right_scraper 1.0.26 → 3.0.0
Sign up to get free protection for your applications and to get access to all the features.
- data/Gemfile +16 -0
- data/README.rdoc +9 -28
- data/Rakefile +51 -39
- data/lib/right_scraper/builders/base.rb +64 -0
- data/lib/right_scraper/builders/filesystem.rb +96 -0
- data/lib/right_scraper/builders/union.rb +57 -0
- data/lib/right_scraper/logger.rb +102 -0
- data/lib/right_scraper/loggers/noisy.rb +85 -0
- data/lib/right_scraper/processes/ssh.rb +188 -0
- data/lib/right_scraper/repositories/base.rb +299 -0
- data/lib/right_scraper/repositories/download.rb +90 -0
- data/lib/right_scraper/repositories/git.rb +92 -0
- data/lib/right_scraper/repositories/mock.rb +70 -0
- data/lib/right_scraper/repositories/svn.rb +96 -0
- data/lib/right_scraper/resources/base.rb +70 -0
- data/{spec/scraper_base_spec.rb → lib/right_scraper/resources/cookbook.rb} +9 -23
- data/lib/right_scraper/resources/workflow.rb +55 -0
- data/lib/right_scraper/retrievers/base.rb +114 -0
- data/lib/right_scraper/retrievers/checkout.rb +79 -0
- data/lib/right_scraper/retrievers/download.rb +97 -0
- data/lib/right_scraper/retrievers/git.rb +140 -0
- data/lib/right_scraper/retrievers/svn.rb +87 -0
- data/lib/right_scraper/scanners/base.rb +111 -0
- data/lib/right_scraper/scanners/cookbook_manifest.rb +59 -0
- data/lib/right_scraper/scanners/cookbook_metadata.rb +69 -0
- data/lib/right_scraper/scanners/cookbook_s3_upload.rb +84 -0
- data/lib/right_scraper/scanners/union.rb +89 -0
- data/lib/right_scraper/scanners/workflow_manifest.rb +86 -0
- data/lib/right_scraper/scanners/workflow_metadata.rb +70 -0
- data/lib/right_scraper/scanners/workflow_s3_upload.rb +85 -0
- data/lib/right_scraper/scraper.rb +81 -57
- data/lib/right_scraper/scraper_logger.rb +61 -0
- data/lib/right_scraper/scrapers/base.rb +262 -0
- data/lib/right_scraper/scrapers/cookbook.rb +73 -0
- data/lib/right_scraper/scrapers/workflow.rb +88 -0
- data/lib/right_scraper/svn_client.rb +101 -0
- data/lib/right_scraper/version.rb +28 -0
- data/lib/right_scraper.rb +35 -11
- data/right_scraper.gemspec +26 -13
- data/right_scraper.rconf +13 -0
- data/spec/builder_spec.rb +50 -0
- data/spec/cookbook_helper.rb +73 -0
- data/spec/cookbook_manifest_spec.rb +55 -0
- data/spec/cookbook_s3_upload_spec.rb +152 -0
- data/spec/download/download_retriever_spec.rb +118 -0
- data/spec/download/download_retriever_spec_helper.rb +72 -0
- data/spec/download/download_spec.rb +130 -0
- data/spec/download/multi_dir_spec.rb +106 -0
- data/spec/download/multi_dir_spec_helper.rb +40 -0
- data/spec/git/cookbook_spec.rb +166 -0
- data/spec/git/demokey +27 -0
- data/spec/git/demokey.pub +1 -0
- data/spec/git/password_key +30 -0
- data/spec/git/password_key.pub +1 -0
- data/spec/git/repository_spec.rb +110 -0
- data/spec/git/retriever_spec.rb +505 -0
- data/spec/git/retriever_spec_helper.rb +112 -0
- data/spec/git/scraper_spec.rb +136 -0
- data/spec/git/ssh_spec.rb +170 -0
- data/spec/git/url_spec.rb +103 -0
- data/spec/logger_spec.rb +185 -0
- data/spec/repository_spec.rb +89 -23
- data/spec/{scraper_spec_helper_base.rb → retriever_spec_helper.rb} +41 -27
- data/spec/scanner_spec.rb +61 -0
- data/spec/scraper_helper.rb +96 -0
- data/spec/scraper_spec.rb +123 -45
- data/spec/spec_helper.rb +87 -14
- data/spec/svn/cookbook_spec.rb +97 -0
- data/spec/svn/multi_svn_spec.rb +64 -0
- data/spec/svn/multi_svn_spec_helper.rb +40 -0
- data/spec/svn/repository_spec.rb +72 -0
- data/spec/svn/retriever_spec.rb +261 -0
- data/spec/svn/scraper_spec.rb +90 -0
- data/spec/svn/{svn_scraper_spec_helper.rb → svn_retriever_spec_helper.rb} +46 -27
- data/spec/svn/url_spec.rb +47 -0
- data/spec/url_spec.rb +164 -0
- metadata +203 -31
- data/lib/right_scraper/linux/process_monitor.rb +0 -84
- data/lib/right_scraper/repository.rb +0 -78
- data/lib/right_scraper/scraper_base.rb +0 -175
- data/lib/right_scraper/scrapers/download_scraper.rb +0 -67
- data/lib/right_scraper/scrapers/git_scraper.rb +0 -283
- data/lib/right_scraper/scrapers/svn_scraper.rb +0 -119
- data/lib/right_scraper/watcher.rb +0 -158
- data/lib/right_scraper/win32/process_monitor.rb +0 -98
- data/spec/download/download_scraper_spec.rb +0 -94
- data/spec/git/git_scraper_spec.rb +0 -165
- data/spec/git/git_scraper_spec_helper.rb +0 -72
- data/spec/rcov.opts +0 -1
- data/spec/spec.opts +0 -2
- data/spec/svn/svn_scraper_spec.rb +0 -148
- data/spec/watcher_spec.rb +0 -74
data/spec/repository_spec.rb
CHANGED
@@ -1,18 +1,18 @@
|
|
1
1
|
#--
|
2
|
-
# Copyright: Copyright (c) 2010 RightScale, Inc.
|
2
|
+
# Copyright: Copyright (c) 2010-2011 RightScale, Inc.
|
3
3
|
#
|
4
4
|
# Permission is hereby granted, free of charge, to any person obtaining
|
5
5
|
# a copy of this software and associated documentation files (the
|
6
6
|
# 'Software'), to deal in the Software without restriction, including
|
7
7
|
# without limitation the rights to use, copy, modify, merge, publish,
|
8
|
-
# distribute, sublicense, and/or sell copies of the Software, and to
|
8
|
+
# distribute, sublicense, and/or sell copies of the Software, and to
|
9
9
|
# permit persons to whom the Software is furnished to do so, subject to
|
10
10
|
# the following conditions:
|
11
11
|
#
|
12
12
|
# The above copyright notice and this permission notice shall be
|
13
13
|
# included in all copies or substantial portions of the Software.
|
14
14
|
#
|
15
|
-
# THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
15
|
+
# THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
16
16
|
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
17
17
|
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
18
18
|
# IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
@@ -21,25 +21,91 @@
|
|
21
21
|
# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
22
22
|
#++
|
23
23
|
|
24
|
-
require File.join(File.dirname(__FILE__), 'spec_helper')
|
25
|
-
require '
|
26
|
-
|
27
|
-
describe
|
28
|
-
|
29
|
-
|
30
|
-
|
31
|
-
|
32
|
-
|
33
|
-
|
34
|
-
|
35
|
-
|
36
|
-
|
37
|
-
|
38
|
-
|
39
|
-
|
40
|
-
|
41
|
-
|
42
|
-
|
24
|
+
require File.expand_path(File.join(File.dirname(__FILE__), 'spec_helper'))
|
25
|
+
require File.expand_path(File.join(File.dirname(__FILE__), '..', 'lib', 'right_scraper', 'repositories', 'mock'))
|
26
|
+
|
27
|
+
describe RightScraper::Repositories::Base do
|
28
|
+
include RightScraper::SpecHelpers::DevelopmentModeEnvironment
|
29
|
+
|
30
|
+
context 'with a repository type that doesn\'t exist' do
|
31
|
+
it 'should throw a comprehensible error when you try to create it' do
|
32
|
+
lambda {
|
33
|
+
RightScraper::Repositories::Base.from_hash(:display_name => 'display_name',
|
34
|
+
:repo_type => :nonexistent,
|
35
|
+
:url => 'url',
|
36
|
+
:tag => 'tag',
|
37
|
+
:first_credential => 'first_credential',
|
38
|
+
:second_credential => 'second_credential')
|
39
|
+
}.should raise_error(/Can't understand how to make nonexistent repos/)
|
40
|
+
end
|
41
|
+
end
|
42
|
+
|
43
|
+
context 'with a mock repository' do
|
44
|
+
before(:each) do
|
45
|
+
@repo = RightScraper::Repositories::Base.from_hash(:display_name => 'display_name',
|
46
|
+
:repo_type => :mock,
|
47
|
+
:url => 'url',
|
48
|
+
:tag => 'tag',
|
49
|
+
:first_credential => 'first_credential',
|
50
|
+
:second_credential => 'second_credential')
|
51
|
+
end
|
52
|
+
|
53
|
+
it 'should be initializable from a hash' do
|
54
|
+
@repo.should be_kind_of(RightScraper::Repositories::Base)
|
55
|
+
@repo.display_name.should == 'display_name'
|
56
|
+
@repo.repo_type.should == :mock
|
57
|
+
@repo.url.should == 'url'
|
58
|
+
@repo.tag.should == 'tag'
|
59
|
+
@repo.first_credential.should == 'first_credential'
|
60
|
+
@repo.second_credential.should == 'second_credential'
|
61
|
+
end
|
62
|
+
|
63
|
+
it 'should know the SHA-1 of its root location' do
|
64
|
+
@repo.repository_hash.should ==
|
65
|
+
Digest::SHA1.hexdigest("1\000mock\000url")
|
66
|
+
end
|
67
|
+
|
68
|
+
it 'should know the SHA-1 of the identifier for this specific checkout' do
|
69
|
+
@repo.checkout_hash.should ==
|
70
|
+
Digest::SHA1.hexdigest("1\000mock\000url")
|
71
|
+
end
|
72
|
+
end
|
73
|
+
end
|
74
|
+
|
75
|
+
describe RightScraper::Repositories::Download do
|
76
|
+
include RightScraper::SpecHelpers::DevelopmentModeEnvironment
|
77
|
+
|
78
|
+
before(:each) do
|
79
|
+
@repo = RightScraper::Repositories::Base.from_hash(:display_name => 'test repo',
|
80
|
+
:repo_type => :download,
|
81
|
+
:url => "http://foo.bar.baz.quux/%20CBLAH",
|
82
|
+
:tag => "412530982323",
|
83
|
+
:first_credential => "foo:b/ar",
|
84
|
+
:second_credential => "foo@bar")
|
85
|
+
end
|
86
|
+
|
87
|
+
it 'should have a tag' do
|
88
|
+
@repo.tag.should == '412530982323'
|
89
|
+
end
|
90
|
+
|
91
|
+
it 'should include the tag in the checkout hash' do
|
92
|
+
@repo.checkout_hash.should_not == @repo.repository_hash
|
93
|
+
oldhash = @repo.checkout_hash
|
94
|
+
@repo.tag = "42398"
|
95
|
+
@repo.checkout_hash.should_not == oldhash
|
96
|
+
end
|
97
|
+
|
98
|
+
it 'should have the same repository hash with or without credentials' do
|
99
|
+
initial_hash = @repo.repository_hash
|
100
|
+
@repo.first_credential = nil
|
101
|
+
@repo.second_credential = nil
|
102
|
+
@repo.repository_hash.should == initial_hash
|
103
|
+
end
|
104
|
+
|
105
|
+
it 'should have the same checkout hash with or without credentials' do
|
106
|
+
initial_hash = @repo.checkout_hash
|
107
|
+
@repo.first_credential = nil
|
108
|
+
@repo.second_credential = nil
|
109
|
+
@repo.checkout_hash.should == initial_hash
|
43
110
|
end
|
44
|
-
|
45
111
|
end
|
@@ -1,5 +1,5 @@
|
|
1
1
|
#--
|
2
|
-
# Copyright: Copyright (c) 2010 RightScale, Inc.
|
2
|
+
# Copyright: Copyright (c) 2010-2011 RightScale, Inc.
|
3
3
|
#
|
4
4
|
# Permission is hereby granted, free of charge, to any person obtaining
|
5
5
|
# a copy of this software and associated documentation files (the
|
@@ -21,33 +21,67 @@
|
|
21
21
|
# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
22
22
|
#++
|
23
23
|
|
24
|
+
require 'digest/sha1'
|
24
25
|
require File.expand_path(File.join(File.dirname(__FILE__), 'spec_helper'))
|
26
|
+
require 'tmpdir'
|
25
27
|
|
26
|
-
module
|
28
|
+
module RightScraper
|
27
29
|
|
28
30
|
# Base class for all scrapers spec helpers
|
29
31
|
# Define helper methods used to manage repositories using each
|
30
32
|
# source control software
|
31
|
-
class
|
33
|
+
class RetrieverSpecHelper
|
32
34
|
|
33
35
|
include SpecHelpers
|
34
36
|
|
37
|
+
def initialize
|
38
|
+
@tmpdir = Dir.mktmpdir
|
39
|
+
FileUtils.mkdir(repo_path)
|
40
|
+
@repo_content = [ { 'folder1' => [ 'file2', 'file3' ] },
|
41
|
+
{ 'folder2' => [ { 'folder3' => [ 'file4' ] } ] },
|
42
|
+
'file1' ]
|
43
|
+
end
|
44
|
+
|
45
|
+
def close
|
46
|
+
FileUtils.remove_entry_secure @tmpdir
|
47
|
+
end
|
48
|
+
|
35
49
|
# Path to test repository
|
36
50
|
#
|
37
51
|
# === Return
|
38
52
|
# repo_path(String):: Path to test repository
|
39
53
|
def repo_path
|
40
|
-
|
54
|
+
File.join(@tmpdir, "repository")
|
41
55
|
end
|
42
56
|
|
43
57
|
# Default test repo content
|
44
58
|
#
|
45
59
|
# === Return
|
46
60
|
# content(String):: Default test repo content
|
47
|
-
|
48
|
-
|
61
|
+
attr_reader :repo_content
|
62
|
+
|
63
|
+
def manifest
|
64
|
+
hash = {}
|
65
|
+
scan(@repo_content, hash, nil)
|
66
|
+
hash['metadata.json'] = Digest::SHA1.hexdigest(@repo_content.to_json + "\n")
|
67
|
+
hash
|
49
68
|
end
|
50
69
|
|
70
|
+
def scan(contents, hash, position)
|
71
|
+
contents.each do |object|
|
72
|
+
if object.instance_of?(Hash)
|
73
|
+
object.each do |key, value|
|
74
|
+
relative_position = position ? File.join(position, key) : key
|
75
|
+
scan(value, hash, relative_position)
|
76
|
+
end
|
77
|
+
else
|
78
|
+
relative_position = position ? File.join(position, object) : object
|
79
|
+
hash[relative_position] = Digest::SHA1.hexdigest(object + "\n")
|
80
|
+
end
|
81
|
+
end
|
82
|
+
end
|
83
|
+
private :scan
|
84
|
+
|
51
85
|
# Test branch content
|
52
86
|
#
|
53
87
|
# === Return
|
@@ -64,26 +98,6 @@ module RightScale
|
|
64
98
|
content = [ { 'additional_folder' => [ 'afile1', 'afile2' ] }, 'afile3' ]
|
65
99
|
end
|
66
100
|
|
67
|
-
# Create test repository following given layout
|
68
|
-
# Delete any previously created repo
|
69
|
-
#
|
70
|
-
# === Return
|
71
|
-
# repo_path(String):: Path to created repositoy
|
72
|
-
#
|
73
|
-
# === Raise
|
74
|
-
# Exception:: If repository initialization fails
|
75
|
-
def setup_test_repo
|
76
|
-
raise 'Not supported'
|
77
|
-
end
|
78
|
-
|
79
|
-
# Delete test repository
|
80
|
-
#
|
81
|
-
# === Return
|
82
|
-
# true:: Always return true
|
83
|
-
def delete_test_repo
|
84
|
-
FileUtils.rm_rf(repo_path)
|
85
|
-
end
|
86
|
-
|
87
101
|
# Commit any non-commited changes of given directory
|
88
102
|
#
|
89
103
|
# === Parameters
|
@@ -129,4 +143,4 @@ module RightScale
|
|
129
143
|
raise 'Not implemented'
|
130
144
|
end
|
131
145
|
end
|
132
|
-
end
|
146
|
+
end
|
@@ -0,0 +1,61 @@
|
|
1
|
+
#--
|
2
|
+
# Copyright: Copyright (c) 2010-2011 RightScale, Inc.
|
3
|
+
#
|
4
|
+
# Permission is hereby granted, free of charge, to any person obtaining
|
5
|
+
# a copy of this software and associated documentation files (the
|
6
|
+
# 'Software'), to deal in the Software without restriction, including
|
7
|
+
# without limitation the rights to use, copy, modify, merge, publish,
|
8
|
+
# distribute, sublicense, and/or sell copies of the Software, and to
|
9
|
+
# permit persons to whom the Software is furnished to do so, subject to
|
10
|
+
# the following conditions:
|
11
|
+
#
|
12
|
+
# The above copyright notice and this permission notice shall be
|
13
|
+
# included in all copies or substantial portions of the Software.
|
14
|
+
#
|
15
|
+
# THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
16
|
+
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
17
|
+
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
18
|
+
# IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
19
|
+
# CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
|
20
|
+
# TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
|
21
|
+
# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
22
|
+
#++
|
23
|
+
|
24
|
+
require File.expand_path(File.join(File.dirname(__FILE__), 'spec_helper'))
|
25
|
+
require File.expand_path(File.join(File.dirname(__FILE__), 'download', 'download_retriever_spec_helper'))
|
26
|
+
|
27
|
+
describe RightScraper::Scanners::Base do
|
28
|
+
include RightScraper::SpecHelpers::DevelopmentModeEnvironment
|
29
|
+
|
30
|
+
before(:each) do
|
31
|
+
@helper = RightScraper::DownloadRetrieverSpecHelper.new
|
32
|
+
@repo = @helper.repo
|
33
|
+
end
|
34
|
+
|
35
|
+
after(:each) do
|
36
|
+
@helper.close
|
37
|
+
end
|
38
|
+
|
39
|
+
it 'should be called correctly' do
|
40
|
+
scanner = flexmock("scanner")
|
41
|
+
scanner.should_receive(:new).with(Hash).once.and_return(scanner)
|
42
|
+
scanner.should_receive(:begin).with(RightScraper::Resources::Cookbook).once
|
43
|
+
scanner.should_receive(:notice_dir).with(nil).once.and_return(true)
|
44
|
+
scanner.should_receive(:notice).with("file1", Proc).once
|
45
|
+
scanner.should_receive(:notice_dir).with("folder1").once.and_return(true)
|
46
|
+
scanner.should_receive(:notice).with("folder1/file2", Proc).once
|
47
|
+
scanner.should_receive(:notice).with("folder1/file3", Proc).once
|
48
|
+
scanner.should_receive(:notice_dir).with("folder2").once.and_return(false)
|
49
|
+
scanner.should_receive(:notice).with("metadata.json", Proc).once
|
50
|
+
scanner.should_receive(:end).with(RightScraper::Resources::Cookbook).once
|
51
|
+
scanner.should_receive(:finish).with().once
|
52
|
+
|
53
|
+
@scraper = RightScraper::Scrapers::Base.scraper(:repo_dir => @helper.download_repo_path,
|
54
|
+
:kind => :cookbook,
|
55
|
+
:scanners => [scanner],
|
56
|
+
:repository => @repo)
|
57
|
+
@scraper.next_resource.should_not be_nil
|
58
|
+
@scraper.next_resource.should be_nil
|
59
|
+
@scraper.close
|
60
|
+
end
|
61
|
+
end
|
@@ -0,0 +1,96 @@
|
|
1
|
+
#--
|
2
|
+
# Copyright: Copyright (c) 2010-2011 RightScale, Inc.
|
3
|
+
#
|
4
|
+
# Permission is hereby granted, free of charge, to any person obtaining
|
5
|
+
# a copy of this software and associated documentation files (the
|
6
|
+
# 'Software'), to deal in the Software without restriction, including
|
7
|
+
# without limitation the rights to use, copy, modify, merge, publish,
|
8
|
+
# distribute, sublicense, and/or sell copies of the Software, and to
|
9
|
+
# permit persons to whom the Software is furnished to do so, subject to
|
10
|
+
# the following conditions:
|
11
|
+
#
|
12
|
+
# The above copyright notice and this permission notice shall be
|
13
|
+
# included in all copies or substantial portions of the Software.
|
14
|
+
#
|
15
|
+
# THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
16
|
+
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
17
|
+
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
18
|
+
# IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
19
|
+
# CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
|
20
|
+
# TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
|
21
|
+
# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
22
|
+
#++
|
23
|
+
|
24
|
+
require File.expand_path(File.join(File.dirname(__FILE__), 'spec_helper'))
|
25
|
+
require 'tmpdir'
|
26
|
+
|
27
|
+
module RightScraper
|
28
|
+
module SpecHelpers
|
29
|
+
module FromScratchScraping
|
30
|
+
def FromScratchScraping.included(mod)
|
31
|
+
mod.module_eval do
|
32
|
+
before(:each) do
|
33
|
+
@basedir = Dir.mktmpdir
|
34
|
+
@retriever = @retriever_class.new(@repo,
|
35
|
+
:basedir => @basedir,
|
36
|
+
:max_bytes => 1024**2,
|
37
|
+
:max_seconds => 20)
|
38
|
+
@retriever.retrieve
|
39
|
+
end
|
40
|
+
|
41
|
+
after(:each) do
|
42
|
+
FileUtils.remove_entry_secure(@basedir)
|
43
|
+
@scraper = nil
|
44
|
+
end
|
45
|
+
end
|
46
|
+
end
|
47
|
+
end
|
48
|
+
|
49
|
+
module CookbookScraping
|
50
|
+
def CookbookScraping.included(mod)
|
51
|
+
mod.module_eval do
|
52
|
+
before(:each) do
|
53
|
+
@scraper = RightScraper::Scrapers::Base.scraper(:repo_dir => @retriever.repo_dir,
|
54
|
+
:kind => :cookbook,
|
55
|
+
:repository => @retriever.repository,
|
56
|
+
:ignorable_paths => @retriever.ignorable_paths)
|
57
|
+
end
|
58
|
+
end
|
59
|
+
end
|
60
|
+
end
|
61
|
+
|
62
|
+
module WorkflowScraping
|
63
|
+
def WorkflowScraping.included(mod)
|
64
|
+
mod.module_eval do
|
65
|
+
before(:each) do
|
66
|
+
@scraper = RightScraper::Scrapers::Base.scraper(:repo_dir => @retriever.repo_dir,
|
67
|
+
:kind => :workflow,
|
68
|
+
:repository => @retriever.repository,
|
69
|
+
:ignorable_paths => @retriever.ignorable_paths)
|
70
|
+
end
|
71
|
+
end
|
72
|
+
end
|
73
|
+
end
|
74
|
+
|
75
|
+
end
|
76
|
+
module ScraperHelper
|
77
|
+
def archive_skeleton(archive)
|
78
|
+
files = Set.new
|
79
|
+
Archive.read_open_memory(archive) do |ar|
|
80
|
+
while entry = ar.next_header
|
81
|
+
files << [entry.pathname, ar.read_data]
|
82
|
+
end
|
83
|
+
end
|
84
|
+
files
|
85
|
+
end
|
86
|
+
|
87
|
+
def check_resource(resource, params={})
|
88
|
+
position = params[:position] || "."
|
89
|
+
resource.should_not == nil
|
90
|
+
resource.repository.should be_an_equal_repo @repo
|
91
|
+
resource.pos.should == position
|
92
|
+
resource.metadata.should == (params[:metadata] || @helper.repo_content)
|
93
|
+
resource.manifest.should == (params[:manifest] || @helper.manifest)
|
94
|
+
end
|
95
|
+
end
|
96
|
+
end
|
data/spec/scraper_spec.rb
CHANGED
@@ -1,18 +1,18 @@
|
|
1
1
|
#--
|
2
|
-
# Copyright: Copyright (c) 2010 RightScale, Inc.
|
2
|
+
# Copyright: Copyright (c) 2010-2011 RightScale, Inc.
|
3
3
|
#
|
4
4
|
# Permission is hereby granted, free of charge, to any person obtaining
|
5
5
|
# a copy of this software and associated documentation files (the
|
6
6
|
# 'Software'), to deal in the Software without restriction, including
|
7
7
|
# without limitation the rights to use, copy, modify, merge, publish,
|
8
|
-
# distribute, sublicense, and/or sell copies of the Software, and to
|
8
|
+
# distribute, sublicense, and/or sell copies of the Software, and to
|
9
9
|
# permit persons to whom the Software is furnished to do so, subject to
|
10
10
|
# the following conditions:
|
11
11
|
#
|
12
12
|
# The above copyright notice and this permission notice shall be
|
13
13
|
# included in all copies or substantial portions of the Software.
|
14
14
|
#
|
15
|
-
# THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
15
|
+
# THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
|
16
16
|
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
17
17
|
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
|
18
18
|
# IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
|
@@ -21,49 +21,127 @@
|
|
21
21
|
# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
22
22
|
#++
|
23
23
|
|
24
|
-
require File.join(File.dirname(__FILE__), 'spec_helper')
|
25
|
-
require File.join(File.dirname(__FILE__), '
|
26
|
-
require
|
27
|
-
require
|
28
|
-
|
29
|
-
|
30
|
-
|
31
|
-
|
32
|
-
|
33
|
-
|
34
|
-
|
35
|
-
|
36
|
-
|
37
|
-
|
38
|
-
|
39
|
-
|
40
|
-
|
41
|
-
|
42
|
-
|
43
|
-
|
44
|
-
|
45
|
-
|
46
|
-
|
47
|
-
|
48
|
-
|
49
|
-
|
50
|
-
|
51
|
-
|
52
|
-
|
24
|
+
require File.expand_path(File.join(File.dirname(__FILE__), 'spec_helper'))
|
25
|
+
require File.expand_path(File.join(File.dirname(__FILE__), 'download', 'download_retriever_spec_helper'))
|
26
|
+
require 'tmpdir'
|
27
|
+
require 'flexmock'
|
28
|
+
|
29
|
+
describe RightScraper::Scraper do
|
30
|
+
include RightScraper::SpecHelpers::DevelopmentModeEnvironment
|
31
|
+
|
32
|
+
include RightScraper::SharedExamples
|
33
|
+
|
34
|
+
shared_examples_for 'scrapes to given base dir' do
|
35
|
+
before(:each) do
|
36
|
+
@scraper = RightScraper::Scraper.new(:basedir => @tmpdir, :kind => :cookbook)
|
37
|
+
end
|
38
|
+
|
39
|
+
after(:each) do
|
40
|
+
FileUtils.remove_entry_secure @tmpdir unless @tmpdir.nil?
|
41
|
+
end
|
42
|
+
|
43
|
+
it 'starts out successful' do
|
44
|
+
@scraper.succeeded?.should be_true
|
45
|
+
@scraper.errors.should == []
|
46
|
+
end
|
47
|
+
|
48
|
+
context 'given a legal download repository' do
|
49
|
+
before(:each) do
|
50
|
+
@helper = RightScraper::DownloadRetrieverSpecHelper.new
|
51
|
+
@repo = @helper.repo
|
52
|
+
end
|
53
|
+
|
54
|
+
after(:each) do
|
55
|
+
@helper.close
|
56
|
+
end
|
57
|
+
|
58
|
+
it_should_behave_like "a normal repository"
|
59
|
+
|
60
|
+
it 'should log correctly as it scrapes' do
|
61
|
+
callback = flexmock("callback")
|
62
|
+
callback.should_receive(:call).with(:begin, :retrieving, "from #{@repo}", nil).once.ordered
|
63
|
+
callback.should_receive(:call).with(:begin, :initialize, String, nil).once.ordered
|
64
|
+
callback.should_receive(:call).with(:commit, :initialize, String, nil).once.ordered
|
65
|
+
callback.should_receive(:call).with(:begin, :downloading, "", nil).once.ordered
|
66
|
+
callback.should_receive(:call).with(:begin, :running_command, String, nil).once.ordered
|
67
|
+
callback.should_receive(:call).with(:commit, :running_command, String, nil).once.ordered
|
68
|
+
callback.should_receive(:call).with(:commit, :downloading, "", nil).once.ordered
|
69
|
+
callback.should_receive(:call).with(:begin, :unpacking, "", nil).once.ordered
|
70
|
+
callback.should_receive(:call).with(:begin, :running_command, String, nil).once.ordered
|
71
|
+
callback.should_receive(:call).with(:commit, :running_command, String, nil).once.ordered
|
72
|
+
callback.should_receive(:call).with(:commit, :unpacking, "", nil).once.ordered
|
73
|
+
callback.should_receive(:call).with(:commit, :retrieving, "from #{@repo}", nil).once.ordered
|
74
|
+
callback.should_receive(:call).with(:begin, :scraping, String, nil).once.ordered
|
75
|
+
callback.should_receive(:call).with(:begin, :finding_next_cookbook, String, nil).once.ordered
|
76
|
+
callback.should_receive(:call).with(:begin, :reading_cookbook, String, nil).once.ordered
|
77
|
+
callback.should_receive(:call).with(:begin, :scanning_filesystem, String, nil).once.ordered
|
78
|
+
callback.should_receive(:call).with(:begin, :metadata_parsing, "", nil).once.ordered
|
79
|
+
callback.should_receive(:call).with(:commit, :metadata_parsing, "", nil).once.ordered
|
80
|
+
callback.should_receive(:call).with(:commit, :scanning_filesystem, String, nil).once.ordered
|
81
|
+
callback.should_receive(:call).with(:commit, :reading_cookbook, String, nil).once.ordered
|
82
|
+
callback.should_receive(:call).with(:commit, :finding_next_cookbook, String, nil).once.ordered
|
83
|
+
callback.should_receive(:call).with(:begin, :next, "", nil).once.ordered
|
84
|
+
callback.should_receive(:call).with(:begin, :searching, "", nil).once.ordered
|
85
|
+
callback.should_receive(:call).with(:commit, :searching, "", nil).once.ordered
|
86
|
+
callback.should_receive(:call).with(:commit, :next, "", nil).once.ordered
|
87
|
+
callback.should_receive(:call).with(:begin, :next, "", nil).once.ordered
|
88
|
+
callback.should_receive(:call).with(:commit, :next, "", nil).once.ordered
|
89
|
+
callback.should_receive(:call).with(:commit, :scraping, String, nil).once.ordered
|
90
|
+
@scraper.scrape(@repo) do |phase, operation, explanation, exception|
|
91
|
+
callback.call(phase, operation, explanation, exception)
|
92
|
+
end
|
93
|
+
@scraper.errors.should == []
|
94
|
+
@scraper.succeeded?.should be_true
|
95
|
+
@scraper.resources.size.should == 1
|
96
|
+
end
|
97
|
+
end
|
98
|
+
|
99
|
+
context 'given several repositories' do
|
100
|
+
it 'should continue to scrape even if errors occur' do
|
101
|
+
GC.start
|
102
|
+
repo = RightScraper::Repositories::Base.from_hash(:display_name => 'illegal repo',
|
103
|
+
:repo_type => :download,
|
104
|
+
:url => "http://example.com/foo")
|
105
|
+
@scraper.scrape(repo)
|
106
|
+
helpers = [RightScraper::DownloadRetrieverSpecHelper,
|
107
|
+
RightScraper::DownloadRetrieverSpecHelper,
|
108
|
+
RightScraper::DownloadRetrieverSpecHelper]
|
109
|
+
helpers.each do |klass|
|
110
|
+
helper = klass.new
|
111
|
+
@scraper.scrape(helper.repo)
|
112
|
+
helper.close
|
113
|
+
end
|
114
|
+
@scraper.succeeded?.should be_false
|
115
|
+
@scraper.resources.size.should == 3
|
116
|
+
@scraper.errors.size.should == 1
|
117
|
+
end
|
118
|
+
end
|
119
|
+
|
120
|
+
it 'catches normal logging behavior' do
|
121
|
+
logger = @scraper.instance_variable_get(:@logger)
|
122
|
+
logger.should_not be_nil
|
123
|
+
logger.info("foo")
|
124
|
+
logger.error("foo")
|
125
|
+
@scraper.succeeded?.should be_false
|
126
|
+
@scraper.errors.should == [[nil, :log, {:severity => Logger::ERROR,
|
127
|
+
:message => "foo",
|
128
|
+
:progname => nil}]]
|
129
|
+
end
|
53
130
|
end
|
54
|
-
|
55
|
-
|
56
|
-
|
57
|
-
|
58
|
-
|
59
|
-
|
131
|
+
|
132
|
+
context 'when base dir is not provided' do
|
133
|
+
before(:each) do
|
134
|
+
@tmpdir = nil
|
135
|
+
end
|
136
|
+
|
137
|
+
it_should_behave_like 'scrapes to given base dir'
|
60
138
|
end
|
61
|
-
|
62
|
-
|
63
|
-
|
64
|
-
|
65
|
-
|
66
|
-
|
139
|
+
|
140
|
+
context 'when base dir is provided' do
|
141
|
+
before(:each) do
|
142
|
+
@tmpdir = Dir.mktmpdir
|
143
|
+
end
|
144
|
+
|
145
|
+
it_should_behave_like 'scrapes to given base dir'
|
67
146
|
end
|
68
|
-
|
69
147
|
end
|