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 +1,639 @@
|
|
1
|
+
/**
|
2
|
+
* Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
* or more contributor license agreements. See the NOTICE file
|
4
|
+
* distributed with this work for additional information
|
5
|
+
* regarding copyright ownership. The ASF licenses this file
|
6
|
+
* to you under the Apache License, Version 2.0 (the
|
7
|
+
* "License"); you may not use this file except in compliance
|
8
|
+
* with the License. You may obtain a copy of the License at
|
9
|
+
*
|
10
|
+
* http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
*
|
12
|
+
* Unless required by applicable law or agreed to in writing,
|
13
|
+
* software distributed under the License is distributed on an
|
14
|
+
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
* KIND, either express or implied. See the License for the
|
16
|
+
* specific language governing permissions and limitations
|
17
|
+
* under the License.
|
18
|
+
*/
|
19
|
+
package nokogiri.internals.c14n;
|
20
|
+
|
21
|
+
import java.net.URI;
|
22
|
+
import java.net.URISyntaxException;
|
23
|
+
import java.util.ArrayList;
|
24
|
+
import java.util.Collection;
|
25
|
+
import java.util.HashMap;
|
26
|
+
import java.util.Iterator;
|
27
|
+
import java.util.List;
|
28
|
+
import java.util.Map;
|
29
|
+
import java.util.SortedSet;
|
30
|
+
import java.util.TreeSet;
|
31
|
+
|
32
|
+
|
33
|
+
import org.w3c.dom.Attr;
|
34
|
+
import org.w3c.dom.Element;
|
35
|
+
import org.w3c.dom.NamedNodeMap;
|
36
|
+
import org.w3c.dom.Node;
|
37
|
+
|
38
|
+
/**
|
39
|
+
* Implements <A HREF="http://www.w3.org/TR/2008/PR-xml-c14n11-20080129/">
|
40
|
+
* Canonical XML Version 1.1</A>, a W3C Proposed Recommendation from 29
|
41
|
+
* January 2008.
|
42
|
+
*
|
43
|
+
* @author Sean Mullan
|
44
|
+
* @author Raul Benito
|
45
|
+
*/
|
46
|
+
public abstract class Canonicalizer11 extends CanonicalizerBase {
|
47
|
+
|
48
|
+
private static final String XMLNS_URI = Constants.NamespaceSpecNS;
|
49
|
+
private static final String XML_LANG_URI = Constants.XML_LANG_SPACE_SpecNS;
|
50
|
+
private final SortedSet<Attr> result = new TreeSet<Attr>(COMPARE);
|
51
|
+
|
52
|
+
private boolean firstCall = true;
|
53
|
+
|
54
|
+
private static class XmlAttrStack {
|
55
|
+
static class XmlsStackElement {
|
56
|
+
int level;
|
57
|
+
boolean rendered = false;
|
58
|
+
List<Attr> nodes = new ArrayList<Attr>();
|
59
|
+
}
|
60
|
+
|
61
|
+
int currentLevel = 0;
|
62
|
+
int lastlevel = 0;
|
63
|
+
XmlsStackElement cur;
|
64
|
+
List<XmlsStackElement> levels = new ArrayList<XmlsStackElement>();
|
65
|
+
|
66
|
+
void push(int level) {
|
67
|
+
currentLevel = level;
|
68
|
+
if (currentLevel == -1) {
|
69
|
+
return;
|
70
|
+
}
|
71
|
+
cur = null;
|
72
|
+
while (lastlevel >= currentLevel) {
|
73
|
+
levels.remove(levels.size() - 1);
|
74
|
+
int newSize = levels.size();
|
75
|
+
if (newSize == 0) {
|
76
|
+
lastlevel = 0;
|
77
|
+
return;
|
78
|
+
}
|
79
|
+
lastlevel = (levels.get(newSize - 1)).level;
|
80
|
+
}
|
81
|
+
}
|
82
|
+
|
83
|
+
void addXmlnsAttr(Attr n) {
|
84
|
+
if (cur == null) {
|
85
|
+
cur = new XmlsStackElement();
|
86
|
+
cur.level = currentLevel;
|
87
|
+
levels.add(cur);
|
88
|
+
lastlevel = currentLevel;
|
89
|
+
}
|
90
|
+
cur.nodes.add(n);
|
91
|
+
}
|
92
|
+
|
93
|
+
void getXmlnsAttr(Collection<Attr> col) throws CanonicalizationException {
|
94
|
+
int size = levels.size() - 1;
|
95
|
+
if (cur == null) {
|
96
|
+
cur = new XmlsStackElement();
|
97
|
+
cur.level = currentLevel;
|
98
|
+
lastlevel = currentLevel;
|
99
|
+
levels.add(cur);
|
100
|
+
}
|
101
|
+
boolean parentRendered = false;
|
102
|
+
if (size == -1) {
|
103
|
+
parentRendered = true;
|
104
|
+
} else {
|
105
|
+
XmlsStackElement e = levels.get(size);
|
106
|
+
if (e.rendered && e.level + 1 == currentLevel) {
|
107
|
+
parentRendered = true;
|
108
|
+
}
|
109
|
+
}
|
110
|
+
if (parentRendered) {
|
111
|
+
col.addAll(cur.nodes);
|
112
|
+
cur.rendered = true;
|
113
|
+
return;
|
114
|
+
}
|
115
|
+
|
116
|
+
Map<String, Attr> loa = new HashMap<String, Attr>();
|
117
|
+
List<Attr> baseAttrs = new ArrayList<Attr>();
|
118
|
+
boolean successiveOmitted = true;
|
119
|
+
for (; size >= 0; size--) {
|
120
|
+
XmlsStackElement e = levels.get(size);
|
121
|
+
if (e.rendered) {
|
122
|
+
successiveOmitted = false;
|
123
|
+
}
|
124
|
+
Iterator<Attr> it = e.nodes.iterator();
|
125
|
+
while (it.hasNext() && successiveOmitted) {
|
126
|
+
Attr n = it.next();
|
127
|
+
if (n.getLocalName().equals("base") && !e.rendered) {
|
128
|
+
baseAttrs.add(n);
|
129
|
+
} else if (!loa.containsKey(n.getName())) {
|
130
|
+
loa.put(n.getName(), n);
|
131
|
+
}
|
132
|
+
}
|
133
|
+
}
|
134
|
+
if (!baseAttrs.isEmpty()) {
|
135
|
+
Iterator<Attr> it = col.iterator();
|
136
|
+
String base = null;
|
137
|
+
Attr baseAttr = null;
|
138
|
+
while (it.hasNext()) {
|
139
|
+
Attr n = it.next();
|
140
|
+
if (n.getLocalName().equals("base")) {
|
141
|
+
base = n.getValue();
|
142
|
+
baseAttr = n;
|
143
|
+
break;
|
144
|
+
}
|
145
|
+
}
|
146
|
+
it = baseAttrs.iterator();
|
147
|
+
while (it.hasNext()) {
|
148
|
+
Attr n = it.next();
|
149
|
+
if (base == null) {
|
150
|
+
base = n.getValue();
|
151
|
+
baseAttr = n;
|
152
|
+
} else {
|
153
|
+
try {
|
154
|
+
base = joinURI(n.getValue(), base);
|
155
|
+
} catch (URISyntaxException e1) {
|
156
|
+
throw new CanonicalizationException(e1);
|
157
|
+
}
|
158
|
+
}
|
159
|
+
}
|
160
|
+
if (base != null && base.length() != 0) {
|
161
|
+
baseAttr.setValue(base);
|
162
|
+
col.add(baseAttr);
|
163
|
+
}
|
164
|
+
}
|
165
|
+
|
166
|
+
cur.rendered = true;
|
167
|
+
col.addAll(loa.values());
|
168
|
+
}
|
169
|
+
}
|
170
|
+
|
171
|
+
private final XmlAttrStack xmlattrStack = new XmlAttrStack();
|
172
|
+
|
173
|
+
/**
|
174
|
+
* Constructor Canonicalizer11
|
175
|
+
*
|
176
|
+
* @param includeComments
|
177
|
+
*/
|
178
|
+
public Canonicalizer11(boolean includeComments) {
|
179
|
+
super(includeComments);
|
180
|
+
}
|
181
|
+
|
182
|
+
/**
|
183
|
+
* Always throws a CanonicalizationException because this is inclusive c14n.
|
184
|
+
*
|
185
|
+
* @param rootNode
|
186
|
+
* @param inclusiveNamespaces
|
187
|
+
* @return none it always fails
|
188
|
+
* @throws CanonicalizationException
|
189
|
+
*/
|
190
|
+
public byte[] engineCanonicalizeSubTree(
|
191
|
+
Node rootNode, String inclusiveNamespaces, CanonicalFilter filter
|
192
|
+
) throws CanonicalizationException {
|
193
|
+
throw new CanonicalizationException("c14n.Canonicalizer.UnsupportedOperation");
|
194
|
+
}
|
195
|
+
|
196
|
+
/**
|
197
|
+
* Returns the Attr[]s to be output for the given element.
|
198
|
+
* <br>
|
199
|
+
* The code of this method is a copy of {@link #handleAttributes(Element,
|
200
|
+
* NameSpaceSymbTable)},
|
201
|
+
* whereas it takes into account that subtree-c14n is -- well --
|
202
|
+
* subtree-based.
|
203
|
+
* So if the element in question isRoot of c14n, it's parent is not in the
|
204
|
+
* node set, as well as all other ancestors.
|
205
|
+
*
|
206
|
+
* @param element
|
207
|
+
* @param ns
|
208
|
+
* @return the Attr[]s to be output
|
209
|
+
* @throws CanonicalizationException
|
210
|
+
* @throws URISyntaxException
|
211
|
+
*/
|
212
|
+
@Override
|
213
|
+
protected Iterator<Attr> handleAttributesSubtree(Element element, NameSpaceSymbTable ns)
|
214
|
+
throws CanonicalizationException {
|
215
|
+
if (!element.hasAttributes() && !firstCall) {
|
216
|
+
return null;
|
217
|
+
}
|
218
|
+
// result will contain the attrs which have to be output
|
219
|
+
final SortedSet<Attr> result = this.result;
|
220
|
+
result.clear();
|
221
|
+
|
222
|
+
if (element.hasAttributes()) {
|
223
|
+
NamedNodeMap attrs = element.getAttributes();
|
224
|
+
int attrsLength = attrs.getLength();
|
225
|
+
|
226
|
+
for (int i = 0; i < attrsLength; i++) {
|
227
|
+
Attr attribute = (Attr) attrs.item(i);
|
228
|
+
String NUri = attribute.getNamespaceURI();
|
229
|
+
String NName = attribute.getLocalName();
|
230
|
+
String NValue = attribute.getValue();
|
231
|
+
|
232
|
+
if (!XMLNS_URI.equals(NUri)) {
|
233
|
+
// It's not a namespace attr node. Add to the result and continue.
|
234
|
+
result.add(attribute);
|
235
|
+
} else if (!(XML.equals(NName) && XML_LANG_URI.equals(NValue))) {
|
236
|
+
// The default mapping for xml must not be output.
|
237
|
+
Node n = ns.addMappingAndRender(NName, NValue, attribute);
|
238
|
+
|
239
|
+
if (n != null) {
|
240
|
+
// Render the ns definition
|
241
|
+
result.add((Attr)n);
|
242
|
+
if (C14nHelper.namespaceIsRelative(attribute)) {
|
243
|
+
Object exArgs[] = {element.getTagName(), NName, attribute.getNodeValue()};
|
244
|
+
throw new CanonicalizationException(
|
245
|
+
"c14n.Canonicalizer.RelativeNamespace", exArgs
|
246
|
+
);
|
247
|
+
}
|
248
|
+
}
|
249
|
+
}
|
250
|
+
}
|
251
|
+
}
|
252
|
+
|
253
|
+
if (firstCall) {
|
254
|
+
// It is the first node of the subtree
|
255
|
+
// Obtain all the namespaces defined in the parents, and added to the output.
|
256
|
+
ns.getUnrenderedNodes(result);
|
257
|
+
// output the attributes in the xml namespace.
|
258
|
+
xmlattrStack.getXmlnsAttr(result);
|
259
|
+
firstCall = false;
|
260
|
+
}
|
261
|
+
|
262
|
+
return result.iterator();
|
263
|
+
}
|
264
|
+
|
265
|
+
/**
|
266
|
+
* Returns the Attr[]s to be output for the given element.
|
267
|
+
* <br>
|
268
|
+
* IMPORTANT: This method expects to work on a modified DOM tree, i.e. a
|
269
|
+
* DOM which has been prepared using
|
270
|
+
* {@link nokogiri.internals.c14n.security.utils.XMLUtils#circumventBug2650(
|
271
|
+
* org.w3c.dom.Document)}.
|
272
|
+
*
|
273
|
+
* @param element
|
274
|
+
* @param ns
|
275
|
+
* @return the Attr[]s to be output
|
276
|
+
* @throws CanonicalizationException
|
277
|
+
*/
|
278
|
+
@Override
|
279
|
+
protected Iterator<Attr> handleAttributes(Element element, NameSpaceSymbTable ns)
|
280
|
+
throws CanonicalizationException {
|
281
|
+
// result will contain the attrs which have to be output
|
282
|
+
xmlattrStack.push(ns.getLevel());
|
283
|
+
boolean isRealVisible = isVisibleDO(element, ns.getLevel()) == 1;
|
284
|
+
final SortedSet<Attr> result = this.result;
|
285
|
+
result.clear();
|
286
|
+
|
287
|
+
if (element.hasAttributes()) {
|
288
|
+
NamedNodeMap attrs = element.getAttributes();
|
289
|
+
int attrsLength = attrs.getLength();
|
290
|
+
|
291
|
+
for (int i = 0; i < attrsLength; i++) {
|
292
|
+
Attr attribute = (Attr) attrs.item(i);
|
293
|
+
String NUri = attribute.getNamespaceURI();
|
294
|
+
String NName = attribute.getLocalName();
|
295
|
+
String NValue = attribute.getValue();
|
296
|
+
|
297
|
+
if (!XMLNS_URI.equals(NUri)) {
|
298
|
+
//A non namespace definition node.
|
299
|
+
if (XML_LANG_URI.equals(NUri)) {
|
300
|
+
if (NName.equals("id")) {
|
301
|
+
if (isRealVisible) {
|
302
|
+
// treat xml:id like any other attribute
|
303
|
+
// (emit it, but don't inherit it)
|
304
|
+
result.add(attribute);
|
305
|
+
}
|
306
|
+
} else {
|
307
|
+
xmlattrStack.addXmlnsAttr(attribute);
|
308
|
+
}
|
309
|
+
} else if (isRealVisible) {
|
310
|
+
//The node is visible add the attribute to the list of output attributes.
|
311
|
+
result.add(attribute);
|
312
|
+
}
|
313
|
+
} else if (!XML.equals(NName) || !XML_LANG_URI.equals(NValue)) {
|
314
|
+
/* except omit namespace node with local name xml, which defines
|
315
|
+
* the xml prefix, if its string value is
|
316
|
+
* http://www.w3.org/XML/1998/namespace.
|
317
|
+
*/
|
318
|
+
// add the prefix binding to the ns symb table.
|
319
|
+
if (isVisible(attribute)) {
|
320
|
+
if (isRealVisible || !ns.removeMappingIfRender(NName)) {
|
321
|
+
// The xpath select this node output it if needed.
|
322
|
+
Node n = ns.addMappingAndRender(NName, NValue, attribute);
|
323
|
+
if (n != null) {
|
324
|
+
result.add((Attr)n);
|
325
|
+
if (C14nHelper.namespaceIsRelative(attribute)) {
|
326
|
+
Object exArgs[] = { element.getTagName(), NName, attribute.getNodeValue() };
|
327
|
+
throw new CanonicalizationException(
|
328
|
+
"c14n.Canonicalizer.RelativeNamespace", exArgs
|
329
|
+
);
|
330
|
+
}
|
331
|
+
}
|
332
|
+
}
|
333
|
+
} else {
|
334
|
+
if (isRealVisible && !XMLNS.equals(NName)) {
|
335
|
+
ns.removeMapping(NName);
|
336
|
+
} else {
|
337
|
+
ns.addMapping(NName, NValue, attribute);
|
338
|
+
}
|
339
|
+
}
|
340
|
+
}
|
341
|
+
}
|
342
|
+
}
|
343
|
+
|
344
|
+
if (isRealVisible) {
|
345
|
+
//The element is visible, handle the xmlns definition
|
346
|
+
Attr xmlns = element.getAttributeNodeNS(XMLNS_URI, XMLNS);
|
347
|
+
Node n = null;
|
348
|
+
if (xmlns == null) {
|
349
|
+
//No xmlns def just get the already defined.
|
350
|
+
n = ns.getMapping(XMLNS);
|
351
|
+
} else if (!isVisible(xmlns)) {
|
352
|
+
//There is a definition but the xmlns is not selected by the xpath.
|
353
|
+
//then xmlns=""
|
354
|
+
n = ns.addMappingAndRender(XMLNS, "", nullNode);
|
355
|
+
}
|
356
|
+
//output the xmlns def if needed.
|
357
|
+
if (n != null) {
|
358
|
+
result.add((Attr)n);
|
359
|
+
}
|
360
|
+
//Float all xml:* attributes of the unselected parent elements to this one.
|
361
|
+
xmlattrStack.getXmlnsAttr(result);
|
362
|
+
ns.getUnrenderedNodes(result);
|
363
|
+
}
|
364
|
+
|
365
|
+
return result.iterator();
|
366
|
+
}
|
367
|
+
|
368
|
+
@Override
|
369
|
+
protected void handleParent(Element e, NameSpaceSymbTable ns) {
|
370
|
+
if (!e.hasAttributes() && e.getNamespaceURI() == null) {
|
371
|
+
return;
|
372
|
+
}
|
373
|
+
xmlattrStack.push(-1);
|
374
|
+
NamedNodeMap attrs = e.getAttributes();
|
375
|
+
int attrsLength = attrs.getLength();
|
376
|
+
for (int i = 0; i < attrsLength; i++) {
|
377
|
+
Attr attribute = (Attr) attrs.item(i);
|
378
|
+
String NName = attribute.getLocalName();
|
379
|
+
String NValue = attribute.getNodeValue();
|
380
|
+
|
381
|
+
if (Constants.NamespaceSpecNS.equals(attribute.getNamespaceURI())) {
|
382
|
+
if (!XML.equals(NName) || !Constants.XML_LANG_SPACE_SpecNS.equals(NValue)) {
|
383
|
+
ns.addMapping(NName, NValue, attribute);
|
384
|
+
}
|
385
|
+
} else if (!"id".equals(NName) && XML_LANG_URI.equals(attribute.getNamespaceURI())) {
|
386
|
+
xmlattrStack.addXmlnsAttr(attribute);
|
387
|
+
}
|
388
|
+
}
|
389
|
+
if (e.getNamespaceURI() != null) {
|
390
|
+
String NName = e.getPrefix();
|
391
|
+
String NValue = e.getNamespaceURI();
|
392
|
+
String Name;
|
393
|
+
if (NName == null || NName.equals("")) {
|
394
|
+
NName = "xmlns";
|
395
|
+
Name = "xmlns";
|
396
|
+
} else {
|
397
|
+
Name = "xmlns:" + NName;
|
398
|
+
}
|
399
|
+
Attr n = e.getOwnerDocument().createAttributeNS("http://www.w3.org/2000/xmlns/", Name);
|
400
|
+
n.setValue(NValue);
|
401
|
+
ns.addMapping(NName, NValue, n);
|
402
|
+
}
|
403
|
+
}
|
404
|
+
|
405
|
+
private static String joinURI(String baseURI, String relativeURI) throws URISyntaxException {
|
406
|
+
String bscheme = null;
|
407
|
+
String bauthority = null;
|
408
|
+
String bpath = "";
|
409
|
+
String bquery = null;
|
410
|
+
|
411
|
+
// pre-parse the baseURI
|
412
|
+
if (baseURI != null) {
|
413
|
+
if (baseURI.endsWith("..")) {
|
414
|
+
baseURI = baseURI + "/";
|
415
|
+
}
|
416
|
+
URI base = new URI(baseURI);
|
417
|
+
bscheme = base.getScheme();
|
418
|
+
bauthority = base.getAuthority();
|
419
|
+
bpath = base.getPath();
|
420
|
+
bquery = base.getQuery();
|
421
|
+
}
|
422
|
+
|
423
|
+
URI r = new URI(relativeURI);
|
424
|
+
String rscheme = r.getScheme();
|
425
|
+
String rauthority = r.getAuthority();
|
426
|
+
String rpath = r.getPath();
|
427
|
+
String rquery = r.getQuery();
|
428
|
+
|
429
|
+
String tscheme, tauthority, tpath, tquery;
|
430
|
+
if (rscheme != null && rscheme.equals(bscheme)) {
|
431
|
+
rscheme = null;
|
432
|
+
}
|
433
|
+
if (rscheme != null) {
|
434
|
+
tscheme = rscheme;
|
435
|
+
tauthority = rauthority;
|
436
|
+
tpath = removeDotSegments(rpath);
|
437
|
+
tquery = rquery;
|
438
|
+
} else {
|
439
|
+
if (rauthority != null) {
|
440
|
+
tauthority = rauthority;
|
441
|
+
tpath = removeDotSegments(rpath);
|
442
|
+
tquery = rquery;
|
443
|
+
} else {
|
444
|
+
if (rpath.length() == 0) {
|
445
|
+
tpath = bpath;
|
446
|
+
if (rquery != null) {
|
447
|
+
tquery = rquery;
|
448
|
+
} else {
|
449
|
+
tquery = bquery;
|
450
|
+
}
|
451
|
+
} else {
|
452
|
+
if (rpath.startsWith("/")) {
|
453
|
+
tpath = removeDotSegments(rpath);
|
454
|
+
} else {
|
455
|
+
if (bauthority != null && bpath.length() == 0) {
|
456
|
+
tpath = "/" + rpath;
|
457
|
+
} else {
|
458
|
+
int last = bpath.lastIndexOf('/');
|
459
|
+
if (last == -1) {
|
460
|
+
tpath = rpath;
|
461
|
+
} else {
|
462
|
+
tpath = bpath.substring(0, last+1) + rpath;
|
463
|
+
}
|
464
|
+
}
|
465
|
+
tpath = removeDotSegments(tpath);
|
466
|
+
}
|
467
|
+
tquery = rquery;
|
468
|
+
}
|
469
|
+
tauthority = bauthority;
|
470
|
+
}
|
471
|
+
tscheme = bscheme;
|
472
|
+
}
|
473
|
+
return new URI(tscheme, tauthority, tpath, tquery, null).toString();
|
474
|
+
}
|
475
|
+
|
476
|
+
private static String removeDotSegments(String path) {
|
477
|
+
|
478
|
+
// 1. The input buffer is initialized with the now-appended path
|
479
|
+
// components then replace occurrences of "//" in the input buffer
|
480
|
+
// with "/" until no more occurrences of "//" are in the input buffer.
|
481
|
+
String input = path;
|
482
|
+
while (input.indexOf("//") > -1) {
|
483
|
+
input = input.replaceAll("//", "/");
|
484
|
+
}
|
485
|
+
|
486
|
+
// Initialize the output buffer with the empty string.
|
487
|
+
StringBuilder output = new StringBuilder();
|
488
|
+
|
489
|
+
// If the input buffer starts with a root slash "/" then move this
|
490
|
+
// character to the output buffer.
|
491
|
+
if (input.charAt(0) == '/') {
|
492
|
+
output.append('/');
|
493
|
+
input = input.substring(1);
|
494
|
+
}
|
495
|
+
|
496
|
+
printStep("1 ", output, input);
|
497
|
+
|
498
|
+
// While the input buffer is not empty, loop as follows
|
499
|
+
while (input.length() != 0) {
|
500
|
+
// 2A. If the input buffer begins with a prefix of "./",
|
501
|
+
// then remove that prefix from the input buffer
|
502
|
+
// else if the input buffer begins with a prefix of "../", then
|
503
|
+
// if also the output does not contain the root slash "/" only,
|
504
|
+
// then move this prefix to the end of the output buffer else
|
505
|
+
// remove that prefix
|
506
|
+
if (input.startsWith("./")) {
|
507
|
+
input = input.substring(2);
|
508
|
+
printStep("2A", output, input);
|
509
|
+
} else if (input.startsWith("../")) {
|
510
|
+
input = input.substring(3);
|
511
|
+
if (!output.toString().equals("/")) {
|
512
|
+
output.append("../");
|
513
|
+
}
|
514
|
+
printStep("2A", output, input);
|
515
|
+
// 2B. if the input buffer begins with a prefix of "/./" or "/.",
|
516
|
+
// where "." is a complete path segment, then replace that prefix
|
517
|
+
// with "/" in the input buffer; otherwise,
|
518
|
+
} else if (input.startsWith("/./")) {
|
519
|
+
input = input.substring(2);
|
520
|
+
printStep("2B", output, input);
|
521
|
+
} else if (input.equals("/.")) {
|
522
|
+
// FIXME: what is complete path segment?
|
523
|
+
input = input.replaceFirst("/.", "/");
|
524
|
+
printStep("2B", output, input);
|
525
|
+
// 2C. if the input buffer begins with a prefix of "/../" or "/..",
|
526
|
+
// where ".." is a complete path segment, then replace that prefix
|
527
|
+
// with "/" in the input buffer and if also the output buffer is
|
528
|
+
// empty, last segment in the output buffer equals "../" or "..",
|
529
|
+
// where ".." is a complete path segment, then append ".." or "/.."
|
530
|
+
// for the latter case respectively to the output buffer else
|
531
|
+
// remove the last segment and its preceding "/" (if any) from the
|
532
|
+
// output buffer and if hereby the first character in the output
|
533
|
+
// buffer was removed and it was not the root slash then delete a
|
534
|
+
// leading slash from the input buffer; otherwise,
|
535
|
+
} else if (input.startsWith("/../")) {
|
536
|
+
input = input.substring(3);
|
537
|
+
if (output.length() == 0) {
|
538
|
+
output.append("/");
|
539
|
+
} else if (output.toString().endsWith("../")) {
|
540
|
+
output.append("..");
|
541
|
+
} else if (output.toString().endsWith("..")) {
|
542
|
+
output.append("/..");
|
543
|
+
} else {
|
544
|
+
int index = output.lastIndexOf("/");
|
545
|
+
if (index == -1) {
|
546
|
+
output = new StringBuilder();
|
547
|
+
if (input.charAt(0) == '/') {
|
548
|
+
input = input.substring(1);
|
549
|
+
}
|
550
|
+
} else {
|
551
|
+
output = output.delete(index, output.length());
|
552
|
+
}
|
553
|
+
}
|
554
|
+
printStep("2C", output, input);
|
555
|
+
} else if (input.equals("/..")) {
|
556
|
+
// FIXME: what is complete path segment?
|
557
|
+
input = input.replaceFirst("/..", "/");
|
558
|
+
if (output.length() == 0) {
|
559
|
+
output.append("/");
|
560
|
+
} else if (output.toString().endsWith("../")) {
|
561
|
+
output.append("..");
|
562
|
+
} else if (output.toString().endsWith("..")) {
|
563
|
+
output.append("/..");
|
564
|
+
} else {
|
565
|
+
int index = output.lastIndexOf("/");
|
566
|
+
if (index == -1) {
|
567
|
+
output = new StringBuilder();
|
568
|
+
if (input.charAt(0) == '/') {
|
569
|
+
input = input.substring(1);
|
570
|
+
}
|
571
|
+
} else {
|
572
|
+
output = output.delete(index, output.length());
|
573
|
+
}
|
574
|
+
}
|
575
|
+
printStep("2C", output, input);
|
576
|
+
// 2D. if the input buffer consists only of ".", then remove
|
577
|
+
// that from the input buffer else if the input buffer consists
|
578
|
+
// only of ".." and if the output buffer does not contain only
|
579
|
+
// the root slash "/", then move the ".." to the output buffer
|
580
|
+
// else delte it.; otherwise,
|
581
|
+
} else if (input.equals(".")) {
|
582
|
+
input = "";
|
583
|
+
printStep("2D", output, input);
|
584
|
+
} else if (input.equals("..")) {
|
585
|
+
if (!output.toString().equals("/")) {
|
586
|
+
output.append("..");
|
587
|
+
}
|
588
|
+
input = "";
|
589
|
+
printStep("2D", output, input);
|
590
|
+
// 2E. move the first path segment (if any) in the input buffer
|
591
|
+
// to the end of the output buffer, including the initial "/"
|
592
|
+
// character (if any) and any subsequent characters up to, but not
|
593
|
+
// including, the next "/" character or the end of the input buffer.
|
594
|
+
} else {
|
595
|
+
int end;
|
596
|
+
int begin = input.indexOf('/');
|
597
|
+
if (begin == 0) {
|
598
|
+
end = input.indexOf('/', 1);
|
599
|
+
} else {
|
600
|
+
end = begin;
|
601
|
+
begin = 0;
|
602
|
+
}
|
603
|
+
String segment;
|
604
|
+
if (end == -1) {
|
605
|
+
segment = input.substring(begin);
|
606
|
+
input = "";
|
607
|
+
} else {
|
608
|
+
segment = input.substring(begin, end);
|
609
|
+
input = input.substring(end);
|
610
|
+
}
|
611
|
+
output.append(segment);
|
612
|
+
printStep("2E", output, input);
|
613
|
+
}
|
614
|
+
}
|
615
|
+
|
616
|
+
// 3. Finally, if the only or last segment of the output buffer is
|
617
|
+
// "..", where ".." is a complete path segment not followed by a slash
|
618
|
+
// then append a slash "/". The output buffer is returned as the result
|
619
|
+
// of remove_dot_segments
|
620
|
+
if (output.toString().endsWith("..")) {
|
621
|
+
output.append('/');
|
622
|
+
printStep("3 ", output, input);
|
623
|
+
}
|
624
|
+
|
625
|
+
return output.toString();
|
626
|
+
}
|
627
|
+
|
628
|
+
private static void printStep(String step, StringBuilder output, String input) {
|
629
|
+
//if (System.getProperty("nokogiri.c14.debug") == "on") { //
|
630
|
+
// System.out.println(" " + step + ": " + output);
|
631
|
+
// if (output.length() == 0) {
|
632
|
+
// System.out.println("\t\t\t\t" + input);
|
633
|
+
// } else {
|
634
|
+
// System.out.println("\t\t\t" + input);
|
635
|
+
// }
|
636
|
+
//}
|
637
|
+
}
|
638
|
+
|
639
|
+
}
|
@@ -0,0 +1,38 @@
|
|
1
|
+
/**
|
2
|
+
* Licensed to the Apache Software Foundation (ASF) under one
|
3
|
+
* or more contributor license agreements. See the NOTICE file
|
4
|
+
* distributed with this work for additional information
|
5
|
+
* regarding copyright ownership. The ASF licenses this file
|
6
|
+
* to you under the Apache License, Version 2.0 (the
|
7
|
+
* "License"); you may not use this file except in compliance
|
8
|
+
* with the License. You may obtain a copy of the License at
|
9
|
+
*
|
10
|
+
* http://www.apache.org/licenses/LICENSE-2.0
|
11
|
+
*
|
12
|
+
* Unless required by applicable law or agreed to in writing,
|
13
|
+
* software distributed under the License is distributed on an
|
14
|
+
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
|
15
|
+
* KIND, either express or implied. See the License for the
|
16
|
+
* specific language governing permissions and limitations
|
17
|
+
* under the License.
|
18
|
+
*/
|
19
|
+
package nokogiri.internals.c14n;
|
20
|
+
|
21
|
+
|
22
|
+
/**
|
23
|
+
* @author Sean Mullan
|
24
|
+
*/
|
25
|
+
public class Canonicalizer11_OmitComments extends Canonicalizer11 {
|
26
|
+
|
27
|
+
public Canonicalizer11_OmitComments() {
|
28
|
+
super(false);
|
29
|
+
}
|
30
|
+
|
31
|
+
public final String engineGetURI() {
|
32
|
+
return Canonicalizer.ALGO_ID_C14N11_OMIT_COMMENTS;
|
33
|
+
}
|
34
|
+
|
35
|
+
public final boolean engineGetIncludeComments() {
|
36
|
+
return false;
|
37
|
+
}
|
38
|
+
}
|