RubyGems - nokogiri - Versions diffs - 1.0.0 → 1.6.8.1 - Mend

nokogiri 1.0.0 → 1.6.8.1

Potentially problematic release.

This version of nokogiri might be problematic. Click here for more details.

Files changed (309) hide show

checksums.yaml +7 -0
data/.autotest +26 -0
data/.cross_rubies +9 -0
data/.editorconfig +17 -0
data/.gemtest +0 -0
data/.travis.yml +51 -0
data/CHANGELOG.rdoc +1160 -0
data/CONTRIBUTING.md +42 -0
data/C_CODING_STYLE.rdoc +33 -0
data/Gemfile +22 -0
data/LICENSE.txt +31 -0
data/Manifest.txt +284 -40
data/README.md +166 -0
data/ROADMAP.md +111 -0
data/Rakefile +310 -199
data/STANDARD_RESPONSES.md +47 -0
data/Y_U_NO_GEMSPEC.md +155 -0
data/appveyor.yml +22 -0
data/bin/nokogiri +118 -0
data/build_all +45 -0
data/dependencies.yml +29 -0
data/ext/nokogiri/depend +358 -0
data/ext/nokogiri/extconf.rb +664 -34
data/ext/nokogiri/html_document.c +120 -33
data/ext/nokogiri/html_document.h +1 -1
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 +145 -0
data/ext/nokogiri/nokogiri.h +131 -0
data/ext/nokogiri/xml_attr.c +94 -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 +23 -19
data/ext/nokogiri/xml_cdata.h +1 -1
data/ext/nokogiri/xml_comment.c +69 -0
data/ext/nokogiri/xml_comment.h +9 -0
data/ext/nokogiri/xml_document.c +501 -54
data/ext/nokogiri/xml_document.h +14 -1
data/ext/nokogiri/xml_document_fragment.c +48 -0
data/ext/nokogiri/xml_document_fragment.h +10 -0
data/ext/nokogiri/xml_dtd.c +109 -24
data/ext/nokogiri/xml_dtd.h +3 -1
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 +60 -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 +117 -0
data/ext/nokogiri/xml_namespace.h +13 -0
data/ext/nokogiri/xml_node.c +1285 -315
data/ext/nokogiri/xml_node.h +4 -6
data/ext/nokogiri/xml_node_set.c +415 -54
data/ext/nokogiri/xml_node_set.h +6 -2
data/ext/nokogiri/xml_processing_instruction.c +56 -0
data/ext/nokogiri/xml_processing_instruction.h +9 -0
data/ext/nokogiri/xml_reader.c +316 -77
data/ext/nokogiri/xml_reader.h +1 -1
data/ext/nokogiri/xml_relax_ng.c +161 -0
data/ext/nokogiri/xml_relax_ng.h +9 -0
data/ext/nokogiri/xml_sax_parser.c +215 -80
data/ext/nokogiri/xml_sax_parser.h +30 -1
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 +115 -0
data/ext/nokogiri/xml_sax_push_parser.h +9 -0
data/ext/nokogiri/xml_schema.c +205 -0
data/ext/nokogiri/xml_schema.h +9 -0
data/ext/nokogiri/xml_syntax_error.c +45 -175
data/ext/nokogiri/xml_syntax_error.h +4 -2
data/ext/nokogiri/xml_text.c +37 -14
data/ext/nokogiri/xml_text.h +1 -1
data/ext/nokogiri/xml_xpath_context.c +230 -13
data/ext/nokogiri/xml_xpath_context.h +2 -1
data/ext/nokogiri/xslt_stylesheet.c +196 -34
data/ext/nokogiri/xslt_stylesheet.h +6 -1
data/lib/nokogiri/css/node.rb +18 -61
data/lib/nokogiri/css/parser.rb +725 -17
data/lib/nokogiri/css/parser.y +126 -63
data/lib/nokogiri/css/parser_extras.rb +91 -0
data/lib/nokogiri/css/syntax_error.rb +7 -0
data/lib/nokogiri/css/tokenizer.rb +148 -5
data/lib/nokogiri/css/tokenizer.rex +31 -39
data/lib/nokogiri/css/xpath_visitor.rb +109 -51
data/lib/nokogiri/css.rb +24 -3
data/lib/nokogiri/decorators/slop.rb +42 -0
data/lib/nokogiri/html/builder.rb +27 -1
data/lib/nokogiri/html/document.rb +329 -3
data/lib/nokogiri/html/document_fragment.rb +39 -0
data/lib/nokogiri/html/element_description.rb +23 -0
data/lib/nokogiri/html/element_description_defaults.rb +671 -0
data/lib/nokogiri/html/entity_lookup.rb +13 -0
data/lib/nokogiri/html/sax/parser.rb +35 -4
data/lib/nokogiri/html/sax/parser_context.rb +16 -0
data/lib/nokogiri/html/sax/push_parser.rb +36 -0
data/lib/nokogiri/html.rb +18 -76
data/lib/nokogiri/syntax_error.rb +4 -0
data/lib/nokogiri/version.rb +106 -1
data/lib/nokogiri/xml/attr.rb +14 -0
data/lib/nokogiri/xml/attribute_decl.rb +18 -0
data/lib/nokogiri/xml/builder.rb +395 -31
data/lib/nokogiri/xml/cdata.rb +4 -2
data/lib/nokogiri/xml/character_data.rb +7 -0
data/lib/nokogiri/xml/document.rb +267 -12
data/lib/nokogiri/xml/document_fragment.rb +149 -0
data/lib/nokogiri/xml/dtd.rb +27 -1
data/lib/nokogiri/xml/element_content.rb +36 -0
data/lib/nokogiri/xml/element_decl.rb +13 -0
data/lib/nokogiri/xml/entity_decl.rb +19 -0
data/lib/nokogiri/xml/namespace.rb +13 -0
data/lib/nokogiri/xml/node/save_options.rb +61 -0
data/lib/nokogiri/xml/node.rb +748 -109
data/lib/nokogiri/xml/node_set.rb +200 -72
data/lib/nokogiri/xml/parse_options.rb +120 -0
data/lib/nokogiri/xml/pp/character_data.rb +18 -0
data/lib/nokogiri/xml/pp/node.rb +56 -0
data/lib/nokogiri/xml/pp.rb +2 -0
data/lib/nokogiri/xml/processing_instruction.rb +8 -0
data/lib/nokogiri/xml/reader.rb +102 -4
data/lib/nokogiri/xml/relax_ng.rb +32 -0
data/lib/nokogiri/xml/sax/document.rb +114 -2
data/lib/nokogiri/xml/sax/parser.rb +97 -7
data/lib/nokogiri/xml/sax/parser_context.rb +16 -0
data/lib/nokogiri/xml/sax/push_parser.rb +60 -0
data/lib/nokogiri/xml/sax.rb +2 -7
data/lib/nokogiri/xml/schema.rb +63 -0
data/lib/nokogiri/xml/searchable.rb +221 -0
data/lib/nokogiri/xml/syntax_error.rb +27 -1
data/lib/nokogiri/xml/text.rb +4 -1
data/lib/nokogiri/xml/xpath/syntax_error.rb +11 -0
data/lib/nokogiri/xml/xpath.rb +4 -0
data/lib/nokogiri/xml/xpath_context.rb +3 -1
data/lib/nokogiri/xml.rb +45 -38
data/lib/nokogiri/xslt/stylesheet.rb +19 -0
data/lib/nokogiri/xslt.rb +47 -2
data/lib/nokogiri.rb +117 -24
data/lib/xsd/xmlparser/nokogiri.rb +102 -0
data/patches/sort-patches-by-date +25 -0
data/ports/archives/libxml2-2.9.4.tar.gz +0 -0
data/ports/archives/libxslt-1.1.29.tar.gz +0 -0
data/suppressions/README.txt +1 -0
data/suppressions/nokogiri_ree-1.8.7.358.supp +61 -0
data/suppressions/nokogiri_ruby-1.8.7.370.supp +0 -0
data/suppressions/nokogiri_ruby-1.9.2.320.supp +28 -0
data/suppressions/nokogiri_ruby-1.9.3.327.supp +28 -0
data/tasks/test.rb +100 -0
data/test/css/test_nthiness.rb +73 -6
data/test/css/test_parser.rb +184 -39
data/test/css/test_tokenizer.rb +72 -19
data/test/css/test_xpath_visitor.rb +44 -2
data/test/decorators/test_slop.rb +20 -0
data/test/files/2ch.html +108 -0
data/test/files/GH_1042.html +18 -0
data/test/files/address_book.rlx +12 -0
data/test/files/address_book.xml +10 -0
data/test/files/atom.xml +344 -0
data/test/files/bar/bar.xsd +4 -0
data/test/files/bogus.xml +0 -0
data/test/files/dont_hurt_em_why.xml +422 -0
data/test/files/encoding.html +82 -0
data/test/files/encoding.xhtml +84 -0
data/test/files/exslt.xml +8 -0
data/test/files/exslt.xslt +35 -0
data/test/files/foo/foo.xsd +4 -0
data/test/files/metacharset.html +10 -0
data/test/files/namespace_pressure_test.xml +1684 -0
data/test/files/noencoding.html +47 -0
data/test/files/po.xml +32 -0
data/test/files/po.xsd +66 -0
data/test/files/saml/saml20assertion_schema.xsd +283 -0
data/test/files/saml/saml20protocol_schema.xsd +302 -0
data/test/files/saml/xenc_schema.xsd +146 -0
data/test/files/saml/xmldsig_schema.xsd +318 -0
data/test/files/shift_jis.html +10 -0
data/test/files/shift_jis.xml +5 -0
data/test/files/shift_jis_no_charset.html +9 -0
data/test/files/slow-xpath.xml +25509 -0
data/test/files/snuggles.xml +3 -0
data/test/files/staff.dtd +10 -0
data/test/files/test_document_url/bar.xml +2 -0
data/test/files/test_document_url/document.dtd +4 -0
data/test/files/test_document_url/document.xml +6 -0
data/test/files/tlm.html +2 -1
data/test/files/to_be_xincluded.xml +2 -0
data/test/files/valid_bar.xml +2 -0
data/test/files/xinclude.xml +4 -0
data/test/helper.rb +124 -13
data/test/html/sax/test_parser.rb +118 -4
data/test/html/sax/test_parser_context.rb +46 -0
data/test/html/sax/test_push_parser.rb +87 -0
data/test/html/test_builder.rb +94 -8
data/test/html/test_document.rb +626 -11
data/test/html/test_document_encoding.rb +145 -0
data/test/html/test_document_fragment.rb +301 -0
data/test/html/test_element_description.rb +105 -0
data/test/html/test_named_characters.rb +14 -0
data/test/html/test_node.rb +212 -0
data/test/html/test_node_encoding.rb +85 -0
data/test/namespaces/test_additional_namespaces_in_builder_doc.rb +14 -0
data/test/namespaces/test_namespaces_aliased_default.rb +24 -0
data/test/namespaces/test_namespaces_in_builder_doc.rb +75 -0
data/test/namespaces/test_namespaces_in_cloned_doc.rb +31 -0
data/test/namespaces/test_namespaces_in_created_doc.rb +75 -0
data/test/namespaces/test_namespaces_in_parsed_doc.rb +80 -0
data/test/namespaces/test_namespaces_preservation.rb +31 -0
data/test/test_convert_xpath.rb +2 -47
data/test/test_css_cache.rb +45 -0
data/test/test_encoding_handler.rb +48 -0
data/test/test_memory_leak.rb +156 -0
data/test/test_nokogiri.rb +103 -1
data/test/test_soap4r_sax.rb +52 -0
data/test/test_xslt_transforms.rb +293 -8
data/test/xml/node/test_save_options.rb +28 -0
data/test/xml/node/test_subclass.rb +44 -0
data/test/xml/sax/test_parser.rb +309 -8
data/test/xml/sax/test_parser_context.rb +115 -0
data/test/xml/sax/test_push_parser.rb +157 -0
data/test/xml/test_attr.rb +67 -0
data/test/xml/test_attribute_decl.rb +86 -0
data/test/xml/test_builder.rb +327 -2
data/test/xml/test_c14n.rb +180 -0
data/test/xml/test_cdata.rb +32 -2
data/test/xml/test_comment.rb +40 -0
data/test/xml/test_document.rb +846 -35
data/test/xml/test_document_encoding.rb +31 -0
data/test/xml/test_document_fragment.rb +271 -0
data/test/xml/test_dtd.rb +153 -9
data/test/xml/test_dtd_encoding.rb +31 -0
data/test/xml/test_element_content.rb +56 -0
data/test/xml/test_element_decl.rb +73 -0
data/test/xml/test_entity_decl.rb +122 -0
data/test/xml/test_entity_reference.rb +251 -0
data/test/xml/test_namespace.rb +96 -0
data/test/xml/test_node.rb +1126 -105
data/test/xml/test_node_attributes.rb +115 -0
data/test/xml/test_node_encoding.rb +69 -0
data/test/xml/test_node_inheritance.rb +32 -0
data/test/xml/test_node_reparenting.rb +549 -0
data/test/xml/test_node_set.rb +668 -9
data/test/xml/test_parse_options.rb +64 -0
data/test/xml/test_processing_instruction.rb +30 -0
data/test/xml/test_reader.rb +589 -0
data/test/xml/test_reader_encoding.rb +134 -0
data/test/xml/test_relax_ng.rb +60 -0
data/test/xml/test_schema.rb +142 -0
data/test/xml/test_syntax_error.rb +30 -0
data/test/xml/test_text.rb +49 -2
data/test/xml/test_unparented_node.rb +440 -0
data/test/xml/test_xinclude.rb +83 -0
data/test/xml/test_xpath.rb +445 -0
data/test/xslt/test_custom_functions.rb +133 -0
data/test/xslt/test_exception_handling.rb +37 -0
data/test_all +107 -0
metadata +459 -115
data/History.txt +0 -6
data/README.ja.txt +0 -86
data/README.txt +0 -87
data/ext/nokogiri/html_sax_parser.c +0 -32
data/ext/nokogiri/html_sax_parser.h +0 -11
data/ext/nokogiri/native.c +0 -40
data/ext/nokogiri/native.h +0 -51
data/ext/nokogiri/xml_xpath.c +0 -46
data/ext/nokogiri/xml_xpath.h +0 -11
data/lib/nokogiri/css/generated_parser.rb +0 -653
data/lib/nokogiri/css/generated_tokenizer.rb +0 -159
data/lib/nokogiri/decorators/hpricot/node.rb +0 -58
data/lib/nokogiri/decorators/hpricot/node_set.rb +0 -14
data/lib/nokogiri/decorators/hpricot/xpath_visitor.rb +0 -17
data/lib/nokogiri/decorators/hpricot.rb +0 -3
data/lib/nokogiri/decorators.rb +0 -1
data/lib/nokogiri/hpricot.rb +0 -47
data/lib/nokogiri/xml/after_handler.rb +0 -18
data/lib/nokogiri/xml/before_handler.rb +0 -32
data/lib/nokogiri/xml/element.rb +0 -6
data/lib/nokogiri/xml/entity_declaration.rb +0 -9
data/nokogiri.gemspec +0 -34
data/test/hpricot/files/basic.xhtml +0 -17
data/test/hpricot/files/boingboing.html +0 -2266
data/test/hpricot/files/cy0.html +0 -3653
data/test/hpricot/files/immob.html +0 -400
data/test/hpricot/files/pace_application.html +0 -1320
data/test/hpricot/files/tenderlove.html +0 -16
data/test/hpricot/files/uswebgen.html +0 -220
data/test/hpricot/files/utf8.html +0 -1054
data/test/hpricot/files/week9.html +0 -1723
data/test/hpricot/files/why.xml +0 -19
data/test/hpricot/load_files.rb +0 -7
data/test/hpricot/test_alter.rb +0 -67
data/test/hpricot/test_builder.rb +0 -27
data/test/hpricot/test_parser.rb +0 -423
data/test/hpricot/test_paths.rb +0 -15
data/test/hpricot/test_preserved.rb +0 -78
data/test/hpricot/test_xml.rb +0 -30
data/test/test_reader.rb +0 -222

data/test/files/snuggles.xml ADDED Viewed

@@ -0,0 +1,3 @@
+<x xmlns:tenderlove='http://tenderlovemaking.com/'>
+  <tenderlove:foo awesome='true'>snuggles!</tenderlove:foo>
+</x>

data/test/files/staff.dtd ADDED Viewed

@@ -0,0 +1,10 @@
+<!ENTITY ent1 "es">
+<!ENTITY ent2 "1900 Dallas Road">
+<!ENTITY ent3 "Texas">
+<!ENTITY ent4 "<entElement domestic='Yes'>Element data</entElement><?PItarget PIdata?>">
+<!ENTITY ent5 PUBLIC "entityURI" "entityFile" NDATA notation1>
+<!ENTITY ent1 "This entity should be discarded">
+<!ELEMENT br EMPTY>
+<!ATTLIST br width CDATA "0">
+<!NOTATION notation1 PUBLIC "notation1File">
+<!NOTATION notation2 SYSTEM "notation2File">

data/test/files/test_document_url/bar.xml ADDED Viewed

	@@ -0,0 +1,2 @@
1	+ <?xml version="1.0" encoding="UTF-8" ?>
2	+ <blah>foobar</blah>

data/test/files/test_document_url/document.dtd ADDED Viewed

@@ -0,0 +1,4 @@
+<!ENTITY  bar SYSTEM "bar.xml">
+<!ELEMENT document (body)>
+<!ELEMENT blah ANY>
+<!ELEMENT body ANY>

data/test/files/test_document_url/document.xml ADDED Viewed

@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="UTF-8" ?>
+<!DOCTYPE document SYSTEM "document.dtd">
+<document>
+  <body>&bar;</body>
+</document>

data/test/files/tlm.html CHANGED Viewed

@@ -46,7 +46,7 @@
 .codesnip-container  {border:1px solid #ccc; background:#eee; padding: 5px;margin:10px;}
 </style>
 <link rel="EditURI" type="application/rsd+xml" title="RSD" href="http://tenderlovemaking.com/xmlrpc.php?rsd" />
-<link rel="wlwmanifest" type="application/wlwmanifest+xml" href="http://tenderlovemaking.com/wp-includes/wlwmanifest.xml" />
+<link rel="wlwmanifest" type="application/wlwmanifest+xml" href="http://tenderlovemaking.com/wp-includes/wlwmanifest.xml" />
 <meta name="generator" content="WordPress 2.6" />
 	<link rel="stylesheet" type="text/css" href="http://tenderlovemaking.com/wp-content/plugins/spell_checker/spell_checker.css" />
@@ -826,6 +826,7 @@ page.<span class="me1">body</span> =~ /&lt;textarea<span class="br0">&#91;</span
 </ul>
 </div>
+<div id="abc.123" class='special.character'>Special character div</div>
 <div id="footer">
 A design by <a href="http://blog.geminigeek.com/wordpress-theme">GeminiGeek</a> &bull; Powered by <a href="http://wordpress.org">Wordpress</a><!--&bull; <a href="#">CSS</a> &bull; <a href="#">xHTML 1.0</a>-->
 </div>

data/test/files/to_be_xincluded.xml ADDED Viewed

	@@ -0,0 +1,2 @@
1	+ <?xml version="1.0" encoding="utf-8"?>
2	+ <included>this snippet is to be included from xinclude.xml</included>

data/test/files/valid_bar.xml ADDED Viewed

	@@ -0,0 +1,2 @@
1	+ <?xml version="1.0" encoding="UTF-8"?>
2	+ <bar />

data/test/files/xinclude.xml ADDED Viewed

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="utf-8"?>
+<test xmlns:xi="http://www.w3.org/2001/XInclude">
+    <xi:include href="to_be_xincluded.xml"/>
+</test>

data/test/helper.rb CHANGED Viewed

@@ -1,23 +1,101 @@
-require 'test/unit'
+#Process.setrlimit(Process::RLIMIT_CORE, Process::RLIM_INFINITY) unless RUBY_PLATFORM =~ /(java|mswin|mingw)/i
+$VERBOSE = true
+require 'minitest/autorun'
+require 'minitest/pride'
+require 'fileutils'
+require 'tempfile'
+require 'pp'
-%w(../lib ../ext).each do |path|
-  $LOAD_PATH.unshift(File.expand_path(File.join(File.dirname(__FILE__), path)))
+require 'nokogiri'
+if ENV['TEST_NOKOGIRI_WITH_LIBXML_RUBY']
+  require 'libxml'
+  warn "#{__FILE__}:#{__LINE__}: loaded libxml-ruby '#{LibXML::XML::VERSION}'"
 end
-require 'nokogiri'
+warn "#{__FILE__}:#{__LINE__}: version info: #{Nokogiri::VERSION_INFO.inspect}"
 module Nokogiri
-  class TestCase < Test::Unit::TestCase
-    ASSETS_DIR = File.join(File.dirname(__FILE__), 'files')
-    XML_FILE = File.join(ASSETS_DIR, 'staff.xml')
-    XSLT_FILE = File.join(ASSETS_DIR, 'staff.xslt')
-    HTML_FILE = File.join(ASSETS_DIR, 'tlm.html')
-    undef :default_test
+  class TestCase < MiniTest::Spec
+    ASSETS_DIR          = File.expand_path File.join(File.dirname(__FILE__), 'files')
+    ADDRESS_SCHEMA_FILE = File.join(ASSETS_DIR, 'address_book.rlx')
+    ADDRESS_XML_FILE    = File.join(ASSETS_DIR, 'address_book.xml')
+    ENCODING_HTML_FILE  = File.join(ASSETS_DIR, 'encoding.html')
+    ENCODING_XHTML_FILE = File.join(ASSETS_DIR, 'encoding.xhtml')
+    EXML_FILE           = File.join(ASSETS_DIR, 'exslt.xml')
+    EXSLT_FILE          = File.join(ASSETS_DIR, 'exslt.xslt')
+    HTML_FILE           = File.join(ASSETS_DIR, 'tlm.html')
+    METACHARSET_FILE    = File.join(ASSETS_DIR, 'metacharset.html')
+    NICH_FILE           = File.join(ASSETS_DIR, '2ch.html')
+    NOENCODING_FILE     = File.join(ASSETS_DIR, 'noencoding.html')
+    PO_SCHEMA_FILE      = File.join(ASSETS_DIR, 'po.xsd')
+    PO_XML_FILE         = File.join(ASSETS_DIR, 'po.xml')
+    SHIFT_JIS_HTML      = File.join(ASSETS_DIR, 'shift_jis.html')
+    SHIFT_JIS_NO_CHARSET= File.join(ASSETS_DIR, 'shift_jis_no_charset.html')
+    SHIFT_JIS_XML       = File.join(ASSETS_DIR, 'shift_jis.xml')
+    SNUGGLES_FILE       = File.join(ASSETS_DIR, 'snuggles.xml')
+    XML_FILE            = File.join(ASSETS_DIR, 'staff.xml')
+    XML_XINCLUDE_FILE   = File.join(ASSETS_DIR, 'xinclude.xml')
+    XML_ATOM_FILE       = File.join(ASSETS_DIR, 'atom.xml')
+    XSLT_FILE           = File.join(ASSETS_DIR, 'staff.xslt')
+    XPATH_FILE          = File.join(ASSETS_DIR, 'slow-xpath.xml')
     def teardown
-      GC.start if ENV['NOKOGIRI_GC']
+      if ENV['NOKOGIRI_GC']
+        STDOUT.putc '!'
+        if RUBY_PLATFORM =~ /java/
+          require 'java'
+          java.lang.System.gc
+        else
+          GC.start
+        end
+      end
+    end
+    def stress_memory_while &block
+      # force the test to explicitly declare a skip
+      raise "JRuby doesn't do GC" if Nokogiri.jruby?
+      old_stress = GC.stress
+      begin
+        GC.stress = true
+        yield
+      ensure
+        GC.stress = old_stress
+      end
     end
+    def assert_indent amount, doc, message = nil
+      nodes = []
+      doc.traverse do |node|
+        nodes << node if node.text? && node.blank?
+      end
+      assert nodes.length > 0
+      nodes.each do |node|
+        len = node.content.gsub(/[\r\n]/, '').length
+        assert_equal(0, len % amount, message)
+      end
+    end
+    def util_decorate(document, decorator_module)
+      document.decorators(XML::Node) << decorator_module
+      document.decorators(XML::NodeSet) << decorator_module
+      document.decorate!
+    end
+    #
+    #  Test::Unit backwards compatibility section
+    #
+    alias :assert_no_match      :refute_match
+    alias :assert_not_nil       :refute_nil
+    alias :assert_raise         :assert_raises
+    alias :assert_not_equal     :refute_equal
+    def assert_not_send send_ary, m = nil
+      recv, msg, *args = send_ary
+      m = message(m) {
+        "Expected #{mu_pp(recv)}.#{msg}(*#{mu_pp(args)}) to return false" }
+      assert !recv.__send__(msg, *args), m
+    end unless method_defined?(:assert_not_send)
   end
   module SAX
@@ -25,7 +103,15 @@ module Nokogiri
       class Doc < XML::SAX::Document
         attr_reader :start_elements, :start_document_called
         attr_reader :end_elements, :end_document_called
-        attr_reader :data, :comments, :cdata_blocks
+        attr_reader :data, :comments, :cdata_blocks, :start_elements_namespace
+        attr_reader :errors, :warnings, :end_elements_namespace
+        attr_reader :xmldecls
+        attr_reader :processing_instructions
+        def xmldecl version, encoding, standalone
+          @xmldecls = [version, encoding, standalone].compact
+          super
+        end
         def start_document
           @start_document_called = true
@@ -37,16 +123,36 @@ module Nokogiri
           super
         end
+        def error error
+          (@errors ||= []) << error
+          super
+        end
+        def warning warning
+          (@warning ||= []) << warning
+          super
+        end
         def start_element *args
           (@start_elements ||= []) << args
           super
         end
+        def start_element_namespace *args
+          (@start_elements_namespace ||= []) << args
+          super
+        end
         def end_element *args
           (@end_elements ||= []) << args
           super
         end
+        def end_element_namespace *args
+          (@end_elements_namespace ||= []) << args
+          super
+        end
         def characters string
           @data ||= []
           @data += [string]
@@ -64,6 +170,11 @@ module Nokogiri
           @cdata_blocks += [string]
           super
         end
+        def processing_instruction name, content
+          @processing_instructions ||= []
+          @processing_instructions << [name, content]
+        end
       end
     end
   end

data/test/html/sax/test_parser.rb CHANGED Viewed

@@ -1,16 +1,71 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', '..', "helper"))
+# -*- coding: utf-8 -*-
+require "helper"
 module Nokogiri
   module HTML
     module SAX
       class TestParser < Nokogiri::SAX::TestCase
         def setup
+          super
           @parser = HTML::SAX::Parser.new(Doc.new)
         end
+        def test_parse_empty_document
+          # This caused a segfault in libxml 2.6.x
+          assert_nil @parser.parse ''
+        end
+        def test_parse_empty_file
+          # Make sure empty files don't break stuff
+          empty_file_name =  File.join(ASSETS_DIR, 'bogus.xml')
+          # assert_nothing_raised do
+            @parser.parse_file empty_file_name
+          # end
+        end
         def test_parse_file
           @parser.parse_file(HTML_FILE)
-          assert_equal 1110, @parser.document.end_elements.length
+          # Take a look at the comment in test_parse_document to know
+          # a possible reason to this difference.
+          if Nokogiri.uses_libxml?
+            assert_equal 1111, @parser.document.end_elements.length
+          else
+            assert_equal 1120, @parser.document.end_elements.length
+          end
+        end
+        def test_parse_file_nil_argument
+          assert_raises(ArgumentError) {
+            @parser.parse_file(nil)
+          }
+        end
+        def test_parse_file_non_existant
+          assert_raise Errno::ENOENT do
+            @parser.parse_file('there_is_no_reasonable_way_this_file_exists')
+          end
+        end
+        def test_parse_file_with_dir
+          assert_raise Errno::EISDIR do
+            @parser.parse_file(File.dirname(__FILE__))
+          end
+        end
+        def test_parse_memory_nil
+          assert_raise ArgumentError do
+            @parser.parse_memory(nil)
+          end
+        end
+        def test_parse_force_encoding
+          @parser.parse_memory(<<-HTML, 'UTF-8')
+          <meta http-equiv="Content-Type" content="text/html; charset=windows-1251">
+          Информация
+          HTML
+          assert_equal("Информация",
+                       @parser.document.data.join.strip)
         end
         def test_parse_document
@@ -18,8 +73,67 @@ module Nokogiri
             <p>Paragraph 1</p>
             <p>Paragraph 2</p>
           eoxml
-          assert_equal([["html", []], ["body", []], ["p", []], ["p", []]],
-                       @parser.document.start_elements)
+          # JRuby version is different because of the internal implementation
+          # JRuby version uses NekoHTML which inserts empty "head" elements.
+          #
+          # Currently following features are set:
+          # "http://cyberneko.org/html/properties/names/elems" => "lower"
+          # "http://cyberneko.org/html/properties/names/attrs" => "lower"
+          if Nokogiri.uses_libxml?
+            assert_equal([["html", []], ["body", []], ["p", []], ["p", []]],
+                         @parser.document.start_elements)
+          else
+            assert_equal([["html", []], ["head", []], ["body", []], ["p", []], ["p", []]],
+                         @parser.document.start_elements)
+          end
+        end
+        def test_parser_attributes
+          html = <<-eohtml
+<html>
+  <head>
+    <title>hello</title>
+  </head>
+<body>
+  <img src="face.jpg" title="daddy &amp; me">
+  <hr noshade size="2">
+</body>
+</html>
+          eohtml
+          block_called = false
+          @parser.parse(html) { |ctx|
+            block_called = true
+            ctx.replace_entities = true
+          }
+          assert block_called
+          noshade_value = if Nokogiri.uses_libxml? && Nokogiri::VERSION_INFO['libxml']['loaded'] < '2.7.7'
+                            ['noshade', 'noshade']
+                          else
+                            ['noshade', nil]
+                          end
+          assert_equal [
+            ['html', []],
+            ['head', []],
+            ['title', []],
+            ['body', []],
+            ['img', [
+                ['src', 'face.jpg'],
+                ['title', 'daddy & me']
+              ]],
+            ['hr', [
+                noshade_value,
+                ['size', '2']
+              ]]
+          ], @parser.document.start_elements
+        end
+        def test_empty_processing_instruction
+          @parser.parse_memory("<strong>this will segfault<?strong>")
         end
       end
     end

data/test/html/sax/test_parser_context.rb ADDED Viewed

@@ -0,0 +1,46 @@
+# -*- coding: utf-8 -*-
+require "helper"
+module Nokogiri
+  module HTML
+    module SAX
+      class TestParserContext < Nokogiri::SAX::TestCase
+        def test_from_io
+          ctx = ParserContext.new StringIO.new('fo'), 'UTF-8'
+          assert ctx
+        end
+        def test_from_string
+          ctx = ParserContext.new 'blah blah'
+          assert ctx
+        end
+        def test_parse_with
+          ctx = ParserContext.new 'blah'
+          assert_raises ArgumentError do
+            ctx.parse_with nil
+          end
+        end
+        def test_parse_with_sax_parser
+          # assert_nothing_raised do
+            xml = "<root />"
+            ctx = ParserContext.new xml
+            parser = Parser.new Doc.new
+            ctx.parse_with parser
+          # end
+        end
+        def test_from_file
+          # assert_nothing_raised do
+            ctx = ParserContext.file HTML_FILE, 'UTF-8'
+            parser = Parser.new Doc.new
+            ctx.parse_with parser
+          # end
+        end
+      end
+    end
+  end
+end

data/test/html/sax/test_push_parser.rb ADDED Viewed

@@ -0,0 +1,87 @@
+# -*- coding: utf-8 -*-
+require "helper"
+module Nokogiri
+  module HTML
+    module SAX
+      class TestPushParser < Nokogiri::SAX::TestCase
+        def setup
+          super
+          @parser = HTML::SAX::PushParser.new(Doc.new)
+        end
+        def test_end_document_called
+          @parser.<<(<<-eoxml)
+            <p id="asdfasdf">
+              <!-- This is a comment -->
+              Paragraph 1
+            </p>
+          eoxml
+          assert ! @parser.document.end_document_called
+          @parser.finish
+          assert @parser.document.end_document_called
+        end
+        def test_start_element
+          @parser.<<(<<-eoxml)
+            <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
+            <html><head><body><p id="asdfasdf">
+          eoxml
+          assert_equal [["html", []], ["head", []], ["body", []], ["p", [["id", "asdfasdf"]]]],
+            @parser.document.start_elements
+          @parser.<<(<<-eoxml)
+              <!-- This is a comment -->
+              Paragraph 1
+            </p></body></html>
+          eoxml
+          assert_equal [' This is a comment '], @parser.document.comments
+          @parser.finish
+        end
+        def test_chevron_partial_html
+          @parser.<<(<<-eoxml)
+            <p id="asdfasdf">
+          eoxml
+          @parser.<<(<<-eoxml)
+              <!-- This is a comment -->
+              Paragraph 1
+            </p>
+          eoxml
+          assert_equal [' This is a comment '], @parser.document.comments
+          @parser.finish
+        end
+        def test_chevron
+          @parser.<<(<<-eoxml)
+            <p id="asdfasdf">
+              <!-- This is a comment -->
+              Paragraph 1
+            </p>
+          eoxml
+          @parser.finish
+          assert_equal [' This is a comment '], @parser.document.comments
+        end
+        def test_default_options
+          assert_equal 0, @parser.options
+        end
+        def test_broken_encoding
+          skip("ultra hard to fix for pure Java version") if Nokogiri.jruby?
+          @parser.options |= XML::ParseOptions::RECOVER
+          # This is ISO_8859-1:
+          @parser.<< "<?xml version='1.0' encoding='UTF-8'?><r>Gau\337</r>"
+          @parser.finish
+          assert(@parser.document.errors.size >= 1)
+          assert_equal "Gau\337", @parser.document.data.join
+          assert_equal [["r"], ["body"], ["html"]], @parser.document.end_elements
+        end
+      end
+    end
+  end
+end

data/test/html/test_builder.rb CHANGED Viewed

@@ -1,8 +1,35 @@
-require File.expand_path(File.join(File.dirname(__FILE__), '..', "helper"))
+require "helper"
 module Nokogiri
   module HTML
     class TestBuilder < Nokogiri::TestCase
+      def test_top_level_function_builds
+        foo = nil
+        Nokogiri() { |xml| foo = xml }
+        assert_instance_of Nokogiri::HTML::Builder, foo
+      end
+      def test_builder_with_explicit_tags
+        html_doc = Nokogiri::HTML::Builder.new {
+          div.slide(:class => 'another_class') {
+            node = Nokogiri::XML::Node.new("id", doc)
+            node.content = "hello"
+            insert(node)
+          }
+        }.doc
+        assert_equal 1, html_doc.css('div.slide > id').length
+        assert_equal 'hello', html_doc.at('div.slide > id').content
+      end
+      def test_hash_as_attributes_for_attribute_method
+        html = Nokogiri::HTML::Builder.new { ||
+          div.slide(:class => 'another_class') {
+            span 'Slide 1'
+          }
+        }.to_html
+        assert_match 'class="slide another_class"', html
+      end
       def test_hash_as_attributes
         builder = Nokogiri::HTML::Builder.new do
           div(:id => 'awesome') {
@@ -10,7 +37,32 @@ module Nokogiri
           }
         end
         assert_equal('<div id="awesome"><h1>america</h1></div>',
-                     builder.to_html.gsub(/\n/, ''))
+                     builder.doc.root.to_html.gsub(/\n/, '').gsub(/>\s*</, '><'))
+      end
+      def test_href_with_attributes
+        uri = 'http://tenderlovemaking.com/'
+        built = Nokogiri::XML::Builder.new {
+          div {
+            a('King Khan & The Shrines', :href => uri)
+          }
+        }
+        assert_equal 'http://tenderlovemaking.com/',
+          built.doc.at('a')[:href]
+      end
+      def test_tag_nesting
+        builder = Nokogiri::HTML::Builder.new do
+          body {
+            span.left ''
+            span.middle {
+              div.icon ''
+            }
+            span.right ''
+          }
+        end
+        assert node = builder.doc.css('span.right').first
+        assert_equal 'middle', node.previous_sibling['class']
       end
       def test_has_ampersand
@@ -22,7 +74,7 @@ module Nokogiri
         end
         assert_equal(
           '<div class="rad" id="thing">&lt;awe&amp;some&gt;<b>hello &amp; world</b></div>',
-                     builder.to_html.gsub(/\n/, ''))
+                     builder.doc.root.to_html.gsub(/\n/, ''))
       end
       def test_multi_tags
@@ -34,7 +86,7 @@ module Nokogiri
         end
         assert_equal(
           '<div class="rad" id="thing">&lt;awesome&gt;<b>hello</b></div>',
-                     builder.doc.to_html.gsub(/\n/, ''))
+                     builder.doc.root.to_html.gsub(/\n/, ''))
       end
       def test_attributes_plus_block
@@ -44,7 +96,7 @@ module Nokogiri
           }
         end
         assert_equal('<div class="rad" id="thing">&lt;awesome&gt;</div>',
-                     builder.doc.to_html.chomp)
+                     builder.doc.root.to_html.chomp)
       end
       def test_builder_adds_attributes
@@ -52,14 +104,14 @@ module Nokogiri
           div.rad.thing! "tender div"
         end
         assert_equal('<div class="rad" id="thing">tender div</div>',
-                     builder.doc.to_html.chomp)
+                     builder.doc.root.to_html.chomp)
       end
       def test_bold_tag
         builder = Nokogiri::HTML::Builder.new do
           b "bold tag"
         end
-        assert_equal('<b>bold tag</b>', builder.doc.to_html.chomp)
+        assert_equal('<b>bold tag</b>', builder.doc.root.to_html.chomp)
       end
       def test_html_then_body_tag
@@ -71,7 +123,41 @@ module Nokogiri
           }
         end
         assert_equal('<html><body><b>bold tag</b></body></html>',
-                     builder.doc.to_html.chomp)
+                     builder.doc.root.to_html.chomp.gsub(/>\s*</, '><'))
+      end
+      def test_instance_eval_with_delegation_to_block_context
+        class << self
+          def foo
+            "foo!"
+          end
+        end
+        builder = Nokogiri::HTML::Builder.new { text foo }
+        assert builder.to_html.include?("foo!")
+      end
+      def test_builder_with_param
+        doc = Nokogiri::HTML::Builder.new { |html|
+          html.body {
+            html.p "hello world"
+          }
+        }.doc
+        assert node = doc.xpath('//body/p').first
+        assert_equal 'hello world', node.content
+      end
+      def test_builder_with_id
+        text = "hello world"
+        doc = Nokogiri::HTML::Builder.new { |html|
+          html.body {
+            html.id_ text
+          }
+        }.doc
+        assert node = doc.xpath('//body/id').first
+        assert_equal text, node.content
       end
     end
   end