ronin-web 0.2.1 → 0.3.0.pre1
Sign up to get free protection for your applications and to get access to all the features.
- data/.document +4 -0
- data/.gemtest +0 -0
- data/.rspec +1 -0
- data/.yardopts +1 -0
- data/COPYING.txt +674 -0
- data/{History.txt → ChangeLog.md} +39 -8
- data/Gemfile +49 -0
- data/{README.txt → README.md} +56 -40
- data/Rakefile +25 -24
- data/bin/ronin-web +18 -5
- data/gemspec.yml +30 -0
- data/lib/ronin/network/{helpers → mixins}/web.rb +74 -39
- data/lib/ronin/web.rb +7 -6
- data/lib/ronin/web/extensions.rb +7 -6
- data/lib/ronin/web/extensions/nokogiri.rb +22 -0
- data/lib/ronin/web/extensions/nokogiri/xml.rb +22 -0
- data/lib/ronin/web/extensions/nokogiri/xml/attr.rb +35 -1
- data/lib/ronin/web/extensions/nokogiri/xml/document.rb +34 -4
- data/lib/ronin/web/extensions/nokogiri/xml/element.rb +36 -2
- data/lib/ronin/web/extensions/nokogiri/xml/node.rb +59 -20
- data/lib/ronin/web/extensions/nokogiri/xml/text.rb +36 -2
- data/lib/ronin/web/middleware.rb +27 -0
- data/lib/ronin/web/middleware/base.rb +144 -0
- data/lib/ronin/web/middleware/directories.rb +177 -0
- data/lib/ronin/web/middleware/files.rb +142 -0
- data/lib/ronin/web/middleware/filters.rb +28 -0
- data/lib/ronin/web/middleware/filters/campaign_filter.rb +77 -0
- data/lib/ronin/web/middleware/filters/ip_filter.rb +73 -0
- data/lib/ronin/web/middleware/filters/path_filter.rb +73 -0
- data/lib/ronin/web/middleware/filters/referer_filter.rb +71 -0
- data/lib/ronin/web/middleware/filters/user_agent_filter.rb +71 -0
- data/lib/ronin/web/middleware/filters/vhost_filter.rb +71 -0
- data/lib/ronin/web/middleware/helpers.rb +145 -0
- data/lib/ronin/web/middleware/proxy.rb +265 -0
- data/lib/ronin/web/middleware/proxy_request.rb +262 -0
- data/lib/ronin/web/middleware/request.rb +58 -0
- data/lib/ronin/web/middleware/response.rb +33 -0
- data/lib/ronin/web/middleware/router.rb +167 -0
- data/lib/ronin/web/middleware/rule.rb +103 -0
- data/lib/ronin/web/proxy.rb +7 -6
- data/lib/ronin/web/proxy/app.rb +7 -6
- data/lib/ronin/web/proxy/base.rb +14 -9
- data/lib/ronin/web/proxy/web.rb +9 -6
- data/lib/ronin/web/server.rb +7 -6
- data/lib/ronin/web/server/app.rb +7 -6
- data/lib/ronin/web/server/base.rb +209 -82
- data/lib/ronin/web/server/web.rb +10 -6
- data/lib/ronin/web/spider.rb +38 -26
- data/lib/ronin/web/version.rb +8 -7
- data/lib/ronin/web/web.rb +122 -80
- data/ronin-web.gemspec +15 -0
- data/spec/helpers/output.rb +3 -0
- data/spec/spec_helper.rb +2 -3
- data/spec/web/extensions/nokogiri_spec.rb +7 -7
- data/spec/web/{server/helpers/server.rb → helpers/rack_app.rb} +3 -14
- data/spec/web/helpers/root.rb +15 -0
- data/spec/web/helpers/root/test1.txt +1 -0
- data/spec/web/helpers/root/test1/index.html +1 -0
- data/spec/web/helpers/root/test1/test1.txt +1 -0
- data/spec/web/helpers/root/test2.txt +1 -0
- data/spec/web/helpers/root/test2/test2.txt +1 -0
- data/spec/web/helpers/root/test3.txt +1 -0
- data/spec/web/helpers/root/test3/test3.txt +1 -0
- data/spec/web/middleware/directories_spec.rb +86 -0
- data/spec/web/middleware/files_spec.rb +57 -0
- data/spec/web/middleware/filters/campaign_filter_spec.rb +30 -0
- data/spec/web/middleware/filters/ip_filter_spec.rb +25 -0
- data/spec/web/middleware/filters/path_filter_spec.rb +29 -0
- data/spec/web/middleware/filters/referer_filter_spec.rb +25 -0
- data/spec/web/middleware/filters/user_agent_filter_spec.rb +25 -0
- data/spec/web/middleware/filters/vhost_filter_spec.rb +23 -0
- data/spec/web/middleware/proxy_spec.rb +67 -0
- data/spec/web/middleware/response_spec.rb +20 -0
- data/spec/web/middleware/router_spec.rb +65 -0
- data/spec/web/middleware/rule_spec.rb +37 -0
- data/spec/web/proxy/base_spec.rb +1 -2
- data/spec/web/server/base_spec.rb +4 -13
- data/spec/web/server/classes/sub_app.rb +2 -2
- data/spec/web/server/classes/test_app.rb +1 -1
- data/spec/web/web_spec.rb +57 -67
- metadata +171 -146
- data.tar.gz.sig +0 -0
- data/Manifest.txt +0 -64
- data/lib/ronin/scanners/web.rb +0 -52
- data/lib/ronin/web/server/files.rb +0 -92
- data/lib/ronin/web/server/helpers.rb +0 -25
- data/lib/ronin/web/server/helpers/files.rb +0 -126
- data/lib/ronin/web/server/helpers/hosts.rb +0 -72
- data/lib/ronin/web/server/helpers/proxy.rb +0 -153
- data/lib/ronin/web/server/helpers/rendering.rb +0 -36
- data/lib/ronin/web/server/hosts.rb +0 -86
- data/lib/ronin/web/server/proxy.rb +0 -116
- data/spec/scanners/web_spec.rb +0 -24
- data/spec/web/helpers/root/index.html +0 -1
- data/spec/web/helpers/root/test.txt +0 -1
- data/spec/web/helpers/server.rb +0 -10
- data/spec/web/server/classes/files/dir/file.txt +0 -1
- data/spec/web/server/classes/files/dir/index.html +0 -1
- data/spec/web/server/classes/files/dir2/file2.txt +0 -1
- data/spec/web/server/classes/files/dir3/page.xml +0 -4
- data/spec/web/server/classes/files/file.txt +0 -1
- data/spec/web/server/classes/files_app.rb +0 -27
- data/spec/web/server/classes/hosts_app.rb +0 -40
- data/spec/web/server/classes/proxy_app.rb +0 -45
- data/spec/web/server/files_spec.rb +0 -74
- data/spec/web/server/hosts_spec.rb +0 -55
- data/spec/web/server/proxy_spec.rb +0 -49
- data/tasks/spec.rb +0 -10
- data/tasks/yard.rb +0 -13
- metadata.gz.sig +0 -3
@@ -2,43 +2,68 @@
|
|
2
2
|
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
3
|
# scraping and spidering functionality.
|
4
4
|
#
|
5
|
-
# Copyright (c) 2006-
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
6
|
#
|
7
|
-
# This
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
8
10
|
# it under the terms of the GNU General Public License as published by
|
9
|
-
# the Free Software Foundation
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
10
12
|
# (at your option) any later version.
|
11
13
|
#
|
12
|
-
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
13
15
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
14
16
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
15
17
|
# GNU General Public License for more details.
|
16
18
|
#
|
17
19
|
# You should have received a copy of the GNU General Public License
|
18
|
-
# along with
|
19
|
-
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
20
21
|
#
|
21
22
|
|
22
|
-
require 'ronin/
|
23
|
+
require 'ronin/ui/output/helpers'
|
23
24
|
require 'ronin/network/http/proxy'
|
24
25
|
require 'ronin/web/web'
|
26
|
+
require 'ronin/mixin'
|
27
|
+
|
28
|
+
require 'parameters'
|
25
29
|
|
26
30
|
module Ronin
|
27
31
|
module Network
|
28
|
-
module
|
32
|
+
module Mixins
|
29
33
|
module Web
|
30
|
-
|
34
|
+
|
35
|
+
mixin UI::Output::Helpers, Parameters
|
36
|
+
|
37
|
+
mixin do
|
38
|
+
# The Web Proxy host to connect to
|
39
|
+
parameter :web_proxy_host, :type => String,
|
40
|
+
:description => 'Web Proxy host'
|
41
|
+
|
42
|
+
# The Web Proxy port to connect to
|
43
|
+
parameter :web_proxy_port, :type => Integer,
|
44
|
+
:description => 'Web Proxy port'
|
45
|
+
|
46
|
+
# The Web Proxy user to authenticate with
|
47
|
+
parameter :web_proxy_user, :type => String,
|
48
|
+
:description => 'Web Proxy authentication user'
|
49
|
+
|
50
|
+
# The Web Proxy password to authenticate with
|
51
|
+
parameter :web_proxy_password, :type => String,
|
52
|
+
:description => 'Web Proxy authentication password'
|
53
|
+
end
|
31
54
|
|
32
55
|
protected
|
33
56
|
|
34
57
|
#
|
35
|
-
# Combines the proxy information set by the
|
36
|
-
#
|
37
|
-
#
|
58
|
+
# Combines the proxy information set by the `web_proxy_host`,
|
59
|
+
# `web_proxy_port`, `web_proxy_user` and `web_proxy_password`
|
60
|
+
# parameters.
|
38
61
|
#
|
39
62
|
# @return [Network::HTTP::Proxy]
|
40
63
|
# The current proxy information.
|
41
64
|
#
|
65
|
+
# @api semipublic
|
66
|
+
#
|
42
67
|
def web_proxy
|
43
68
|
HTTP::Proxy.new(
|
44
69
|
:host => @web_proxy_host,
|
@@ -57,12 +82,14 @@ module Ronin
|
|
57
82
|
# @option options [Hash] :proxy (web_proxy)
|
58
83
|
# Proxy information.
|
59
84
|
#
|
60
|
-
# @option options [String] :user_agent (
|
85
|
+
# @option options [String] :user_agent (web_user_agent)
|
61
86
|
# User-Agent string to use.
|
62
87
|
#
|
63
|
-
# @return [
|
88
|
+
# @return [Mechanize]
|
64
89
|
# The agent.
|
65
90
|
#
|
91
|
+
# @api semipublic
|
92
|
+
#
|
66
93
|
def web_agent(options={},&block)
|
67
94
|
unless @web_agent
|
68
95
|
options[:proxy] ||= web_proxy
|
@@ -86,23 +113,26 @@ module Ronin
|
|
86
113
|
# @option options [Hash] :proxy (web_proxy)
|
87
114
|
# Proxy information.
|
88
115
|
#
|
89
|
-
# @option options [String] :user_agent (
|
116
|
+
# @option options [String] :user_agent (web_user_agent)
|
90
117
|
# User-Agent string to use.
|
91
118
|
#
|
92
119
|
# @yield [page]
|
93
120
|
# If a block is given, it will be passed the page for the
|
94
121
|
# requested URL.
|
95
122
|
#
|
96
|
-
# @yieldparam [
|
123
|
+
# @yieldparam [Mechanize::Page] page
|
97
124
|
# The requested page.
|
98
125
|
#
|
99
|
-
# @return [
|
126
|
+
# @return [Mechanize::Page]
|
100
127
|
# The requested page.
|
101
128
|
#
|
102
|
-
|
129
|
+
# @api semipublic
|
130
|
+
#
|
131
|
+
def web_get(url,options={})
|
132
|
+
print_info "Requesting #{url}"
|
103
133
|
page = web_agent(options).get(url)
|
104
134
|
|
105
|
-
|
135
|
+
yield page if block_given?
|
106
136
|
return page
|
107
137
|
end
|
108
138
|
|
@@ -119,7 +149,7 @@ module Ronin
|
|
119
149
|
# @option options [Hash] :proxy (web_proxy)
|
120
150
|
# Proxy information.
|
121
151
|
#
|
122
|
-
# @option options [String] :user_agent (
|
152
|
+
# @option options [String] :user_agent (web_user_agent)
|
123
153
|
# User-Agent string to use.
|
124
154
|
#
|
125
155
|
# @yield [body]
|
@@ -131,13 +161,14 @@ module Ronin
|
|
131
161
|
# @return [String]
|
132
162
|
# The requested body of the page.
|
133
163
|
#
|
134
|
-
|
135
|
-
|
136
|
-
|
164
|
+
# @api semipublic
|
165
|
+
#
|
166
|
+
def web_get_body(url,options={})
|
167
|
+
page = web_get(url,options)
|
168
|
+
body = page.body
|
137
169
|
|
138
|
-
|
139
|
-
|
140
|
-
end
|
170
|
+
yield body if block_given?
|
171
|
+
return body
|
141
172
|
end
|
142
173
|
|
143
174
|
#
|
@@ -156,26 +187,29 @@ module Ronin
|
|
156
187
|
# @option options [Hash] :proxy (web_proxy)
|
157
188
|
# Proxy information.
|
158
189
|
#
|
159
|
-
# @option options [String] :user_agent (
|
190
|
+
# @option options [String] :user_agent (web_user_agent)
|
160
191
|
# User-Agent string to use.
|
161
192
|
#
|
162
193
|
# @yield [page]
|
163
194
|
# If a block is given, it will be passed the page for the
|
164
195
|
# requested URL.
|
165
196
|
#
|
166
|
-
# @yieldparam [
|
197
|
+
# @yieldparam [Mechanize::Page] page
|
167
198
|
# The requested page.
|
168
199
|
#
|
169
|
-
# @return [
|
200
|
+
# @return [Mechanize::Page]
|
170
201
|
# The requested page.
|
171
202
|
#
|
172
|
-
|
203
|
+
# @api semipublic
|
204
|
+
#
|
205
|
+
def web_post(url,options={})
|
173
206
|
query = {}
|
174
207
|
query.merge!(options[:query]) if options[:query]
|
175
208
|
|
209
|
+
print_info "Posting #{url}"
|
176
210
|
page = web_agent(options).post(url)
|
177
211
|
|
178
|
-
|
212
|
+
yield page if block_given?
|
179
213
|
return page
|
180
214
|
end
|
181
215
|
|
@@ -195,25 +229,26 @@ module Ronin
|
|
195
229
|
# @option options [Hash] :proxy (web_proxy)
|
196
230
|
# Proxy information.
|
197
231
|
#
|
198
|
-
# @option options [String] :user_agent (
|
232
|
+
# @option options [String] :user_agent (web_user_agent)
|
199
233
|
# User-Agent string to use.
|
200
234
|
#
|
201
235
|
# @yield [body]
|
202
236
|
# If a block is given, it will be passed the body of the page.
|
203
237
|
#
|
204
|
-
# @yieldparam [
|
238
|
+
# @yieldparam [Mechanize::Page] page
|
205
239
|
# The body of the requested page.
|
206
240
|
#
|
207
241
|
# @return [String]
|
208
242
|
# The requested body of the page.
|
209
243
|
#
|
210
|
-
|
211
|
-
|
212
|
-
|
244
|
+
# @api semipublic
|
245
|
+
#
|
246
|
+
def web_post_body(url,options={})
|
247
|
+
page = web_post(url,options)
|
248
|
+
body = page.body
|
213
249
|
|
214
|
-
|
215
|
-
|
216
|
-
end
|
250
|
+
yield body if block_given?
|
251
|
+
return body
|
217
252
|
end
|
218
253
|
end
|
219
254
|
end
|
data/lib/ronin/web.rb
CHANGED
@@ -2,21 +2,22 @@
|
|
2
2
|
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
3
|
# scraping and spidering functionality.
|
4
4
|
#
|
5
|
-
# Copyright (c) 2006-
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
6
|
#
|
7
|
-
# This
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
8
10
|
# it under the terms of the GNU General Public License as published by
|
9
|
-
# the Free Software Foundation
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
10
12
|
# (at your option) any later version.
|
11
13
|
#
|
12
|
-
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
13
15
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
14
16
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
15
17
|
# GNU General Public License for more details.
|
16
18
|
#
|
17
19
|
# You should have received a copy of the GNU General Public License
|
18
|
-
# along with
|
19
|
-
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
20
21
|
#
|
21
22
|
|
22
23
|
require 'ronin/web/extensions'
|
data/lib/ronin/web/extensions.rb
CHANGED
@@ -2,21 +2,22 @@
|
|
2
2
|
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
3
|
# scraping and spidering functionality.
|
4
4
|
#
|
5
|
-
# Copyright (c) 2006-
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
6
|
#
|
7
|
-
# This
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
8
10
|
# it under the terms of the GNU General Public License as published by
|
9
|
-
# the Free Software Foundation
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
10
12
|
# (at your option) any later version.
|
11
13
|
#
|
12
|
-
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
13
15
|
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
14
16
|
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
15
17
|
# GNU General Public License for more details.
|
16
18
|
#
|
17
19
|
# You should have received a copy of the GNU General Public License
|
18
|
-
# along with
|
19
|
-
# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
20
21
|
#
|
21
22
|
|
22
23
|
require 'ronin/web/extensions/nokogiri'
|
@@ -1 +1,23 @@
|
|
1
|
+
#
|
2
|
+
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
|
+
# scraping and spidering functionality.
|
4
|
+
#
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
|
+
#
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
10
|
+
# it under the terms of the GNU General Public License as published by
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
12
|
+
# (at your option) any later version.
|
13
|
+
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
15
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
16
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
17
|
+
# GNU General Public License for more details.
|
18
|
+
#
|
19
|
+
# You should have received a copy of the GNU General Public License
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
21
|
+
#
|
22
|
+
|
1
23
|
require 'ronin/web/extensions/nokogiri/xml'
|
@@ -1,3 +1,25 @@
|
|
1
|
+
#
|
2
|
+
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
|
+
# scraping and spidering functionality.
|
4
|
+
#
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
|
+
#
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
10
|
+
# it under the terms of the GNU General Public License as published by
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
12
|
+
# (at your option) any later version.
|
13
|
+
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
15
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
16
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
17
|
+
# GNU General Public License for more details.
|
18
|
+
#
|
19
|
+
# You should have received a copy of the GNU General Public License
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
21
|
+
#
|
22
|
+
|
1
23
|
require 'ronin/web/extensions/nokogiri/xml/node'
|
2
24
|
require 'ronin/web/extensions/nokogiri/xml/text'
|
3
25
|
require 'ronin/web/extensions/nokogiri/xml/attr'
|
@@ -1,10 +1,44 @@
|
|
1
|
+
#
|
2
|
+
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
|
+
# scraping and spidering functionality.
|
4
|
+
#
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
|
+
#
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
10
|
+
# it under the terms of the GNU General Public License as published by
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
12
|
+
# (at your option) any later version.
|
13
|
+
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
15
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
16
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
17
|
+
# GNU General Public License for more details.
|
18
|
+
#
|
19
|
+
# You should have received a copy of the GNU General Public License
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
21
|
+
#
|
22
|
+
|
1
23
|
require 'nokogiri'
|
2
24
|
|
3
25
|
module Nokogiri
|
4
26
|
module XML
|
5
27
|
class Attr < Node
|
6
28
|
|
7
|
-
|
29
|
+
#
|
30
|
+
# Determines if the attribute is similar to another attribute.
|
31
|
+
#
|
32
|
+
# @param [Nokogiri::XML::Attr] other
|
33
|
+
# The other attribute.
|
34
|
+
#
|
35
|
+
# @return [Boolean]
|
36
|
+
# Specifies if the attribute is similar, in indentity or value,
|
37
|
+
# to the other attribute.
|
38
|
+
#
|
39
|
+
# @api public
|
40
|
+
#
|
41
|
+
def ==(other)
|
8
42
|
super(other) && (self.value == other.value)
|
9
43
|
end
|
10
44
|
|
@@ -1,3 +1,25 @@
|
|
1
|
+
#
|
2
|
+
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
|
+
# scraping and spidering functionality.
|
4
|
+
#
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
|
+
#
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
10
|
+
# it under the terms of the GNU General Public License as published by
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
12
|
+
# (at your option) any later version.
|
13
|
+
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
15
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
16
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
17
|
+
# GNU General Public License for more details.
|
18
|
+
#
|
19
|
+
# You should have received a copy of the GNU General Public License
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
21
|
+
#
|
22
|
+
|
1
23
|
require 'nokogiri'
|
2
24
|
|
3
25
|
module Nokogiri
|
@@ -5,12 +27,20 @@ module Nokogiri
|
|
5
27
|
class Document < Node
|
6
28
|
|
7
29
|
#
|
8
|
-
#
|
30
|
+
# Calculates the sum of all sub-children of the document.
|
31
|
+
#
|
32
|
+
# @return [Integer]
|
33
|
+
# The total number of children and sub-children of the document.
|
34
|
+
# Returns `0` if the document has no root element.
|
35
|
+
#
|
36
|
+
# @api public
|
9
37
|
#
|
10
38
|
def total_children
|
11
|
-
|
12
|
-
|
13
|
-
|
39
|
+
if root
|
40
|
+
1 + root.total_children
|
41
|
+
else
|
42
|
+
0
|
43
|
+
end
|
14
44
|
end
|
15
45
|
|
16
46
|
end
|
@@ -1,10 +1,44 @@
|
|
1
|
+
#
|
2
|
+
# Ronin Web - A Ruby library for Ronin that provides support for web
|
3
|
+
# scraping and spidering functionality.
|
4
|
+
#
|
5
|
+
# Copyright (c) 2006-2011 Hal Brodigan (postmodern.mod3 at gmail.com)
|
6
|
+
#
|
7
|
+
# This file is part of Ronin Web.
|
8
|
+
#
|
9
|
+
# Ronin is free software: you can redistribute it and/or modify
|
10
|
+
# it under the terms of the GNU General Public License as published by
|
11
|
+
# the Free Software Foundation, either version 3 of the License, or
|
12
|
+
# (at your option) any later version.
|
13
|
+
#
|
14
|
+
# Ronin is distributed in the hope that it will be useful,
|
15
|
+
# but WITHOUT ANY WARRANTY; without even the implied warranty of
|
16
|
+
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
17
|
+
# GNU General Public License for more details.
|
18
|
+
#
|
19
|
+
# You should have received a copy of the GNU General Public License
|
20
|
+
# along with Ronin. If not, see <http://www.gnu.org/licenses/>.
|
21
|
+
#
|
22
|
+
|
1
23
|
require 'nokogiri'
|
2
24
|
|
3
25
|
module Nokogiri
|
4
26
|
module XML
|
5
27
|
class Element < Node
|
6
28
|
|
7
|
-
|
29
|
+
#
|
30
|
+
# Determines if the element is similar to another element.
|
31
|
+
#
|
32
|
+
# @param [Nokogiri::XML::Element] other
|
33
|
+
# The other element.
|
34
|
+
#
|
35
|
+
# @return [Boolean]
|
36
|
+
# Specifies whether the element is equal, in identity or value, to
|
37
|
+
# another element.
|
38
|
+
#
|
39
|
+
# @api public
|
40
|
+
#
|
41
|
+
def ==(other)
|
8
42
|
return false unless super(other)
|
9
43
|
return false unless attribute_nodes.length == other.attribute_nodes.length
|
10
44
|
|
@@ -15,7 +49,7 @@ module Nokogiri
|
|
15
49
|
return false unless attr1.similar?(attr2)
|
16
50
|
end
|
17
51
|
|
18
|
-
true
|
52
|
+
return true
|
19
53
|
end
|
20
54
|
|
21
55
|
end
|