tidy-ext 0.1.7

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 (65) hide show
  1. data/.gitignore +4 -0
  2. data/LICENSE +50 -0
  3. data/README +12 -0
  4. data/Rakefile +60 -0
  5. data/VERSION +1 -0
  6. data/ext/tidy/access.c +3310 -0
  7. data/ext/tidy/access.h +279 -0
  8. data/ext/tidy/alloc.c +107 -0
  9. data/ext/tidy/attrask.c +209 -0
  10. data/ext/tidy/attrdict.c +2398 -0
  11. data/ext/tidy/attrdict.h +122 -0
  12. data/ext/tidy/attrget.c +213 -0
  13. data/ext/tidy/attrs.c +1911 -0
  14. data/ext/tidy/attrs.h +374 -0
  15. data/ext/tidy/buffio.c +232 -0
  16. data/ext/tidy/buffio.h +118 -0
  17. data/ext/tidy/charsets.c +1032 -0
  18. data/ext/tidy/charsets.h +14 -0
  19. data/ext/tidy/clean.c +2674 -0
  20. data/ext/tidy/clean.h +87 -0
  21. data/ext/tidy/config.c +1746 -0
  22. data/ext/tidy/config.h +153 -0
  23. data/ext/tidy/entities.c +419 -0
  24. data/ext/tidy/entities.h +24 -0
  25. data/ext/tidy/extconf.rb +5 -0
  26. data/ext/tidy/fileio.c +106 -0
  27. data/ext/tidy/fileio.h +46 -0
  28. data/ext/tidy/forward.h +69 -0
  29. data/ext/tidy/iconvtc.c +105 -0
  30. data/ext/tidy/iconvtc.h +15 -0
  31. data/ext/tidy/istack.c +373 -0
  32. data/ext/tidy/lexer.c +3825 -0
  33. data/ext/tidy/lexer.h +617 -0
  34. data/ext/tidy/localize.c +1882 -0
  35. data/ext/tidy/mappedio.c +329 -0
  36. data/ext/tidy/mappedio.h +16 -0
  37. data/ext/tidy/message.h +207 -0
  38. data/ext/tidy/parser.c +4408 -0
  39. data/ext/tidy/parser.h +76 -0
  40. data/ext/tidy/platform.h +636 -0
  41. data/ext/tidy/pprint.c +2276 -0
  42. data/ext/tidy/pprint.h +93 -0
  43. data/ext/tidy/ruby-tidy.c +195 -0
  44. data/ext/tidy/streamio.c +1407 -0
  45. data/ext/tidy/streamio.h +222 -0
  46. data/ext/tidy/tagask.c +286 -0
  47. data/ext/tidy/tags.c +955 -0
  48. data/ext/tidy/tags.h +235 -0
  49. data/ext/tidy/tidy-int.h +129 -0
  50. data/ext/tidy/tidy.h +1097 -0
  51. data/ext/tidy/tidyenum.h +622 -0
  52. data/ext/tidy/tidylib.c +1751 -0
  53. data/ext/tidy/tmbstr.c +306 -0
  54. data/ext/tidy/tmbstr.h +92 -0
  55. data/ext/tidy/utf8.c +539 -0
  56. data/ext/tidy/utf8.h +52 -0
  57. data/ext/tidy/version.h +14 -0
  58. data/ext/tidy/win32tc.c +795 -0
  59. data/ext/tidy/win32tc.h +19 -0
  60. data/spec/spec_helper.rb +5 -0
  61. data/spec/tidy/compat_spec.rb +44 -0
  62. data/spec/tidy/remote_uri_spec.rb +14 -0
  63. data/spec/tidy/test1.html +5 -0
  64. data/spec/tidy/tidy_spec.rb +34 -0
  65. metadata +125 -0
data/ext/tidy/win32tc.h ADDED
@@ -0,0 +1,19 @@
1
+ #ifndef __WIN32TC_H__
2
+ #define __WIN32TC_H__
3
+ #ifdef TIDY_WIN32_MLANG_SUPPORT
4
+
5
+ /* win32tc.h -- Interface to Win32 transcoding routines
6
+
7
+ (c) 1998-2006 (W3C) MIT, ERCIM, Keio University
8
+ See tidy.h for the copyright notice.
9
+
10
+ $Id: win32tc.h,v 1.3 2006/12/29 16:31:09 arnaud02 Exp $
11
+ */
12
+
13
+ uint TY_(Win32MLangGetCPFromName)(TidyAllocator *allocator,ctmbstr encoding);
14
+ Bool TY_(Win32MLangInitInputTranscoder)(StreamIn * in, uint wincp);
15
+ void TY_(Win32MLangUninitInputTranscoder)(StreamIn * in);
16
+ int TY_(Win32MLangGetChar)(byte firstByte, StreamIn * in, uint * bytesRead);
17
+
18
+ #endif /* TIDY_WIN32_MLANG_SUPPORT */
19
+ #endif /* __WIN32TC_H__ */
data/spec/spec_helper.rb ADDED
@@ -0,0 +1,5 @@
1
+ $:.unshift File.expand_path(File.dirname(__FILE__) + "/../lib")
2
+
3
+ require 'rubygems'
4
+ require 'spec'
5
+ require 'tidy'
data/spec/tidy/compat_spec.rb ADDED
@@ -0,0 +1,44 @@
1
+ require File.join(File.dirname(__FILE__), "..", "spec_helper")
2
+
3
+ describe "tidy compatibility methods" do
4
+
5
+ subject { Tidy.new }
6
+
7
+ it "should be defined" do
8
+ defined?(Tidy).should be_true
9
+ end
10
+
11
+ it "should have an open method" do
12
+ Tidy.should respond_to :open
13
+ end
14
+
15
+ it "should have a clean method" do
16
+ subject.should respond_to :clean
17
+ end
18
+
19
+ it "should have an errors attribute" do
20
+ subject.should respond_to :errors
21
+ end
22
+
23
+ it "should have a path attribute" do
24
+ Tidy.should respond_to :path
25
+ Tidy.path = "hello world"
26
+ Tidy.path.should == "hello world"
27
+ end
28
+
29
+ it "should parse a string" do
30
+ tidy = Tidy.open({}) do |tidy|
31
+ xml = tidy.clean("<html><body>String</body></html>")
32
+ xml.should_not be_empty
33
+ end
34
+ end
35
+
36
+ it "should be able to parse more than one string consecutively" do
37
+ tidy = Tidy.open({}) do |tidy|
38
+ errors1, html1 = tidy.clean("<html><body>String</body></html>")
39
+ errors2, html2 = tidy.clean("<html><head><title>hello</title></head><body>String</body></html>")
40
+ errors1.should_not == errors2
41
+ end
42
+ end
43
+
44
+ end
data/spec/tidy/remote_uri_spec.rb ADDED
@@ -0,0 +1,14 @@
1
+ require File.join(File.dirname(__FILE__), "..", "spec_helper")
2
+ require 'open-uri'
3
+
4
+ describe "tidy class methods" do
5
+
6
+ subject { Tidy.new }
7
+
8
+ it "should parse the google.com uri" do
9
+ uri = open("http://www.google.com")
10
+ page = uri.read
11
+ errors, html = subject.parse(page)
12
+ end
13
+
14
+ end
data/spec/tidy/test1.html ADDED
@@ -0,0 +1,5 @@
1
+ <html>
2
+ <body>
3
+ Hello World!
4
+ </body>
5
+ </html>
data/spec/tidy/tidy_spec.rb ADDED
@@ -0,0 +1,34 @@
1
+ require File.join(File.dirname(__FILE__), "..", "spec_helper")
2
+
3
+ describe "tidy class methods" do
4
+
5
+ subject { Tidy.new }
6
+
7
+ it "should have a new method" do
8
+ Tidy.should respond_to :new
9
+ end
10
+
11
+ it "should parse a string" do
12
+ errors, html = subject.parse("<html><body>String</body></html>")
13
+ errors.should_not be_nil
14
+ html.should_not be_nil
15
+ end
16
+
17
+ it "should parse a file" do
18
+ file = File.new(File.join(File.dirname(__FILE__),'test1.html'))
19
+ errors, html = subject.parse(file)
20
+ errors.should_not be_nil
21
+ html.should_not be_nil
22
+ end
23
+
24
+ it "should respond to access" do
25
+ subject.should respond_to(:access)
26
+ end
27
+
28
+ it "should be able to parse more than one string consecutively" do
29
+ errors1, html1 = subject.parse("<html><body>String</body></html>")
30
+ errors2, html2 = subject.parse("<html><head><title>hello</title></head><body>String</body></html>")
31
+ errors1.should_not == errors2
32
+ end
33
+
34
+ end
metadata ADDED
@@ -0,0 +1,125 @@
1
+ --- !ruby/object:Gem::Specification
2
+ name: tidy-ext
3
+ version: !ruby/object:Gem::Version
4
+ prerelease: false
5
+ segments:
6
+ - 0
7
+ - 1
8
+ - 7
9
+ version: 0.1.7
10
+ platform: ruby
11
+ authors:
12
+ - Carl Douglas
13
+ autorequire:
14
+ bindir: bin
15
+ cert_chain: []
16
+
17
+ date: 2010-06-10 00:00:00 +10:00
18
+ default_executable:
19
+ dependencies: []
20
+
21
+ description: " Tidies up web pages.\n"
22
+ email: carl.douglas@gmail.com
23
+ executables: []
24
+
25
+ extensions:
26
+ - ext/tidy/extconf.rb
27
+ extra_rdoc_files:
28
+ - LICENSE
29
+ files:
30
+ - .gitignore
31
+ - LICENSE
32
+ - README
33
+ - Rakefile
34
+ - VERSION
35
+ - ext/tidy/access.c
36
+ - ext/tidy/access.h
37
+ - ext/tidy/alloc.c
38
+ - ext/tidy/attrask.c
39
+ - ext/tidy/attrdict.c
40
+ - ext/tidy/attrdict.h
41
+ - ext/tidy/attrget.c
42
+ - ext/tidy/attrs.c
43
+ - ext/tidy/attrs.h
44
+ - ext/tidy/buffio.c
45
+ - ext/tidy/buffio.h
46
+ - ext/tidy/charsets.c
47
+ - ext/tidy/charsets.h
48
+ - ext/tidy/clean.c
49
+ - ext/tidy/clean.h
50
+ - ext/tidy/config.c
51
+ - ext/tidy/config.h
52
+ - ext/tidy/entities.c
53
+ - ext/tidy/entities.h
54
+ - ext/tidy/extconf.rb
55
+ - ext/tidy/fileio.c
56
+ - ext/tidy/fileio.h
57
+ - ext/tidy/forward.h
58
+ - ext/tidy/iconvtc.c
59
+ - ext/tidy/iconvtc.h
60
+ - ext/tidy/istack.c
61
+ - ext/tidy/lexer.c
62
+ - ext/tidy/lexer.h
63
+ - ext/tidy/localize.c
64
+ - ext/tidy/mappedio.c
65
+ - ext/tidy/mappedio.h
66
+ - ext/tidy/message.h
67
+ - ext/tidy/parser.c
68
+ - ext/tidy/parser.h
69
+ - ext/tidy/platform.h
70
+ - ext/tidy/pprint.c
71
+ - ext/tidy/pprint.h
72
+ - ext/tidy/ruby-tidy.c
73
+ - ext/tidy/streamio.c
74
+ - ext/tidy/streamio.h
75
+ - ext/tidy/tagask.c
76
+ - ext/tidy/tags.c
77
+ - ext/tidy/tags.h
78
+ - ext/tidy/tidy-int.h
79
+ - ext/tidy/tidy.h
80
+ - ext/tidy/tidyenum.h
81
+ - ext/tidy/tidylib.c
82
+ - ext/tidy/tmbstr.c
83
+ - ext/tidy/tmbstr.h
84
+ - ext/tidy/utf8.c
85
+ - ext/tidy/utf8.h
86
+ - ext/tidy/version.h
87
+ - ext/tidy/win32tc.c
88
+ - ext/tidy/win32tc.h
89
+ - spec/spec_helper.rb
90
+ - spec/tidy/compat_spec.rb
91
+ - spec/tidy/remote_uri_spec.rb
92
+ - spec/tidy/test1.html
93
+ - spec/tidy/tidy_spec.rb
94
+ has_rdoc: true
95
+ homepage: http://github.com/carld/tidy
96
+ licenses: []
97
+
98
+ post_install_message:
99
+ rdoc_options:
100
+ - --charset=UTF-8
101
+ require_paths:
102
+ - lib
103
+ required_ruby_version: !ruby/object:Gem::Requirement
104
+ requirements:
105
+ - - ">="
106
+ - !ruby/object:Gem::Version
107
+ segments:
108
+ - 0
109
+ version: "0"
110
+ required_rubygems_version: !ruby/object:Gem::Requirement
111
+ requirements:
112
+ - - ">="
113
+ - !ruby/object:Gem::Version
114
+ segments:
115
+ - 0
116
+ version: "0"
117
+ requirements: []
118
+
119
+ rubyforge_project:
120
+ rubygems_version: 1.3.6
121
+ signing_key:
122
+ specification_version: 3
123
+ summary: HTML Tidy library implemented as a Ruby extension.
124
+ test_files: []
125
+