModule;
|
50
|
-
import org.jruby.RubyObjectAdapter;
|
51
|
-
import org.jruby.anno.JRubyClass;
|
52
|
-
import org.jruby.anno.JRubyMethod;
|
53
|
-
import org.jruby.exceptions.RaiseException;
|
54
|
-
import org.jruby.javasupport.JavaEmbedUtils;
|
55
|
-
import org.jruby.runtime.ThreadContext;
|
56
|
-
import org.jruby.runtime.builtin.IRubyObject;
|
57
|
-
import org.xml.sax.ContentHandler;
|
58
|
-
import org.xml.sax.ErrorHandler;
|
59
|
-
import org.xml.sax.SAXException;
|
60
|
-
import org.xml.sax.SAXNotRecognizedException;
|
61
|
-
import org.xml.sax.SAXNotSupportedException;
|
62
|
-
import org.xml.sax.SAXParseException;
|
63
|
-
|
64
|
-
/**
|
65
|
-
* Base class for the SAX parsers.
|
66
|
-
*
|
67
|
-
* @author Patrick Mahoney <pat@polycrystal.org>
|
68
|
-
* @author Yoko Harada <yokolet@gmail.com>
|
69
|
-
*/
|
70
|
-
@JRubyClass(name="Nokogiri::XML::SAX::ParserContext")
|
71
|
-
public class XmlSaxParserContext extends ParserContext {
|
72
|
-
protected static final String FEATURE_NAMESPACES =
|
73
|
-
"http://xml.org/sax/features/namespaces";
|
74
|
-
protected static final String FEATURE_NAMESPACE_PREFIXES =
|
75
|
-
"http://xml.org/sax/features/namespace-prefixes";
|
76
|
-
protected static final String FEATURE_LOAD_EXTERNAL_DTD =
|
77
|
-
"http://apache.org/xml/features/nonvalidating/load-external-dtd";
|
78
|
-
|
79
|
-
protected AbstractSAXParser parser;
|
80
|
-
|
81
|
-
protected NokogiriHandler handler = null;
|
82
|
-
private IRubyObject replaceEntities;
|
83
|
-
|
84
|
-
public XmlSaxParserContext(final Ruby ruby, RubyClass rubyClass) {
|
85
|
-
super(ruby, rubyClass);
|
86
|
-
}
|
87
|
-
|
88
|
-
protected void initialize(Ruby runtime) {
|
89
|
-
replaceEntities = runtime.getTrue();
|
90
|
-
try {
|
91
|
-
parser = createParser();
|
92
|
-
} catch (SAXException se) {
|
93
|
-
throw RaiseException.createNativeRaiseException(runtime, se);
|
94
|
-
}
|
95
|
-
}
|
96
|
-
|
97
|
-
/**
|
98
|
-
* Create and return a copy of this object.
|
99
|
-
*
|
100
|
-
* @return a clone of this object
|
101
|
-
*/
|
102
|
-
@Override
|
103
|
-
public Object clone() throws CloneNotSupportedException {
|
104
|
-
return super.clone();
|
105
|
-
}
|
106
|
-
|
107
|
-
protected AbstractSAXParser createParser() throws SAXException {
|
108
|
-
XmlSaxParser parser = new XmlSaxParser();
|
109
|
-
parser.setFeature(FEATURE_NAMESPACE_PREFIXES, true);
|
110
|
-
parser.setFeature(FEATURE_LOAD_EXTERNAL_DTD, false);
|
111
|
-
return parser;
|
112
|
-
}
|
113
|
-
|
114
|
-
/**
|
115
|
-
* Create a new parser context that will parse the string
|
116
|
-
* <code>data</code>.
|
117
|
-
*/
|
118
|
-
@JRubyMethod(name="memory", meta=true)
|
119
|
-
public static IRubyObject parse_memory(ThreadContext context,
|
120
|
-
IRubyObject klazz,
|
121
|
-
IRubyObject data) {
|
122
|
-
XmlSaxParserContext ctx = (XmlSaxParserContext) NokogiriService.XML_SAXPARSER_CONTEXT_ALLOCATOR.allocate(context.getRuntime(), (RubyClass) klazz);
|
123
|
-
ctx.initialize(context.getRuntime());
|
124
|
-
ctx.setInputSource(context, data);
|
125
|
-
return ctx;
|
126
|
-
}
|
127
|
-
|
128
|
-
/**
|
129
|
-
* Create a new parser context that will read from the file
|
130
|
-
* <code>data</code> and parse.
|
131
|
-
*/
|
132
|
-
@JRubyMethod(name="file", meta=true)
|
133
|
-
public static IRubyObject parse_file(ThreadContext context,
|
134
|
-
IRubyObject klazz,
|
135
|
-
IRubyObject data) {
|
136
|
-
XmlSaxParserContext ctx = (XmlSaxParserContext) NokogiriService.XML_SAXPARSER_CONTEXT_ALLOCATOR.allocate(context.getRuntime(), (RubyClass) klazz);
|
137
|
-
ctx.initialize(context.getRuntime());
|
138
|
-
ctx.setInputSourceFile(context, data);
|
139
|
-
return ctx;
|
140
|
-
}
|
141
|
-
|
142
|
-
/**
|
143
|
-
* Create a new parser context that will read from the IO or
|
144
|
-
* StringIO <code>data</code> and parse.
|
145
|
-
*
|
146
|
-
* TODO: Currently ignores encoding <code>enc</code>.
|
147
|
-
*/
|
148
|
-
@JRubyMethod(name="io", meta=true)
|
149
|
-
public static IRubyObject parse_io(ThreadContext context,
|
150
|
-
IRubyObject klazz,
|
151
|
-
IRubyObject data,
|
152
|
-
IRubyObject enc) {
|
153
|
-
//int encoding = (int)enc.convertToInteger().getLongValue();
|
154
|
-
XmlSaxParserContext ctx = (XmlSaxParserContext) NokogiriService.XML_SAXPARSER_CONTEXT_ALLOCATOR.allocate(context.getRuntime(), (RubyClass) klazz);
|
155
|
-
ctx.initialize(context.getRuntime());
|
156
|
-
ctx.setInputSource(context, data);
|
157
|
-
return ctx;
|
158
|
-
}
|
159
|
-
|
160
|
-
/**
|
161
|
-
* Create a new parser context that will read from a raw input
|
162
|
-
* stream. Not a JRuby method. Meant to be run in a separate
|
163
|
-
* thread by XmlSaxPushParser.
|
164
|
-
*/
|
165
|
-
public static IRubyObject parse_stream(ThreadContext context,
|
166
|
-
IRubyObject klazz,
|
167
|
-
InputStream stream) {
|
168
|
-
XmlSaxParserContext ctx = (XmlSaxParserContext) NokogiriService.XML_SAXPARSER_CONTEXT_ALLOCATOR.allocate(context.getRuntime(), (RubyClass) klazz);
|
169
|
-
ctx.initialize(context.getRuntime());
|
170
|
-
ctx.setInputSource(stream);
|
171
|
-
return ctx;
|
172
|
-
}
|
173
|
-
|
174
|
-
/**
|
175
|
-
* Set a property of the underlying parser.
|
176
|
-
*/
|
177
|
-
protected void setProperty(String key, Object val)
|
178
|
-
throws SAXNotRecognizedException, SAXNotSupportedException {
|
179
|
-
parser.setProperty(key, val);
|
180
|
-
}
|
181
|
-
|
182
|
-
protected void setContentHandler(ContentHandler handler) {
|
183
|
-
parser.setContentHandler(handler);
|
184
|
-
}
|
185
|
-
|
186
|
-
protected void setErrorHandler(ErrorHandler handler) {
|
187
|
-
parser.setErrorHandler(handler);
|
188
|
-
}
|
189
|
-
|
190
|
-
public NokogiriHandler getNokogiriHandler() {
|
191
|
-
return handler;
|
192
|
-
}
|
193
|
-
|
194
|
-
/**
|
195
|
-
* Perform any initialization prior to parsing with the handler
|
196
|
-
* <code>handlerRuby</code>. Convenience hook for subclasses.
|
197
|
-
*/
|
198
|
-
protected void preParse(ThreadContext context,
|
199
|
-
IRubyObject handlerRuby,
|
200
|
-
NokogiriHandler handler) {
|
201
|
-
((XmlSaxParser) parser).setXmlDeclHandler(handler);
|
202
|
-
}
|
203
|
-
|
204
|
-
protected void postParse(ThreadContext context,
|
205
|
-
IRubyObject handlerRuby,
|
206
|
-
NokogiriHandler handler) {
|
207
|
-
// noop
|
208
|
-
}
|
209
|
-
|
210
|
-
protected void do_parse() throws SAXException, IOException {
|
211
|
-
parser.parse(getInputSource());
|
212
|
-
}
|
213
|
-
|
214
|
-
@JRubyMethod
|
215
|
-
public IRubyObject parse_with(ThreadContext context,
|
216
|
-
IRubyObject handlerRuby) {
|
217
|
-
Ruby ruby = context.getRuntime();
|
218
|
-
|
219
|
-
if(!invoke(context, handlerRuby, "respond_to?",
|
220
|
-
ruby.newSymbol("document")).isTrue()) {
|
221
|
-
String msg = "argument must respond_to document";
|
222
|
-
throw ruby.newArgumentError(msg);
|
223
|
-
}
|
224
|
-
|
225
|
-
handler = new NokogiriHandler(ruby, handlerRuby);
|
226
|
-
preParse(context, handlerRuby, handler);
|
227
|
-
|
228
|
-
setContentHandler(handler);
|
229
|
-
setErrorHandler(handler);
|
230
|
-
|
231
|
-
try{
|
232
|
-
setProperty("http://xml.org/sax/properties/lexical-handler",
|
233
|
-
handler);
|
234
|
-
} catch(Exception ex) {
|
235
|
-
throw ruby.newRuntimeError(
|
236
|
-
"Problem while creating XML SAX Parser: " + ex.toString());
|
237
|
-
}
|
238
|
-
|
239
|
-
try{
|
240
|
-
try {
|
241
|
-
do_parse();
|
242
|
-
} catch(SAXParseException spe) {
|
243
|
-
// A bad document (<foo><bar></foo>) should call the
|
244
|
-
// error handler instead of raising a SAX exception.
|
245
|
-
|
246
|
-
// However, an EMPTY document should raise a
|
247
|
-
// RuntimeError. This is a bit kludgy, but AFAIK SAX
|
248
|
-
// doesn't distinguish between empty and bad whereas
|
249
|
-
// Nokogiri does.
|
250
|
-
String message = spe.getMessage();
|
251
|
-
if ("Premature end of file.".matches(message)) {
|
252
|
-
throw ruby.newRuntimeError(
|
253
|
-
"couldn't parse document: " + message);
|
254
|
-
} else {
|
255
|
-
handler.error(spe);
|
256
|
-
}
|
257
|
-
|
258
|
-
}
|
259
|
-
} catch(SAXException se) {
|
260
|
-
throw RaiseException.createNativeRaiseException(ruby, se);
|
261
|
-
} catch(IOException ioe) {
|
262
|
-
throw ruby.newIOErrorFromException(ioe);
|
263
|
-
}
|
264
|
-
|
265
|
-
postParse(context, handlerRuby, handler);
|
266
|
-
|
267
|
-
//maybeTrimLeadingAndTrailingWhitespace(context, handlerRuby);
|
268
|
-
|
269
|
-
return ruby.getNil();
|
270
|
-
}
|
271
|
-
|
272
|
-
/**
|
273
|
-
* Can take a boolean assignment.
|
274
|
-
*
|
275
|
-
* @param context
|
276
|
-
* @param value
|
277
|
-
* @return
|
278
|
-
*/
|
279
|
-
@JRubyMethod(name = "replace_entities=")
|
280
|
-
public IRubyObject set_replace_entities(ThreadContext context,
|
281
|
-
IRubyObject value) {
|
282
|
-
if (!value.isTrue()) replaceEntities = context.getRuntime().getFalse();
|
283
|
-
else replaceEntities = context.getRuntime().getTrue();
|
284
|
-
|
285
|
-
return this;
|
286
|
-
}
|
287
|
-
|
288
|
-
@JRubyMethod(name="replace_entities")
|
289
|
-
public IRubyObject get_replace_entities(ThreadContext context) {
|
290
|
-
return replaceEntities;
|
291
|
-
}
|
292
|
-
|
293
|
-
|
294
|
-
/**
|
295
|
-
* If the handler's document is a FragmentHandler, attempt to trim
|
296
|
-
* leading and trailing whitespace.
|
297
|
-
*
|
298
|
-
* This is a bit hackish and depends heavily on the internals of
|
299
|
-
* FragmentHandler.
|
300
|
-
*/
|
301
|
-
protected void maybeTrimLeadingAndTrailingWhitespace(ThreadContext context,
|
302
|
-
IRubyObject parser) {
|
303
|
-
final String path = "Nokogiri::XML::FragmentHandler";
|
304
|
-
RubyObjectAdapter adapter = JavaEmbedUtils.newObjectAdapter();
|
305
|
-
RubyModule mod =
|
306
|
-
context.getRuntime().getClassFromPath(path);
|
307
|
-
|
308
|
-
IRubyObject handler = adapter.getInstanceVariable(parser, "@document");
|
309
|
-
if (handler == null || handler.isNil() || !adapter.isKindOf(handler, mod))
|
310
|
-
return;
|
311
|
-
IRubyObject stack = adapter.getInstanceVariable(handler, "@stack");
|
312
|
-
if (stack == null || stack.isNil())
|
313
|
-
return;
|
314
|
-
// doc is finally a DocumentFragment whose nodes we can check
|
315
|
-
IRubyObject doc = adapter.callMethod(stack, "first");
|
316
|
-
if (doc == null || doc.isNil())
|
317
|
-
return;
|
318
|
-
|
319
|
-
IRubyObject children;
|
320
|
-
|
321
|
-
for (;;) {
|
322
|
-
children = adapter.callMethod(doc, "children");
|
323
|
-
IRubyObject first = adapter.callMethod(children, "first");
|
324
|
-
if (isWhitespaceText(context, first))
|
325
|
-
adapter.callMethod(first, "unlink");
|
326
|
-
else
|
327
|
-
break;
|
328
|
-
}
|
329
|
-
|
330
|
-
for (;;) {
|
331
|
-
children = adapter.callMethod(doc, "children");
|
332
|
-
IRubyObject last = adapter.callMethod(children, "last");
|
333
|
-
if (isWhitespaceText(context, last))
|
334
|
-
adapter.callMethod(last, "unlink");
|
335
|
-
else
|
336
|
-
break;
|
337
|
-
}
|
338
|
-
|
339
|
-
// While we have a document, normalize it.
|
340
|
-
((XmlNode) doc).normalize();
|
341
|
-
}
|
342
|
-
|
343
|
-
protected boolean isWhitespaceText(ThreadContext context, IRubyObject obj) {
|
344
|
-
if (obj == null || obj.isNil()) return false;
|
345
|
-
|
346
|
-
XmlNode node = (XmlNode) obj;
|
347
|
-
if (!(node instanceof XmlText))
|
348
|
-
return false;
|
349
|
-
|
350
|
-
String content = rubyStringToString(node.content(context));
|
351
|
-
return content.trim().length() == 0;
|
352
|
-
}
|
353
|
-
|
354
|
-
@JRubyMethod(name="column")
|
355
|
-
public IRubyObject column(ThreadContext context) {
|
356
|
-
Integer number = handler.getColumn();
|
357
|
-
if (number == null) return context.getRuntime().getNil();
|
358
|
-
else return RubyFixnum.newFixnum(context.getRuntime(), number.longValue());
|
359
|
-
}
|
360
|
-
|
361
|
-
@JRubyMethod(name="line")
|
362
|
-
public IRubyObject line(ThreadContext context) {
|
363
|
-
Integer number = handler.getLine();
|
364
|
-
if (number == null) return context.getRuntime().getNil();
|
365
|
-
else return RubyFixnum.newFixnum(context.getRuntime(), number.longValue());
|
366
|
-
}
|
367
|
-
}
|