html-proofer 3.18.6 → 3.19.2

Sign up to get free protection for your applications and to get access to all the features.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: e2d6deb85601ff3f5e5df02999f1b55f29231e5658f1d4fa9986c002b3cc0263
4
- data.tar.gz: e20a007144a866b41b97eb36a964ca1372d693ade77acc69a9c9e706b353c8ef
3
+ metadata.gz: d06dbda8bf9baad3be96b5565fcb86de0892d4241c6b7ede08c3c2d7203f6752
4
+ data.tar.gz: da895d696b7b9d1f3ca9c2e504e0c375ce9942f53dd5c2174f80570d32e3ad5b
5
5
  SHA512:
6
- metadata.gz: c51154450eb20c04c0b7038effb12ab8d258457ea5560421b5a2a0739342504afebb8fd092fe4378838a076636ff2665eb691f513b8b34b49af2e89ffba988c2
7
- data.tar.gz: fcb7bf492588fabc991caf1f779fca14a7b553d3ce00dbf55ba24aebf3c39f7613af6a408d94c3a3faaeb21dceb8265ece91ca9bf4320aee7d4f3dca133ba4de
6
+ metadata.gz: 693f677cb91b9b0e79135ef27ed2771d34d7e5fee5bd368b20b5791a4e25a468ba769d299c2cd1417eccd989d9ac1290cfaf9dfd90b9534fa0af0440382e4e0f
7
+ data.tar.gz: c6c33f309e3f8b00dc1721653f2e2099c647c9bd8d3b4ecaab4ec7b97161925911d1e6e3c8cabb30b9b9eb2274fed7238ed338da7d85d25b7c6981d537f6435d
data/bin/htmlproofer CHANGED
@@ -48,6 +48,7 @@ Mercenary.program(:htmlproofer) do |p|
48
48
  p.option 'storage_dir', '--storage-dir PATH', String, 'Directory where to store the cache log (default: "tmp/.htmlproofer")'
49
49
  p.option 'timeframe', '--timeframe <time>', String, 'A string representing the caching timeframe.'
50
50
  p.option 'typhoeus_config', '--typhoeus-config CONFIG', String, 'JSON-formatted string of Typhoeus config. Will override the html-proofer defaults.'
51
+ p.option 'hydra_config', '--hydra-config CONFIG', String, 'JSON-formatted string of Hydra config. Will override the html-proofer defaults.'
51
52
  p.option 'url_ignore', '--url-ignore link1,[link2,...]', Array, 'A comma-separated list of Strings or RegExps containing URLs that are safe to ignore. It affects all HTML attributes. Note that non-HTTP(S) URIs are always ignored'
52
53
  p.option 'url_swap', '--url-swap re:string,[re:string,...]', Array, 'A comma-separated list containing key-value pairs of `RegExp => String`. It transforms URLs that match `RegExp` into `String` via `gsub`. The escape sequences `\\:` should be used to produce literal `:`s.'
53
54
  p.option 'root_dir', '--root-dir PATH', String, 'The absolute path to the directory serving your html-files.'
@@ -87,7 +88,8 @@ Mercenary.program(:htmlproofer) do |p|
87
88
  options[:validation][:report_eof_tags] = opts['report_eof_tags'] unless opts['report_eof_tags'].nil?
88
89
  options[:validation][:report_mismatched_tags] = opts['report_mismatched_tags'] unless opts['report_mismatched_tags'].nil?
89
90
 
90
- options[:typhoeus] = HTMLProofer::Configuration.parse_json_option('typhoeus_config', opts['typhoeus_config']) unless opts['typhoeus_config'].nil?
91
+ options[:typhoeus] = HTMLProofer::Configuration.parse_json_option('typhoeus_config', opts['typhoeus_config'], symbolize_names: false) unless opts['typhoeus_config'].nil?
92
+ options[:hydra] = HTMLProofer::Configuration.parse_json_option('hydra_config', opts['hydra_config']) unless opts['hydra_config'].nil?
91
93
 
92
94
  unless opts['timeframe'].nil?
93
95
  options[:cache] ||= {}
@@ -3,6 +3,7 @@
3
3
  require_relative 'utils'
4
4
  require 'date'
5
5
  require 'json'
6
+ require 'uri'
6
7
 
7
8
  module HTMLProofer
8
9
  class Cache
@@ -11,6 +12,8 @@ module HTMLProofer
11
12
  DEFAULT_STORAGE_DIR = File.join('tmp', '.htmlproofer')
12
13
  DEFAULT_CACHE_FILE_NAME = 'cache.log'
13
14
 
15
+ URI_REGEXP = URI::DEFAULT_PARSER.make_regexp
16
+
14
17
  attr_reader :exists, :cache_log, :storage_dir, :cache_file
15
18
 
16
19
  def initialize(logger, options)
@@ -30,6 +33,8 @@ module HTMLProofer
30
33
  end
31
34
 
32
35
  def within_timeframe?(time)
36
+ return false if time.nil?
37
+
33
38
  (@parsed_timeframe..@cache_time).cover?(Time.parse(time))
34
39
  end
35
40
 
@@ -71,10 +76,14 @@ module HTMLProofer
71
76
  @cache_log[clean_url(url)] = data
72
77
  end
73
78
 
74
- def detect_url_changes(found)
75
- existing_urls = @cache_log.keys.map { |url| clean_url(url) }
79
+ def detect_url_changes(found, type)
76
80
  found_urls = found.keys.map { |url| clean_url(url) }
77
81
 
82
+ # if there were no urls, bail
83
+ return {} if found_urls.empty?
84
+
85
+ existing_urls = @cache_log.keys.map { |url| clean_url(url) }
86
+
78
87
  # prepare to add new URLs detected
79
88
  additions = found.reject do |url, _|
80
89
  url = clean_url(url)
@@ -91,19 +100,20 @@ module HTMLProofer
91
100
  @logger.log :info, "Adding #{new_link_text} to the cache..."
92
101
 
93
102
  # remove from cache URLs that no longer exist
94
- del = 0
103
+ deletions = 0
95
104
  @cache_log.delete_if do |url, _|
96
105
  url = clean_url(url)
106
+
97
107
  if found_urls.include?(url)
98
108
  false
99
- else
109
+ elsif url_matches_type?(url, type)
100
110
  @logger.log :debug, "Removing #{url} from cache check"
101
- del += 1
111
+ deletions += 1
102
112
  true
103
113
  end
104
114
  end
105
115
 
106
- del_link_text = pluralize(del, 'link', 'links')
116
+ del_link_text = pluralize(deletions, 'link', 'links')
107
117
  @logger.log :info, "Removing #{del_link_text} from the cache..."
108
118
 
109
119
  additions
@@ -113,21 +123,22 @@ module HTMLProofer
113
123
  # caches need access to this file. Write a proper versioned
114
124
  # schema in the future
115
125
  def write
116
- file = {}
117
- file = JSON.parse(File.read(cache_file)) if File.exist?(cache_file)
118
- File.write(cache_file, file.merge(@cache_log).to_json)
126
+ File.write(cache_file, @cache_log.to_json)
119
127
  end
120
128
 
121
129
  def load?
122
130
  @load.nil?
123
131
  end
124
132
 
125
- def retrieve_urls(urls)
126
- urls_to_check = detect_url_changes(urls)
133
+ def retrieve_urls(urls, type)
134
+ urls_to_check = detect_url_changes(urls, type)
135
+
127
136
  @cache_log.each_pair do |url, cache|
128
137
  next if within_timeframe?(cache['time']) && cache['message'].empty? # these were successes to skip
129
138
 
130
- urls_to_check[url] = cache['filenames'] # recheck expired links
139
+ if url_matches_type?(url, type)
140
+ urls_to_check[url] = cache['filenames'] # recheck expired links
141
+ end
131
142
  end
132
143
  urls_to_check
133
144
  end
@@ -154,9 +165,9 @@ module HTMLProofer
154
165
 
155
166
  @cache_file = File.join(storage_dir, cache_file_name)
156
167
 
157
- return unless File.exist?(cache_file)
168
+ return unless File.exist?(@cache_file)
158
169
 
159
- contents = File.read(cache_file)
170
+ contents = File.read(@cache_file)
160
171
  @cache_log = contents.empty? ? {} : JSON.parse(contents)
161
172
  end
162
173
 
@@ -174,5 +185,10 @@ module HTMLProofer
174
185
  @cache_datetime - Rational(measurement / 24.0)
175
186
  end.to_time
176
187
  end
188
+
189
+ def url_matches_type?(url, type)
190
+ return true if type == :internal && url !~ URI_REGEXP
191
+ return true if type == :external && url =~ URI_REGEXP
192
+ end
177
193
  end
178
194
  end
@@ -67,7 +67,7 @@ module HTMLProofer
67
67
  end
68
68
  end
69
69
 
70
- def self.parse_json_option(option_name, config)
70
+ def self.parse_json_option(option_name, config, symbolize_names: true)
71
71
  raise ArgumentError, 'Must provide an option name in string format.' unless option_name.is_a?(String)
72
72
  raise ArgumentError, 'Must provide an option name in string format.' if option_name.strip.empty?
73
73
 
@@ -78,7 +78,7 @@ module HTMLProofer
78
78
  return {} if config.strip.empty?
79
79
 
80
80
  begin
81
- JSON.parse(config)
81
+ JSON.parse(config, { symbolize_names: symbolize_names })
82
82
  rescue StandardError
83
83
  raise ArgumentError, "Option '#{option_name} did not contain valid JSON."
84
84
  end
@@ -125,7 +125,7 @@ module HTMLProofer
125
125
  end
126
126
 
127
127
  external_urls = check.external_urls
128
- external_urls = check.external_urls.map { |url, file| [swap(url, @options[:url_swap]), file] }.to_h if @options[:url_swap]
128
+ external_urls = check.external_urls.transform_keys { |url| swap(url, @options[:url_swap]) } if @options[:url_swap]
129
129
  result[:external_urls].merge!(external_urls)
130
130
  result[:failures].concat(check.issues)
131
131
  end
@@ -238,7 +238,7 @@ module HTMLProofer
238
238
  end
239
239
 
240
240
  def load_internal_cache
241
- urls_to_check = @cache.retrieve_urls(@internal_urls)
241
+ urls_to_check = @cache.retrieve_urls(@internal_urls, :internal)
242
242
  cache_text = pluralize(urls_to_check.count, 'internal link', 'internal links')
243
243
  @logger.log :info, "Found #{cache_text} in the cache..."
244
244
 
@@ -26,7 +26,7 @@ module HTMLProofer
26
26
  @external_urls = remove_query_values
27
27
 
28
28
  if @cache.use_cache?
29
- urls_to_check = @cache.retrieve_urls(@external_urls)
29
+ urls_to_check = @cache.retrieve_urls(@external_urls, :external)
30
30
  external_link_checker(urls_to_check)
31
31
  @cache.write
32
32
  else
@@ -1,5 +1,5 @@
1
1
  # frozen_string_literal: true
2
2
 
3
3
  module HTMLProofer
4
- VERSION = '3.18.6'
4
+ VERSION = '3.19.2'
5
5
  end
metadata CHANGED
@@ -1,14 +1,14 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: html-proofer
3
3
  version: !ruby/object:Gem::Version
4
- version: 3.18.6
4
+ version: 3.19.2
5
5
  platform: ruby
6
6
  authors:
7
7
  - Garen Torikian
8
8
  autorequire:
9
9
  bindir: bin
10
10
  cert_chain: []
11
- date: 2021-02-21 00:00:00.000000000 Z
11
+ date: 2021-06-24 00:00:00.000000000 Z
12
12
  dependencies:
13
13
  - !ruby/object:Gem::Dependency
14
14
  name: addressable
@@ -287,7 +287,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
287
287
  requirements:
288
288
  - - ">="
289
289
  - !ruby/object:Gem::Version
290
- version: 2.4.10
290
+ version: 2.6.0
291
291
  - - "<"
292
292
  - !ruby/object:Gem::Version
293
293
  version: '4.0'