hpricot 0.8.2-java

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (52) hide show
  1. data/CHANGELOG +88 -0
  2. data/COPYING +18 -0
  3. data/README +275 -0
  4. data/Rakefile +272 -0
  5. data/ext/fast_xs/FastXsService.java +1030 -0
  6. data/ext/fast_xs/extconf.rb +4 -0
  7. data/ext/fast_xs/fast_xs.c +201 -0
  8. data/ext/hpricot_scan/HpricotCss.java +831 -0
  9. data/ext/hpricot_scan/HpricotScanService.java +2086 -0
  10. data/ext/hpricot_scan/extconf.rb +6 -0
  11. data/ext/hpricot_scan/hpricot_common.rl +76 -0
  12. data/ext/hpricot_scan/hpricot_css.c +3503 -0
  13. data/ext/hpricot_scan/hpricot_css.java.rl +155 -0
  14. data/ext/hpricot_scan/hpricot_css.rl +115 -0
  15. data/ext/hpricot_scan/hpricot_scan.c +6927 -0
  16. data/ext/hpricot_scan/hpricot_scan.h +79 -0
  17. data/ext/hpricot_scan/hpricot_scan.java.rl +1152 -0
  18. data/ext/hpricot_scan/hpricot_scan.rl +788 -0
  19. data/extras/mingw-rbconfig.rb +176 -0
  20. data/lib/fast_xs.jar +0 -0
  21. data/lib/hpricot.rb +26 -0
  22. data/lib/hpricot/blankslate.rb +63 -0
  23. data/lib/hpricot/builder.rb +216 -0
  24. data/lib/hpricot/elements.rb +510 -0
  25. data/lib/hpricot/htmlinfo.rb +691 -0
  26. data/lib/hpricot/inspect.rb +103 -0
  27. data/lib/hpricot/modules.rb +40 -0
  28. data/lib/hpricot/parse.rb +38 -0
  29. data/lib/hpricot/tag.rb +219 -0
  30. data/lib/hpricot/tags.rb +164 -0
  31. data/lib/hpricot/traverse.rb +839 -0
  32. data/lib/hpricot/xchar.rb +94 -0
  33. data/lib/hpricot_scan.jar +0 -0
  34. data/test/files/basic.xhtml +17 -0
  35. data/test/files/boingboing.html +2266 -0
  36. data/test/files/cy0.html +3653 -0
  37. data/test/files/immob.html +400 -0
  38. data/test/files/pace_application.html +1320 -0
  39. data/test/files/tenderlove.html +16 -0
  40. data/test/files/uswebgen.html +220 -0
  41. data/test/files/utf8.html +1054 -0
  42. data/test/files/week9.html +1723 -0
  43. data/test/files/why.xml +19 -0
  44. data/test/load_files.rb +7 -0
  45. data/test/nokogiri-bench.rb +64 -0
  46. data/test/test_alter.rb +96 -0
  47. data/test/test_builder.rb +37 -0
  48. data/test/test_parser.rb +428 -0
  49. data/test/test_paths.rb +25 -0
  50. data/test/test_preserved.rb +88 -0
  51. data/test/test_xml.rb +28 -0
  52. metadata +112 -0
data/test/test_paths.rb ADDED
@@ -0,0 +1,25 @@
1
+ #!/usr/bin/env ruby
2
+
3
+ require 'test/unit'
4
+ require 'hpricot'
5
+ require 'load_files'
6
+
7
+ class TestParser < Test::Unit::TestCase
8
+ def test_roundtrip
9
+ @basic = Hpricot.parse(TestFiles::BASIC)
10
+ %w[link link[2] body #link1 a p.ohmy].each do |css_sel|
11
+ ele = @basic.at(css_sel)
12
+ assert_equal ele, @basic.at(ele.css_path)
13
+ assert_equal ele, @basic.at(ele.xpath)
14
+ end
15
+ end
16
+ def test_attr_brackets
17
+ doc = Hpricot('<input name="vendor[porkpies]"/>')
18
+ assert_equal 1, (doc/'input[@name^="vendor[porkpies]"]').length
19
+ assert_equal 1, (doc/'input[@name="vendor[porkpies]"]').length
20
+ assert_equal 0, (doc/'input[@name$="]]]]]"]').length
21
+
22
+ doc = Hpricot('<input name="vendor[porkpies][meaty]"/>')
23
+ assert_equal 1, (doc/'input[@name^="vendor[porkpies][meaty]"]').length
24
+ end
25
+ end
data/test/test_preserved.rb ADDED
@@ -0,0 +1,88 @@
1
+ # -*- coding: utf-8 -*-
2
+ #!/usr/bin/env ruby
3
+
4
+ require 'test/unit'
5
+ require 'hpricot'
6
+ require 'load_files'
7
+
8
+ unless "".respond_to?(:lines)
9
+ require 'enumerator'
10
+ class String
11
+ def lines
12
+ Enumerable::Enumerator.new(self, :each_line)
13
+ end
14
+ end
15
+ end
16
+
17
+ class TestPreserved < Test::Unit::TestCase
18
+ def assert_roundtrip str
19
+ doc = Hpricot(str)
20
+ yield doc if block_given?
21
+ str2 = doc.to_original_html
22
+ if RUBY_VERSION =~ /^1.9/
23
+ str2.force_encoding('UTF-8')
24
+ end
25
+ str.lines.zip(str2.lines).each do |s1, s2|
26
+ assert_equal s1, s2
27
+ end
28
+ end
29
+
30
+ def assert_html str1, str2
31
+ doc = Hpricot(str2)
32
+ yield doc if block_given?
33
+ assert_equal str1, doc.to_original_html
34
+ end
35
+
36
+ def test_simple
37
+ str = "<p>Hpricot is a <b>you know <i>uh</b> fine thing.</p>"
38
+ assert_html str, str
39
+ assert_html "<p class=\"new\">Hpricot is a <b>you know <i>uh</b> fine thing.</p>", str do |doc|
40
+ (doc/:p).set('class', 'new')
41
+ end
42
+ end
43
+
44
+ def test_parent
45
+ str = "<html><base href='/'><head><title>Test</title></head><body><div id='wrap'><p>Paragraph one.</p><p>Paragraph two.</p></div></body></html>"
46
+ assert_html str, str
47
+ assert_html "<html><base href='/'><body><div id=\"all\"><div><p>Paragraph one.</p></div><div><p>Paragraph two.</p></div></div></body></html>", str do |doc|
48
+ (doc/:head).remove
49
+ (doc/:div).set('id', 'all')
50
+ (doc/:p).wrap('<div></div>')
51
+ end
52
+ end
53
+
54
+ def test_escaping_of_contents
55
+ doc = Hpricot(TestFiles::BOINGBOING)
56
+ assert_equal "Fukuda’s Automatic Door opens around your body as you pass through it. The idea is to save energy and keep the room clean.", doc.at("img[@alt='200606131240']").next.to_s.strip
57
+ end
58
+
59
+ def test_files
60
+ assert_roundtrip TestFiles::BASIC
61
+ assert_roundtrip TestFiles::BOINGBOING
62
+ assert_roundtrip TestFiles::CY0
63
+ end
64
+
65
+ def test_fixup_link
66
+ doc = %{<?xml version="1.0" encoding="UTF-8"?><rss><channel><link>ht</link></channel></rss>}
67
+ assert_roundtrip doc
68
+ assert_equal Hpricot(doc).to_s,
69
+ %{<?xml version="1.0" encoding="UTF-8"?><rss><channel><link />ht</channel></rss>}
70
+ assert_equal Hpricot.XML(doc).to_s,
71
+ %{<?xml version="1.0" encoding="UTF-8"?><rss><channel><link>ht</link></channel></rss>}
72
+ end
73
+
74
+ def test_escaping_of_attrs
75
+ # ampersands in URLs
76
+ str = %{<a href="http://google.com/search?q=hpricot&amp;l=en">Google</a>}
77
+ link = (doc = Hpricot(str)).at(:a)
78
+ assert_equal "http://google.com/search?q=hpricot&l=en", link['href']
79
+ assert_equal "http://google.com/search?q=hpricot&l=en", link.attributes['href']
80
+ assert_equal "http://google.com/search?q=hpricot&l=en", link.get_attribute('href')
81
+ assert_equal "http://google.com/search?q=hpricot&amp;l=en", link.raw_attributes['href']
82
+ assert_equal str, doc.to_html
83
+
84
+ # alter the url
85
+ link['href'] = "javascript:alert(\"AGGA-KA-BOO!\")"
86
+ assert_equal %{<a href="javascript:alert(&quot;AGGA-KA-BOO!&quot;)">Google</a>}, doc.to_html
87
+ end
88
+ end
data/test/test_xml.rb ADDED
@@ -0,0 +1,28 @@
1
+ #!/usr/bin/env ruby
2
+
3
+ require 'test/unit'
4
+ require 'hpricot'
5
+ require 'load_files'
6
+
7
+ class TestParser < Test::Unit::TestCase
8
+ # normally, the link tags are empty HTML tags.
9
+ # contributed by laudney.
10
+ def test_normally_empty
11
+ doc = Hpricot::XML("<rss><channel><title>this is title</title><link>http://fake.com</link></channel></rss>")
12
+ assert_equal "this is title", (doc/:rss/:channel/:title).text
13
+ assert_equal "http://fake.com", (doc/:rss/:channel/:link).text
14
+ end
15
+
16
+ # make sure XML doesn't get downcased
17
+ def test_casing
18
+ doc = Hpricot::XML(TestFiles::WHY)
19
+ assert_equal "hourly", (doc.at "sy:updatePeriod").inner_html
20
+ assert_equal 1, (doc/"guid[@isPermaLink]").length
21
+ end
22
+
23
+ # be sure tags named "text" are ok
24
+ def test_text_tags
25
+ doc = Hpricot::XML("<feed><title>City Poisoned</title><text>Rita Lee has poisoned Brazil.</text></feed>")
26
+ assert_equal "City Poisoned", (doc/"title").text
27
+ end
28
+ end
metadata ADDED
@@ -0,0 +1,112 @@
1
+ --- !ruby/object:Gem::Specification
2
+ name: hpricot
3
+ version: !ruby/object:Gem::Version
4
+ version: 0.8.2
5
+ platform: java
6
+ authors:
7
+ - why the lucky stiff
8
+ autorequire:
9
+ bindir: bin
10
+ cert_chain: []
11
+
12
+ date: 2009-11-05 00:00:00 -06:00
13
+ default_executable:
14
+ dependencies: []
15
+
16
+ description: a swift, liberal HTML parser with a fantastic library
17
+ email: why@ruby-lang.org
18
+ executables: []
19
+
20
+ extensions: []
21
+
22
+ extra_rdoc_files:
23
+ - README
24
+ - CHANGELOG
25
+ - COPYING
26
+ files:
27
+ - CHANGELOG
28
+ - COPYING
29
+ - README
30
+ - Rakefile
31
+ - test/files/basic.xhtml
32
+ - test/files/boingboing.html
33
+ - test/files/cy0.html
34
+ - test/files/immob.html
35
+ - test/files/pace_application.html
36
+ - test/files/tenderlove.html
37
+ - test/files/uswebgen.html
38
+ - test/files/utf8.html
39
+ - test/files/week9.html
40
+ - test/files/why.xml
41
+ - test/load_files.rb
42
+ - test/nokogiri-bench.rb
43
+ - test/test_alter.rb
44
+ - test/test_builder.rb
45
+ - test/test_parser.rb
46
+ - test/test_paths.rb
47
+ - test/test_preserved.rb
48
+ - test/test_xml.rb
49
+ - lib/hpricot/blankslate.rb
50
+ - lib/hpricot/builder.rb
51
+ - lib/hpricot/elements.rb
52
+ - lib/hpricot/htmlinfo.rb
53
+ - lib/hpricot/inspect.rb
54
+ - lib/hpricot/modules.rb
55
+ - lib/hpricot/parse.rb
56
+ - lib/hpricot/tag.rb
57
+ - lib/hpricot/tags.rb
58
+ - lib/hpricot/traverse.rb
59
+ - lib/hpricot/xchar.rb
60
+ - lib/hpricot.rb
61
+ - extras/mingw-rbconfig.rb
62
+ - ext/hpricot_scan/hpricot_scan.h
63
+ - ext/fast_xs/FastXsService.java
64
+ - ext/hpricot_scan/HpricotCss.java
65
+ - ext/hpricot_scan/HpricotScanService.java
66
+ - ext/fast_xs/fast_xs.c
67
+ - ext/hpricot_scan/hpricot_css.c
68
+ - ext/hpricot_scan/hpricot_scan.c
69
+ - ext/fast_xs/extconf.rb
70
+ - ext/hpricot_scan/extconf.rb
71
+ - ext/hpricot_scan/hpricot_common.rl
72
+ - ext/hpricot_scan/hpricot_css.java.rl
73
+ - ext/hpricot_scan/hpricot_css.rl
74
+ - ext/hpricot_scan/hpricot_scan.java.rl
75
+ - ext/hpricot_scan/hpricot_scan.rl
76
+ - lib/hpricot_scan.jar
77
+ - lib/fast_xs.jar
78
+ has_rdoc: true
79
+ homepage: http://code.whytheluckystiff.net/hpricot/
80
+ licenses: []
81
+
82
+ post_install_message:
83
+ rdoc_options:
84
+ - --quiet
85
+ - --title
86
+ - The Hpricot Reference
87
+ - --main
88
+ - README
89
+ - --inline-source
90
+ require_paths:
91
+ - lib
92
+ required_ruby_version: !ruby/object:Gem::Requirement
93
+ requirements:
94
+ - - ">="
95
+ - !ruby/object:Gem::Version
96
+ version: "0"
97
+ version:
98
+ required_rubygems_version: !ruby/object:Gem::Requirement
99
+ requirements:
100
+ - - ">="
101
+ - !ruby/object:Gem::Version
102
+ version: "0"
103
+ version:
104
+ requirements: []
105
+
106
+ rubyforge_project: hobix
107
+ rubygems_version: 1.3.5
108
+ signing_key:
109
+ specification_version: 3
110
+ summary: a swift, liberal HTML parser with a fantastic library
111
+ test_files: []
112
+