Class: Downloader

Inherits:
Object
  • Object
show all
Defined in:
lib/uniprop/downloader.rb

Class Method Summary collapse

Class Method Details

.cache_file(url, name, cache_dir = nil) ⇒ Object 



186
187
188
189
190
191
192
193
194
195
196
197
 
# File 'lib/uniprop/downloader.rb', line 186

def self.cache_file(url, name, cache_dir = nil)
  case cache_dir
  when false
    return nil
  when nil
    cache_dir = ENV['CACHE_DIR']
    if !cache_dir or cache_dir.empty?
      cache_dir = ".downloaded-cache"
    end
  end
  Pathname.new(cache_dir) + (name || File.basename(URI(url).path))
end

.download(url, name, dir = nil, since = true, options = {}) ⇒ Object 



84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
 
# File 'lib/uniprop/downloader.rb', line 84

def self.download(url, name, dir = nil, since = true, options = {})
  options = options.dup
  url = URI(url)
  dryrun = options.delete(:dryrun)
  options.delete(:unicode_beta) # just to be on the safe side for gems and gcc

  if name
    file = Pathname.new(under(dir, name))
  else
    name = File.basename(url.path)
  end
  cache_save = options.delete(:cache_save) {
    ENV["CACHE_SAVE"] != "no"
  }
  cache = cache_file(url, name, options.delete(:cache_dir))
  file ||= cache
  if since.nil? and file.exist?
    if $VERBOSE
      $stdout.puts "#{file} already exists"
      $stdout.flush
    end
    if cache_save
      save_cache(cache, file, name)
    end
    return file.to_path
  end
  if dryrun
    puts "Download #{url} into #{file}"
    return
  end
  if link_cache(cache, file, name, $VERBOSE)
    return file.to_path
  end
  if !https? and URI::HTTPS === url
    warn "*** using http instead of https ***"
    url.scheme = 'http'
    url = URI(url.to_s)
  end
  if $VERBOSE
    $stdout.print "downloading #{name} ... "
    $stdout.flush
  end
  mtime = nil
  options = options.merge(http_options(file, since.nil? ? true : since))
  begin
    data = with_retry(10) do
      data = url.read(options)
      if mtime = data.meta["last-modified"]
        mtime = Time.httpdate(mtime)
      end
      data
    end
  rescue OpenURI::HTTPError => http_error
    if http_error.message =~ /^304 / # 304 Not Modified
      if $VERBOSE
        $stdout.puts "not modified"
        $stdout.flush
      end
      return file.to_path
    end
    raise
  rescue Timeout::Error
    if since.nil? and file.exist?
      puts "Request for #{url} timed out, using old version."
      return file.to_path
    end
    raise
  rescue SocketError
    if since.nil? and file.exist?
      puts "No network connection, unable to download #{url}, using old version."
      return file.to_path
    end
    raise
  end
  dest = (cache_save && cache && !cache.exist? ? cache : file)
  dest.parent.mkpath
  dest.open("wb", 0600) do |f|
    data.scrub!('?') # invalid byte sequence in UTF-8 対策
    f.write(data)
    f.chmod(mode_for(data))
  end
  if mtime
    dest.utime(mtime, mtime)
  end
  if $VERBOSE
    $stdout.puts "done"
    $stdout.flush
  end
  if dest.eql?(cache)
    link_cache(cache, file, name)
  elsif cache_save
    save_cache(cache, file, name)
  end
  return file.to_path
rescue => e
  raise "failed to download #{name}\n#{e.class}: #{e.message}: #{url}"
end

.http_options(file, since) ⇒ Object 



52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
 
# File 'lib/uniprop/downloader.rb', line 52

def self.http_options(file, since)
  options = {}
  if since
    case since
    when true
      since = (File.mtime(file).httpdate rescue nil)
    when Time
      since = since.httpdate
    end
    if since
      options['If-Modified-Since'] = since
    end
  end
  options['Accept-Encoding'] = 'identity' # to disable Net::HTTP::GenericRequest#decode_content
  options
end

.httpdate(date) ⇒ Object 



69
70
71
72
73
74
75
76
77
78
79
80
81
82
 
# File 'lib/uniprop/downloader.rb', line 69

def self.httpdate(date)
  Time.httpdate(date)
rescue ArgumentError => e
  # Some hosts (e.g., zlib.net) return similar to RFC 850 but 4
  # digit year, sometimes.
  /\A\s*
   (?:Mon|Tues|Wednes|Thurs|Fri|Satur|Sun)day,\x20
   (\d\d)-(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)-(\d{4})\x20
   (\d\d):(\d\d):(\d\d)\x20
   GMT
   \s*\z/ix =~ date or raise
  warn e.message
  Time.utc($3, $2, $1, $4, $5, $6)
end

.https ⇒ Object 



44
45
46
 
# File 'lib/uniprop/downloader.rb', line 44

def self.https
  @@https
end

.https=(https) ⇒ Object 



36
37
38
 
# File 'lib/uniprop/downloader.rb', line 36

def self.https=(https)
  @@https = https
end

.https? ⇒ Boolean

Returns:

  • (Boolean) 


40
41
42
 
# File 'lib/uniprop/downloader.rb', line 40

def self.https?
  @@https == 'https'
end

.link_cache(cache, file, name, verbose = false) ⇒ Object 



199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
 
# File 'lib/uniprop/downloader.rb', line 199

def self.link_cache(cache, file, name, verbose = false)
  return false unless cache and cache.exist?
  return true if cache.eql?(file)
  if /cygwin/ !~ RUBY_PLATFORM or /winsymlink:nativestrict/ =~ ENV['CYGWIN']
    begin
      file.make_symlink(cache.relative_path_from(file.parent))
    rescue SystemCallError
    else
      if verbose
        $stdout.puts "made symlink #{name} to #{cache}"
        $stdout.flush
      end
      return true
    end
  end
  begin
    file.make_link(cache)
  rescue SystemCallError
  else
    if verbose
      $stdout.puts "made link #{name} to #{cache}"
      $stdout.flush
    end
    return true
  end
end

.mode_for(data) ⇒ Object 



48
49
50
 
# File 'lib/uniprop/downloader.rb', line 48

def self.mode_for(data)
  /\A#!/ =~ data ? 0755 : 0644
end

.save_cache(cache, file, name) ⇒ Object 



226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
 
# File 'lib/uniprop/downloader.rb', line 226

def self.save_cache(cache, file, name)
  return unless cache or cache.eql?(file)
  begin
    st = cache.stat
  rescue
    begin
      file.rename(cache)
    rescue
      return
    end
  else
    return unless st.mtime > file.lstat.mtime
    file.unlink
  end
  link_cache(cache, file, name)
end

.under(dir, name) ⇒ Object 



182
183
184
 
# File 'lib/uniprop/downloader.rb', line 182

def self.under(dir, name)
  dir ? File.join(dir, File.basename(name)) : name
end