nokogiri 1.12.4-java → 1.13.2-java
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.
- checksums.yaml +4 -4
- data/Gemfile +2 -0
- data/README.md +9 -7
- data/bin/nokogiri +63 -50
- data/dependencies.yml +11 -62
- data/ext/java/nokogiri/Html4SaxPushParser.java +14 -19
- data/ext/java/nokogiri/XmlDocumentFragment.java +4 -3
- data/ext/java/nokogiri/XmlNamespace.java +0 -8
- data/ext/java/nokogiri/XmlNode.java +23 -2
- data/ext/java/nokogiri/XmlNodeSet.java +0 -4
- data/ext/java/nokogiri/XmlReader.java +12 -0
- data/ext/java/nokogiri/XmlSaxParserContext.java +31 -93
- data/ext/java/nokogiri/XmlSaxPushParser.java +17 -25
- data/ext/java/nokogiri/internals/NokogiriEntityResolver.java +1 -1
- data/ext/java/nokogiri/internals/NokogiriErrorHandler.java +29 -8
- data/ext/java/nokogiri/internals/NokogiriHandler.java +6 -24
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler.java +9 -8
- data/ext/java/nokogiri/internals/NokogiriNonStrictErrorHandler4NekoHtml.java +10 -9
- data/ext/java/nokogiri/internals/NokogiriStrictErrorHandler.java +7 -6
- data/ext/java/nokogiri/internals/SaveContextVisitor.java +2 -0
- data/ext/java/nokogiri/internals/XmlDomParserContext.java +16 -25
- data/ext/nokogiri/extconf.rb +64 -37
- data/ext/nokogiri/xml_document.c +35 -35
- data/ext/nokogiri/xml_document_fragment.c +0 -2
- data/ext/nokogiri/xml_dtd.c +2 -2
- data/ext/nokogiri/xml_encoding_handler.c +25 -11
- data/ext/nokogiri/xml_node.c +638 -333
- data/ext/nokogiri/xml_reader.c +37 -11
- data/ext/nokogiri/xml_xpath_context.c +72 -49
- data/ext/nokogiri/xslt_stylesheet.c +107 -9
- data/lib/nokogiri/class_resolver.rb +67 -0
- data/lib/nokogiri/css/node.rb +9 -8
- data/lib/nokogiri/css/parser.rb +360 -341
- data/lib/nokogiri/css/parser.y +249 -244
- data/lib/nokogiri/css/parser_extras.rb +20 -20
- data/lib/nokogiri/css/syntax_error.rb +1 -0
- data/lib/nokogiri/css/tokenizer.rb +4 -3
- data/lib/nokogiri/css/tokenizer.rex +3 -2
- data/lib/nokogiri/css/xpath_visitor.rb +179 -82
- data/lib/nokogiri/css.rb +38 -6
- data/lib/nokogiri/decorators/slop.rb +8 -7
- data/lib/nokogiri/extension.rb +1 -1
- data/lib/nokogiri/gumbo.rb +1 -0
- data/lib/nokogiri/html.rb +16 -10
- data/lib/nokogiri/html4/builder.rb +1 -0
- data/lib/nokogiri/html4/document.rb +84 -75
- data/lib/nokogiri/html4/document_fragment.rb +11 -7
- data/lib/nokogiri/html4/element_description.rb +1 -0
- data/lib/nokogiri/html4/element_description_defaults.rb +426 -520
- data/lib/nokogiri/html4/entity_lookup.rb +2 -1
- data/lib/nokogiri/html4/sax/parser.rb +2 -1
- data/lib/nokogiri/html4/sax/parser_context.rb +1 -0
- data/lib/nokogiri/html4/sax/push_parser.rb +7 -7
- data/lib/nokogiri/html4.rb +11 -5
- data/lib/nokogiri/html5/document.rb +24 -10
- data/lib/nokogiri/html5/document_fragment.rb +5 -2
- data/lib/nokogiri/html5/node.rb +6 -3
- data/lib/nokogiri/html5.rb +68 -64
- data/lib/nokogiri/jruby/dependencies.rb +10 -9
- data/lib/nokogiri/nokogiri.jar +0 -0
- data/lib/nokogiri/syntax_error.rb +1 -0
- data/lib/nokogiri/version/constant.rb +2 -1
- data/lib/nokogiri/version/info.rb +19 -13
- data/lib/nokogiri/version.rb +1 -0
- data/lib/nokogiri/xml/attr.rb +5 -3
- data/lib/nokogiri/xml/attribute_decl.rb +2 -1
- data/lib/nokogiri/xml/builder.rb +32 -32
- data/lib/nokogiri/xml/cdata.rb +2 -1
- data/lib/nokogiri/xml/character_data.rb +1 -0
- data/lib/nokogiri/xml/document.rb +139 -103
- data/lib/nokogiri/xml/document_fragment.rb +41 -38
- data/lib/nokogiri/xml/dtd.rb +3 -2
- data/lib/nokogiri/xml/element_content.rb +1 -0
- data/lib/nokogiri/xml/element_decl.rb +2 -1
- data/lib/nokogiri/xml/entity_decl.rb +3 -2
- data/lib/nokogiri/xml/entity_reference.rb +1 -0
- data/lib/nokogiri/xml/namespace.rb +2 -0
- data/lib/nokogiri/xml/node/save_options.rb +7 -4
- data/lib/nokogiri/xml/node.rb +512 -348
- data/lib/nokogiri/xml/node_set.rb +46 -54
- data/lib/nokogiri/xml/notation.rb +12 -0
- data/lib/nokogiri/xml/parse_options.rb +11 -7
- data/lib/nokogiri/xml/pp/character_data.rb +8 -6
- data/lib/nokogiri/xml/pp/node.rb +24 -26
- data/lib/nokogiri/xml/pp.rb +1 -0
- data/lib/nokogiri/xml/processing_instruction.rb +2 -1
- data/lib/nokogiri/xml/reader.rb +17 -19
- data/lib/nokogiri/xml/relax_ng.rb +1 -0
- data/lib/nokogiri/xml/sax/document.rb +20 -19
- data/lib/nokogiri/xml/sax/parser.rb +36 -34
- data/lib/nokogiri/xml/sax/parser_context.rb +7 -3
- data/lib/nokogiri/xml/sax/push_parser.rb +5 -5
- data/lib/nokogiri/xml/sax.rb +1 -0
- data/lib/nokogiri/xml/schema.rb +7 -6
- data/lib/nokogiri/xml/searchable.rb +93 -62
- data/lib/nokogiri/xml/syntax_error.rb +4 -4
- data/lib/nokogiri/xml/text.rb +1 -0
- data/lib/nokogiri/xml/xpath/syntax_error.rb +2 -1
- data/lib/nokogiri/xml/xpath.rb +12 -0
- data/lib/nokogiri/xml/xpath_context.rb +2 -3
- data/lib/nokogiri/xml.rb +3 -3
- data/lib/nokogiri/xslt/stylesheet.rb +1 -0
- data/lib/nokogiri/xslt.rb +21 -13
- data/lib/nokogiri.rb +19 -16
- data/lib/xsd/xmlparser/nokogiri.rb +25 -24
- metadata +102 -30
@@ -1,33 +1,23 @@
|
|
1
1
|
package nokogiri;
|
2
2
|
|
3
|
-
import
|
4
|
-
|
5
|
-
import java.io.IOException;
|
6
|
-
import java.io.InputStream;
|
7
|
-
|
3
|
+
import nokogiri.internals.*;
|
8
4
|
import org.apache.xerces.parsers.AbstractSAXParser;
|
9
5
|
import org.jruby.Ruby;
|
10
6
|
import org.jruby.RubyClass;
|
11
7
|
import org.jruby.RubyFixnum;
|
12
|
-
import org.jruby.RubyModule;
|
13
|
-
import org.jruby.RubyObjectAdapter;
|
14
8
|
import org.jruby.anno.JRubyClass;
|
15
9
|
import org.jruby.anno.JRubyMethod;
|
16
10
|
import org.jruby.exceptions.RaiseException;
|
17
|
-
import org.jruby.
|
11
|
+
import org.jruby.runtime.Helpers;
|
18
12
|
import org.jruby.runtime.ThreadContext;
|
19
13
|
import org.jruby.runtime.builtin.IRubyObject;
|
20
|
-
import org.xml.sax.ContentHandler;
|
21
|
-
import org.xml.sax.ErrorHandler;
|
22
14
|
import org.xml.sax.SAXException;
|
23
|
-
import org.xml.sax.SAXNotRecognizedException;
|
24
|
-
import org.xml.sax.SAXNotSupportedException;
|
25
15
|
import org.xml.sax.SAXParseException;
|
26
16
|
|
27
|
-
import
|
28
|
-
import
|
29
|
-
|
30
|
-
import
|
17
|
+
import java.io.IOException;
|
18
|
+
import java.io.InputStream;
|
19
|
+
|
20
|
+
import static org.jruby.runtime.Helpers.invoke;
|
31
21
|
|
32
22
|
/**
|
33
23
|
* Base class for the SAX parsers.
|
@@ -51,6 +41,7 @@ public class XmlSaxParserContext extends ParserContext
|
|
51
41
|
protected AbstractSAXParser parser;
|
52
42
|
|
53
43
|
protected NokogiriHandler handler;
|
44
|
+
protected NokogiriErrorHandler errorHandler;
|
54
45
|
private boolean replaceEntities = true;
|
55
46
|
private boolean recovery = false;
|
56
47
|
|
@@ -168,31 +159,12 @@ public class XmlSaxParserContext extends ParserContext
|
|
168
159
|
return (XmlSaxParserContext) NokogiriService.XML_SAXPARSER_CONTEXT_ALLOCATOR.allocate(runtime, klazz);
|
169
160
|
}
|
170
161
|
|
171
|
-
/**
|
172
|
-
* Set a property of the underlying parser.
|
173
|
-
*/
|
174
|
-
protected void
|
175
|
-
setProperty(String key, Object val)
|
176
|
-
throws SAXNotRecognizedException, SAXNotSupportedException
|
177
|
-
{
|
178
|
-
parser.setProperty(key, val);
|
179
|
-
}
|
180
|
-
|
181
|
-
protected void
|
182
|
-
setContentHandler(ContentHandler handler)
|
183
|
-
{
|
184
|
-
parser.setContentHandler(handler);
|
185
|
-
}
|
186
|
-
|
187
|
-
protected void
|
188
|
-
setErrorHandler(ErrorHandler handler)
|
189
|
-
{
|
190
|
-
parser.setErrorHandler(handler);
|
191
|
-
}
|
192
|
-
|
193
162
|
public final NokogiriHandler
|
194
163
|
getNokogiriHandler() { return handler; }
|
195
164
|
|
165
|
+
public final NokogiriErrorHandler
|
166
|
+
getNokogiriErrorHandler() { return errorHandler; }
|
167
|
+
|
196
168
|
/**
|
197
169
|
* Perform any initialization prior to parsing with the handler
|
198
170
|
* <code>handlerRuby</code>. Convenience hook for subclasses.
|
@@ -223,6 +195,17 @@ public class XmlSaxParserContext extends ParserContext
|
|
223
195
|
parser.parse(getInputSource());
|
224
196
|
}
|
225
197
|
|
198
|
+
protected static Options
|
199
|
+
defaultParseOptions(ThreadContext context)
|
200
|
+
{
|
201
|
+
return new ParserContext.Options(
|
202
|
+
RubyFixnum.fix2long(Helpers.invoke(context,
|
203
|
+
((RubyClass)context.getRuntime().getClassFromPath("Nokogiri::XML::ParseOptions"))
|
204
|
+
.getConstant("DEFAULT_XML"),
|
205
|
+
"to_i"))
|
206
|
+
);
|
207
|
+
}
|
208
|
+
|
226
209
|
@JRubyMethod
|
227
210
|
public IRubyObject
|
228
211
|
parse_with(ThreadContext context, IRubyObject handlerRuby)
|
@@ -233,14 +216,19 @@ public class XmlSaxParserContext extends ParserContext
|
|
233
216
|
throw runtime.newArgumentError("argument must respond_to document");
|
234
217
|
}
|
235
218
|
|
236
|
-
|
237
|
-
|
219
|
+
/* TODO: how should we pass in parse options? */
|
220
|
+
ParserContext.Options options = defaultParseOptions(context);
|
221
|
+
|
222
|
+
errorHandler = new NokogiriStrictErrorHandler(runtime, options.noError, options.noWarning);
|
223
|
+
handler = new NokogiriHandler(runtime, handlerRuby, errorHandler);
|
238
224
|
|
239
|
-
|
240
|
-
|
225
|
+
preParse(runtime, handlerRuby, handler);
|
226
|
+
parser.setContentHandler(handler);
|
227
|
+
parser.setErrorHandler(handler);
|
228
|
+
parser.setEntityResolver(new NokogiriEntityResolver(runtime, errorHandler, options));
|
241
229
|
|
242
230
|
try {
|
243
|
-
setProperty("http://xml.org/sax/properties/lexical-handler", handler);
|
231
|
+
parser.setProperty("http://xml.org/sax/properties/lexical-handler", handler);
|
244
232
|
} catch (Exception ex) {
|
245
233
|
throw runtime.newRuntimeError("Problem while creating XML SAX Parser: " + ex.toString());
|
246
234
|
}
|
@@ -270,8 +258,6 @@ public class XmlSaxParserContext extends ParserContext
|
|
270
258
|
|
271
259
|
postParse(runtime, handlerRuby, handler);
|
272
260
|
|
273
|
-
//maybeTrimLeadingAndTrailingWhitespace(context, handlerRuby);
|
274
|
-
|
275
261
|
return runtime.getNil();
|
276
262
|
}
|
277
263
|
|
@@ -319,53 +305,6 @@ public class XmlSaxParserContext extends ParserContext
|
|
319
305
|
return context.runtime.newBoolean(recovery);
|
320
306
|
}
|
321
307
|
|
322
|
-
/**
|
323
|
-
* If the handler's document is a FragmentHandler, attempt to trim
|
324
|
-
* leading and trailing whitespace.
|
325
|
-
*
|
326
|
-
* This is a bit hackish and depends heavily on the internals of
|
327
|
-
* FragmentHandler.
|
328
|
-
*/
|
329
|
-
protected void
|
330
|
-
maybeTrimLeadingAndTrailingWhitespace(ThreadContext context, IRubyObject parser)
|
331
|
-
{
|
332
|
-
RubyObjectAdapter adapter = JavaEmbedUtils.newObjectAdapter();
|
333
|
-
RubyModule mod = context.getRuntime().getClassFromPath("Nokogiri::XML::FragmentHandler");
|
334
|
-
|
335
|
-
IRubyObject handler = adapter.getInstanceVariable(parser, "@document");
|
336
|
-
if (handler == null || handler.isNil() || !adapter.isKindOf(handler, mod)) {
|
337
|
-
return;
|
338
|
-
}
|
339
|
-
IRubyObject stack = adapter.getInstanceVariable(handler, "@stack");
|
340
|
-
if (stack == null || stack.isNil()) {
|
341
|
-
return;
|
342
|
-
}
|
343
|
-
// doc is finally a DocumentFragment whose nodes we can check
|
344
|
-
IRubyObject doc = adapter.callMethod(stack, "first");
|
345
|
-
if (doc == null || doc.isNil()) {
|
346
|
-
return;
|
347
|
-
}
|
348
|
-
|
349
|
-
IRubyObject children;
|
350
|
-
|
351
|
-
for (;;) {
|
352
|
-
children = adapter.callMethod(doc, "children");
|
353
|
-
IRubyObject first = adapter.callMethod(children, "first");
|
354
|
-
if (NokogiriHelpers.isBlank(first)) { adapter.callMethod(first, "unlink"); }
|
355
|
-
else { break; }
|
356
|
-
}
|
357
|
-
|
358
|
-
for (;;) {
|
359
|
-
children = adapter.callMethod(doc, "children");
|
360
|
-
IRubyObject last = adapter.callMethod(children, "last");
|
361
|
-
if (NokogiriHelpers.isBlank(last)) { adapter.callMethod(last, "unlink"); }
|
362
|
-
else { break; }
|
363
|
-
}
|
364
|
-
|
365
|
-
// While we have a document, normalize it.
|
366
|
-
((XmlNode) doc).normalize();
|
367
|
-
}
|
368
|
-
|
369
308
|
@JRubyMethod(name = "column")
|
370
309
|
public IRubyObject
|
371
310
|
column(ThreadContext context)
|
@@ -383,5 +322,4 @@ public class XmlSaxParserContext extends ParserContext
|
|
383
322
|
if (number == null) { return context.getRuntime().getNil(); }
|
384
323
|
return RubyFixnum.newFixnum(context.getRuntime(), number.longValue());
|
385
324
|
}
|
386
|
-
|
387
325
|
}
|
@@ -1,20 +1,9 @@
|
|
1
1
|
package nokogiri;
|
2
2
|
|
3
|
-
import
|
4
|
-
import static org.jruby.runtime.Helpers.invoke;
|
5
|
-
|
6
|
-
import java.io.ByteArrayInputStream;
|
7
|
-
import java.io.IOException;
|
8
|
-
import java.io.InputStream;
|
9
|
-
import java.util.concurrent.ExecutionException;
|
10
|
-
import java.util.concurrent.ExecutorService;
|
11
|
-
import java.util.concurrent.Executors;
|
12
|
-
import java.util.concurrent.Future;
|
13
|
-
import java.util.concurrent.FutureTask;
|
14
|
-
import java.util.concurrent.ThreadFactory;
|
15
|
-
|
3
|
+
import nokogiri.internals.*;
|
16
4
|
import org.jruby.Ruby;
|
17
5
|
import org.jruby.RubyClass;
|
6
|
+
import org.jruby.RubyException;
|
18
7
|
import org.jruby.RubyObject;
|
19
8
|
import org.jruby.anno.JRubyClass;
|
20
9
|
import org.jruby.anno.JRubyMethod;
|
@@ -22,11 +11,14 @@ import org.jruby.exceptions.RaiseException;
|
|
22
11
|
import org.jruby.runtime.ThreadContext;
|
23
12
|
import org.jruby.runtime.builtin.IRubyObject;
|
24
13
|
|
25
|
-
import
|
26
|
-
import
|
27
|
-
import
|
28
|
-
import
|
29
|
-
import
|
14
|
+
import java.io.ByteArrayInputStream;
|
15
|
+
import java.io.IOException;
|
16
|
+
import java.io.InputStream;
|
17
|
+
import java.util.List;
|
18
|
+
import java.util.concurrent.*;
|
19
|
+
|
20
|
+
import static nokogiri.internals.NokogiriHelpers.getNokogiriClass;
|
21
|
+
import static org.jruby.runtime.Helpers.invoke;
|
30
22
|
|
31
23
|
/**
|
32
24
|
* Class for Nokogiri::XML::SAX::PushParser
|
@@ -159,7 +151,8 @@ public class XmlSaxPushParser extends RubyObject
|
|
159
151
|
|
160
152
|
if (!options.recover && parserTask.getErrorCount() > errorCount0) {
|
161
153
|
terminateTask(context.runtime);
|
162
|
-
|
154
|
+
ex = parserTask.getLastError().toThrowable();
|
155
|
+
throw ex;
|
163
156
|
}
|
164
157
|
|
165
158
|
return this;
|
@@ -278,16 +271,15 @@ public class XmlSaxPushParser extends RubyObject
|
|
278
271
|
getErrorCount()
|
279
272
|
{
|
280
273
|
// check for null because thread may not have started yet
|
281
|
-
if (parser.
|
282
|
-
return parser.
|
274
|
+
if (parser.getNokogiriErrorHandler() == null) { return 0; }
|
275
|
+
return parser.getNokogiriErrorHandler().getErrors().size();
|
283
276
|
}
|
284
277
|
|
285
|
-
synchronized final
|
278
|
+
synchronized final RubyException
|
286
279
|
getLastError()
|
287
280
|
{
|
288
|
-
|
281
|
+
List<RubyException> errors = parser.getNokogiriErrorHandler().getErrors();
|
282
|
+
return errors.get(errors.size() - 1);
|
289
283
|
}
|
290
|
-
|
291
284
|
}
|
292
|
-
|
293
285
|
}
|
@@ -85,7 +85,7 @@ public class NokogiriEntityResolver implements EntityResolver2
|
|
85
85
|
private void
|
86
86
|
addError(String errorMessage)
|
87
87
|
{
|
88
|
-
if (handler != null) { handler.
|
88
|
+
if (handler != null) { handler.addError(new Exception(errorMessage)); }
|
89
89
|
}
|
90
90
|
|
91
91
|
/**
|
@@ -1,11 +1,15 @@
|
|
1
1
|
package nokogiri.internals;
|
2
2
|
|
3
|
-
import
|
4
|
-
import java.util.List;
|
5
|
-
|
3
|
+
import nokogiri.XmlSyntaxError;
|
6
4
|
import org.apache.xerces.xni.parser.XMLErrorHandler;
|
5
|
+
import org.jruby.Ruby;
|
6
|
+
import org.jruby.RubyException;
|
7
|
+
import org.jruby.exceptions.RaiseException;
|
7
8
|
import org.xml.sax.ErrorHandler;
|
8
9
|
|
10
|
+
import java.util.ArrayList;
|
11
|
+
import java.util.List;
|
12
|
+
|
9
13
|
/**
|
10
14
|
* Super class of error handlers.
|
11
15
|
*
|
@@ -17,23 +21,40 @@ import org.xml.sax.ErrorHandler;
|
|
17
21
|
*/
|
18
22
|
public abstract class NokogiriErrorHandler implements ErrorHandler, XMLErrorHandler
|
19
23
|
{
|
20
|
-
|
24
|
+
private final Ruby runtime;
|
25
|
+
protected final List<RubyException> errors;
|
21
26
|
protected boolean noerror;
|
22
27
|
protected boolean nowarning;
|
23
28
|
|
24
29
|
public
|
25
|
-
NokogiriErrorHandler(boolean noerror, boolean nowarning)
|
30
|
+
NokogiriErrorHandler(Ruby runtime, boolean noerror, boolean nowarning)
|
26
31
|
{
|
27
|
-
this.
|
32
|
+
this.runtime = runtime;
|
33
|
+
this.errors = new ArrayList<RubyException>(4);
|
28
34
|
this.noerror = noerror;
|
29
35
|
this.nowarning = nowarning;
|
30
36
|
}
|
31
37
|
|
32
|
-
List<
|
38
|
+
public List<RubyException>
|
33
39
|
getErrors() { return errors; }
|
34
40
|
|
35
41
|
public void
|
36
|
-
addError(Exception ex)
|
42
|
+
addError(Exception ex)
|
43
|
+
{
|
44
|
+
addError(XmlSyntaxError.createXMLSyntaxError(runtime, ex));
|
45
|
+
}
|
46
|
+
|
47
|
+
public void
|
48
|
+
addError(RubyException ex)
|
49
|
+
{
|
50
|
+
errors.add(ex);
|
51
|
+
}
|
52
|
+
|
53
|
+
public void
|
54
|
+
addError(RaiseException ex)
|
55
|
+
{
|
56
|
+
addError(ex.getException());
|
57
|
+
}
|
37
58
|
|
38
59
|
protected boolean
|
39
60
|
usesNekoHtml(String domain)
|
@@ -38,25 +38,19 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler
|
|
38
38
|
private final Ruby runtime;
|
39
39
|
private final RubyClass attrClass;
|
40
40
|
private final IRubyObject object;
|
41
|
-
|
42
|
-
/**
|
43
|
-
* Stores parse errors with the most-recent error last.
|
44
|
-
*
|
45
|
-
* TODO: should these be stored in the document 'errors' array?
|
46
|
-
* Currently only string messages are stored there.
|
47
|
-
*/
|
48
|
-
private final LinkedList<RaiseException> errors = new LinkedList<RaiseException>();
|
41
|
+
private NokogiriErrorHandler errorHandler;
|
49
42
|
|
50
43
|
private Locator locator;
|
51
44
|
private boolean needEmptyAttrCheck;
|
52
45
|
|
53
46
|
public
|
54
|
-
NokogiriHandler(Ruby runtime, IRubyObject object)
|
47
|
+
NokogiriHandler(Ruby runtime, IRubyObject object, NokogiriErrorHandler errorHandler)
|
55
48
|
{
|
56
49
|
assert object != null;
|
57
50
|
this.runtime = runtime;
|
58
51
|
this.attrClass = (RubyClass) runtime.getClassFromPath("Nokogiri::XML::SAX::Parser::Attribute");
|
59
52
|
this.object = object;
|
53
|
+
this.errorHandler = errorHandler;
|
60
54
|
charactersBuilder = new StringBuilder();
|
61
55
|
String objectName = object.getMetaClass().getName();
|
62
56
|
if ("Nokogiri::HTML4::SAX::Parser".equals(objectName)) { needEmptyAttrCheck = true; }
|
@@ -253,9 +247,9 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler
|
|
253
247
|
try {
|
254
248
|
final String msg = ex.getMessage();
|
255
249
|
call("error", runtime.newString(msg == null ? "" : msg));
|
256
|
-
addError(
|
250
|
+
errorHandler.addError(ex);
|
257
251
|
} catch (RaiseException e) {
|
258
|
-
addError(e);
|
252
|
+
errorHandler.addError(e);
|
259
253
|
throw e;
|
260
254
|
}
|
261
255
|
}
|
@@ -282,22 +276,10 @@ public class NokogiriHandler extends DefaultHandler2 implements XmlDeclHandler
|
|
282
276
|
call("warning", runtime.newString(msg == null ? "" : msg));
|
283
277
|
}
|
284
278
|
|
285
|
-
protected synchronized void
|
286
|
-
addError(RaiseException e)
|
287
|
-
{
|
288
|
-
errors.add(e);
|
289
|
-
}
|
290
|
-
|
291
279
|
public synchronized int
|
292
280
|
getErrorCount()
|
293
281
|
{
|
294
|
-
return
|
295
|
-
}
|
296
|
-
|
297
|
-
public synchronized RaiseException
|
298
|
-
getLastError()
|
299
|
-
{
|
300
|
-
return errors.getLast();
|
282
|
+
return errorHandler.getErrors().size();
|
301
283
|
}
|
302
284
|
|
303
285
|
private void
|
@@ -1,6 +1,7 @@
|
|
1
1
|
package nokogiri.internals;
|
2
2
|
|
3
3
|
import org.apache.xerces.xni.parser.XMLParseException;
|
4
|
+
import org.jruby.Ruby;
|
4
5
|
import org.xml.sax.SAXException;
|
5
6
|
import org.xml.sax.SAXParseException;
|
6
7
|
|
@@ -13,21 +14,21 @@ import org.xml.sax.SAXParseException;
|
|
13
14
|
public class NokogiriNonStrictErrorHandler extends NokogiriErrorHandler
|
14
15
|
{
|
15
16
|
public
|
16
|
-
NokogiriNonStrictErrorHandler(boolean noerror, boolean nowarning)
|
17
|
+
NokogiriNonStrictErrorHandler(Ruby runtime, boolean noerror, boolean nowarning)
|
17
18
|
{
|
18
|
-
super(noerror, nowarning);
|
19
|
+
super(runtime, noerror, nowarning);
|
19
20
|
}
|
20
21
|
|
21
22
|
public void
|
22
23
|
warning(SAXParseException ex) throws SAXException
|
23
24
|
{
|
24
|
-
|
25
|
+
addError(ex);
|
25
26
|
}
|
26
27
|
|
27
28
|
public void
|
28
29
|
error(SAXParseException ex) throws SAXException
|
29
30
|
{
|
30
|
-
|
31
|
+
addError(ex);
|
31
32
|
}
|
32
33
|
|
33
34
|
public void
|
@@ -38,7 +39,7 @@ public class NokogiriNonStrictErrorHandler extends NokogiriErrorHandler
|
|
38
39
|
// found in the prolog, instead it will keep calling this method and we'll
|
39
40
|
// keep inserting the error in the document errors array until we run
|
40
41
|
// out of memory
|
41
|
-
|
42
|
+
addError(ex);
|
42
43
|
String message = ex.getMessage();
|
43
44
|
|
44
45
|
// The problem with Xerces is that some errors will cause the
|
@@ -53,19 +54,19 @@ public class NokogiriNonStrictErrorHandler extends NokogiriErrorHandler
|
|
53
54
|
public void
|
54
55
|
error(String domain, String key, XMLParseException e)
|
55
56
|
{
|
56
|
-
|
57
|
+
addError(e);
|
57
58
|
}
|
58
59
|
|
59
60
|
public void
|
60
61
|
fatalError(String domain, String key, XMLParseException e)
|
61
62
|
{
|
62
|
-
|
63
|
+
addError(e);
|
63
64
|
}
|
64
65
|
|
65
66
|
public void
|
66
67
|
warning(String domain, String key, XMLParseException e)
|
67
68
|
{
|
68
|
-
|
69
|
+
addError(e);
|
69
70
|
}
|
70
71
|
|
71
72
|
/*
|
@@ -1,6 +1,7 @@
|
|
1
1
|
package nokogiri.internals;
|
2
2
|
|
3
3
|
import org.apache.xerces.xni.parser.XMLParseException;
|
4
|
+
import org.jruby.Ruby;
|
4
5
|
import org.xml.sax.SAXException;
|
5
6
|
import org.xml.sax.SAXParseException;
|
6
7
|
|
@@ -20,15 +21,15 @@ public class NokogiriNonStrictErrorHandler4NekoHtml extends NokogiriErrorHandler
|
|
20
21
|
{
|
21
22
|
|
22
23
|
public
|
23
|
-
NokogiriNonStrictErrorHandler4NekoHtml(boolean nowarning)
|
24
|
+
NokogiriNonStrictErrorHandler4NekoHtml(Ruby runtime, boolean nowarning)
|
24
25
|
{
|
25
|
-
super(false, nowarning);
|
26
|
+
super(runtime, false, nowarning);
|
26
27
|
}
|
27
28
|
|
28
29
|
public
|
29
|
-
NokogiriNonStrictErrorHandler4NekoHtml(boolean noerror, boolean nowarning)
|
30
|
+
NokogiriNonStrictErrorHandler4NekoHtml(Ruby runtime, boolean noerror, boolean nowarning)
|
30
31
|
{
|
31
|
-
super(noerror, nowarning);
|
32
|
+
super(runtime, noerror, nowarning);
|
32
33
|
}
|
33
34
|
|
34
35
|
public void
|
@@ -40,13 +41,13 @@ public class NokogiriNonStrictErrorHandler4NekoHtml extends NokogiriErrorHandler
|
|
40
41
|
public void
|
41
42
|
error(SAXParseException ex) throws SAXException
|
42
43
|
{
|
43
|
-
|
44
|
+
addError(ex);
|
44
45
|
}
|
45
46
|
|
46
47
|
public void
|
47
48
|
fatalError(SAXParseException ex) throws SAXException
|
48
49
|
{
|
49
|
-
|
50
|
+
addError(ex);
|
50
51
|
}
|
51
52
|
|
52
53
|
/**
|
@@ -64,7 +65,7 @@ public class NokogiriNonStrictErrorHandler4NekoHtml extends NokogiriErrorHandler
|
|
64
65
|
public void
|
65
66
|
error(String domain, String key, XMLParseException e)
|
66
67
|
{
|
67
|
-
|
68
|
+
addError(e);
|
68
69
|
}
|
69
70
|
|
70
71
|
/**
|
@@ -82,7 +83,7 @@ public class NokogiriNonStrictErrorHandler4NekoHtml extends NokogiriErrorHandler
|
|
82
83
|
public void
|
83
84
|
fatalError(String domain, String key, XMLParseException e)
|
84
85
|
{
|
85
|
-
|
86
|
+
addError(e);
|
86
87
|
}
|
87
88
|
|
88
89
|
/**
|
@@ -100,7 +101,7 @@ public class NokogiriNonStrictErrorHandler4NekoHtml extends NokogiriErrorHandler
|
|
100
101
|
public void
|
101
102
|
warning(String domain, String key, XMLParseException e)
|
102
103
|
{
|
103
|
-
|
104
|
+
addError(e);
|
104
105
|
}
|
105
106
|
|
106
107
|
}
|
@@ -1,6 +1,7 @@
|
|
1
1
|
package nokogiri.internals;
|
2
2
|
|
3
3
|
import org.apache.xerces.xni.parser.XMLParseException;
|
4
|
+
import org.jruby.Ruby;
|
4
5
|
import org.xml.sax.SAXException;
|
5
6
|
import org.xml.sax.SAXParseException;
|
6
7
|
|
@@ -14,23 +15,23 @@ import org.xml.sax.SAXParseException;
|
|
14
15
|
public class NokogiriStrictErrorHandler extends NokogiriErrorHandler
|
15
16
|
{
|
16
17
|
public
|
17
|
-
NokogiriStrictErrorHandler(boolean noerror, boolean nowarning)
|
18
|
+
NokogiriStrictErrorHandler(Ruby runtime, boolean noerror, boolean nowarning)
|
18
19
|
{
|
19
|
-
super(noerror, nowarning);
|
20
|
+
super(runtime, noerror, nowarning);
|
20
21
|
}
|
21
22
|
|
22
23
|
public void
|
23
24
|
warning(SAXParseException spex) throws SAXException
|
24
25
|
{
|
25
26
|
if (!nowarning) { throw spex; }
|
26
|
-
else {
|
27
|
+
else { addError(spex); }
|
27
28
|
}
|
28
29
|
|
29
30
|
public void
|
30
31
|
error(SAXParseException spex) throws SAXException
|
31
32
|
{
|
32
33
|
if (!noerror) { throw spex; }
|
33
|
-
else {
|
34
|
+
else { addError(spex); }
|
34
35
|
}
|
35
36
|
|
36
37
|
public void
|
@@ -43,7 +44,7 @@ public class NokogiriStrictErrorHandler extends NokogiriErrorHandler
|
|
43
44
|
error(String domain, String key, XMLParseException e) throws XMLParseException
|
44
45
|
{
|
45
46
|
if (!noerror) { throw e; }
|
46
|
-
else {
|
47
|
+
else { addError(e); }
|
47
48
|
}
|
48
49
|
|
49
50
|
public void
|
@@ -56,6 +57,6 @@ public class NokogiriStrictErrorHandler extends NokogiriErrorHandler
|
|
56
57
|
warning(String domain, String key, XMLParseException e) throws XMLParseException
|
57
58
|
{
|
58
59
|
if (!nowarning) { throw e; }
|
59
|
-
|
60
|
+
else { addError(e); }
|
60
61
|
}
|
61
62
|
}
|
@@ -673,6 +673,8 @@ public class SaveContextVisitor
|
|
673
673
|
if (!isEmpty(name) && noEmpty) {
|
674
674
|
buffer.append("</").append(name).append('>');
|
675
675
|
}
|
676
|
+
} else if (asXhtml && !isEmpty(name)) {
|
677
|
+
buffer.append("</").append(name).append('>');
|
676
678
|
}
|
677
679
|
if (needBreakInClosing(element)) {
|
678
680
|
if (!containsText(element)) { indentation.pop(); }
|
@@ -1,30 +1,24 @@
|
|
1
1
|
package nokogiri.internals;
|
2
2
|
|
3
|
-
import
|
4
|
-
import
|
5
|
-
|
6
|
-
import java.io.IOException;
|
7
|
-
import java.util.ArrayList;
|
8
|
-
import java.util.List;
|
9
|
-
|
3
|
+
import nokogiri.XmlDocument;
|
4
|
+
import nokogiri.XmlDtd;
|
5
|
+
import nokogiri.XmlSyntaxError;
|
10
6
|
import org.apache.xerces.parsers.DOMParser;
|
11
|
-
import org.jruby
|
12
|
-
import org.jruby.RubyArray;
|
13
|
-
import org.jruby.RubyClass;
|
14
|
-
import org.jruby.RubyFixnum;
|
7
|
+
import org.jruby.*;
|
15
8
|
import org.jruby.exceptions.RaiseException;
|
16
|
-
import org.jruby.runtime.ThreadContext;
|
17
9
|
import org.jruby.runtime.Helpers;
|
10
|
+
import org.jruby.runtime.ThreadContext;
|
18
11
|
import org.jruby.runtime.builtin.IRubyObject;
|
19
12
|
import org.w3c.dom.Document;
|
20
13
|
import org.w3c.dom.Node;
|
21
14
|
import org.w3c.dom.NodeList;
|
22
15
|
import org.xml.sax.SAXException;
|
23
16
|
|
24
|
-
import
|
25
|
-
import
|
26
|
-
import
|
27
|
-
|
17
|
+
import java.io.IOException;
|
18
|
+
import java.util.ArrayList;
|
19
|
+
import java.util.List;
|
20
|
+
|
21
|
+
import static nokogiri.internals.NokogiriHelpers.isBlank;
|
28
22
|
|
29
23
|
/**
|
30
24
|
* Parser class for XML DOM processing. This class actually parses XML document
|
@@ -48,7 +42,6 @@ public class XmlDomParserContext extends ParserContext
|
|
48
42
|
protected static final String FEATURE_NOT_EXPAND_ENTITY =
|
49
43
|
"http://apache.org/xml/features/dom/create-entity-ref-nodes";
|
50
44
|
protected static final String FEATURE_VALIDATION = "http://xml.org/sax/features/validation";
|
51
|
-
private static final String XINCLUDE_FEATURE_ID = "http://apache.org/xml/features/xinclude";
|
52
45
|
private static final String SECURITY_MANAGER = "http://apache.org/xml/properties/security-manager";
|
53
46
|
|
54
47
|
protected ParserContext.Options options;
|
@@ -69,17 +62,17 @@ public class XmlDomParserContext extends ParserContext
|
|
69
62
|
this.options = new ParserContext.Options(RubyFixnum.fix2long(options));
|
70
63
|
java_encoding = NokogiriHelpers.getValidEncodingOrNull(encoding);
|
71
64
|
ruby_encoding = encoding;
|
72
|
-
initErrorHandler();
|
65
|
+
initErrorHandler(runtime);
|
73
66
|
initParser(runtime);
|
74
67
|
}
|
75
68
|
|
76
69
|
protected void
|
77
|
-
initErrorHandler()
|
70
|
+
initErrorHandler(Ruby runtime)
|
78
71
|
{
|
79
72
|
if (options.recover) {
|
80
|
-
errorHandler = new NokogiriNonStrictErrorHandler(options.noError, options.noWarning);
|
73
|
+
errorHandler = new NokogiriNonStrictErrorHandler(runtime, options.noError, options.noWarning);
|
81
74
|
} else {
|
82
|
-
errorHandler = new NokogiriStrictErrorHandler(options.noError, options.noWarning);
|
75
|
+
errorHandler = new NokogiriStrictErrorHandler(runtime, options.noError, options.noWarning);
|
83
76
|
}
|
84
77
|
}
|
85
78
|
|
@@ -161,12 +154,10 @@ public class XmlDomParserContext extends ParserContext
|
|
161
154
|
mapErrors(ThreadContext context, NokogiriErrorHandler errorHandler)
|
162
155
|
{
|
163
156
|
final Ruby runtime = context.runtime;
|
164
|
-
final List<
|
157
|
+
final List<RubyException> errors = errorHandler.getErrors();
|
165
158
|
final IRubyObject[] errorsAry = new IRubyObject[errors.size()];
|
166
159
|
for (int i = 0; i < errors.size(); i++) {
|
167
|
-
|
168
|
-
xmlSyntaxError.setException(errors.get(i));
|
169
|
-
errorsAry[i] = xmlSyntaxError;
|
160
|
+
errorsAry[i] = errors.get(i);
|
170
161
|
}
|
171
162
|
return runtime.newArrayNoCopy(errorsAry);
|
172
163
|
}
|