5 # :title: rbot HTTP provider
7 # Author:: Tom Gilbert <tom@linuxbrit.co.uk>
8 # Author:: Giuseppe "Oblomov" Bilotta <giuseppe.bilotta@gmail.com>
9 # Author:: Dmitry "jsn" Kim <dmitry point kim at gmail point com>
18 error "Couldn't load 'net/https': #{e}"
19 error "Secured HTTP connections will fail"
20 # give a nicer error than "undefined method `use_ssl='"
21 ::Net::HTTP.class_eval <<-EOC
22 define_method :use_ssl= do |val|
23 # does anybody really set it to false?
25 raise _("I can't do secure HTTP, sorry (%{msg})") % {
32 # To handle Gzipped pages
38 attr_accessor :no_cache
39 unless method_defined? :raw_body
43 def body_charset(str=self.raw_body)
44 ctype = self['content-type'] || 'text/html'
45 return nil unless ctype =~ /^text/i || ctype =~ /x(ht)?ml/i
47 charsets = ['ISO-8859-1'] # should be in config
49 if ctype.match(/charset=["']?([^\s"']+)["']?/i)
51 debug "charset #{charsets.last} added from header"
55 when /<\?xml\s[^>]*encoding=['"]([^\s"'>]+)["'][^>]*\?>/i
57 debug "xml charset #{charsets.last} added from xml pi"
58 when /<(meta\s[^>]*http-equiv=["']?Content-Type["']?[^>]*)>/i
60 if meta =~ /charset=['"]?([^\s'";]+)['"]?/
62 debug "html charset #{charsets.last} added from meta"
69 charsets = self.body_charset(str) or return str
71 charsets.reverse_each do |charset|
73 debug "try decoding using #{charset}"
74 str.force_encoding(charset)
75 tmp = str.encode('UTF-16le', :invalid => :replace, :replace => '').encode('UTF-8')
81 error 'failed to use encoding'
89 def decompress_body(str)
90 method = self['content-encoding']
94 when /gzip/ # Matches gzip, x-gzip, and the non-rfc-compliant gzip;q=\d sent by some servers
95 debug "gunzipping body"
97 return Zlib::GzipReader.new(StringIO.new(str)).read
98 rescue Zlib::Error => e
99 # If we can't unpack the whole stream (e.g. because we're doing a
101 debug "full gunzipping failed (#{e}), trying to recover as much as possible"
103 ret.force_encoding(Encoding::ASCII_8BIT)
105 Zlib::GzipReader.new(StringIO.new(str)).each_byte { |byte|
113 debug "inflating body"
114 # From http://www.koders.com/ruby/fid927B4382397E5115AC0ABE21181AB5C1CBDD5C17.aspx?s=thread:
115 # -MAX_WBITS stops zlib from looking for a zlib header
116 inflater = Zlib::Inflate.new(-Zlib::MAX_WBITS)
118 return inflater.inflate(str)
119 rescue Zlib::Error => e
122 # debug "full inflation failed (#{e}), trying to recover as much as possible"
124 when /^(?:iso-8859-\d+|windows-\d+|utf-8|utf8)$/i
125 # B0rked servers (Freshmeat being one of them) sometimes return the charset
126 # in the content-encoding; in this case we assume that the document has
127 # a standard content-encoding
128 old_hsh = self.to_hash
129 self['content-type']= self['content-type']+"; charset="+method.downcase
130 warning "Charset vs content-encoding confusion, trying to recover: from\n#{old_hsh.pretty_inspect}to\n#{self.to_hash.pretty_inspect}"
134 raise "Unhandled content encoding #{method}"
139 return self.body_to_utf(self.decompress_body(self.raw_body))
142 # Read chunks from the body until we have at least _size_ bytes, yielding
143 # the partial text at each chunk. Return the partial body.
144 def partial_body(size=0, &block)
149 debug "using body() as partial"
151 yield self.body_to_utf(self.decompress_body(partial)) if block_given?
153 debug "disabling cache"
155 self.read_body { |chunk|
157 yield self.body_to_utf(self.decompress_body(partial)) if block_given?
158 break if size and size > 0 and partial.length >= size
162 return self.body_to_utf(self.decompress_body(partial))
167 Net::HTTP.version_1_2
172 # class for making http requests easier (mainly for plugins to use)
173 # this class can check the bot proxy configuration to determine if a proxy
174 # needs to be used, which includes support for per-url proxy configuration.
176 Bot::Config.register Bot::Config::IntegerValue.new('http.read_timeout',
177 :default => 10, :desc => "Default read timeout for HTTP connections")
178 Bot::Config.register Bot::Config::IntegerValue.new('http.open_timeout',
179 :default => 20, :desc => "Default open timeout for HTTP connections")
180 Bot::Config.register Bot::Config::BooleanValue.new('http.use_proxy',
181 :default => false, :desc => "should a proxy be used for HTTP requests?")
182 Bot::Config.register Bot::Config::StringValue.new('http.proxy_uri', :default => false,
183 :desc => "Proxy server to use for HTTP requests (URI, e.g http://proxy.host:port)")
184 Bot::Config.register Bot::Config::StringValue.new('http.proxy_user',
186 :desc => "User for authenticating with the http proxy (if required)")
187 Bot::Config.register Bot::Config::StringValue.new('http.proxy_pass',
189 :desc => "Password for authenticating with the http proxy (if required)")
190 Bot::Config.register Bot::Config::ArrayValue.new('http.proxy_include',
192 :desc => "List of regexps to check against a URI's hostname/ip to see if we should use the proxy to access this URI. All URIs are proxied by default if the proxy is set, so this is only required to re-include URIs that might have been excluded by the exclude list. e.g. exclude /.*\.foo\.com/, include bar\.foo\.com")
193 Bot::Config.register Bot::Config::ArrayValue.new('http.proxy_exclude',
195 :desc => "List of regexps to check against a URI's hostname/ip to see if we should use avoid the proxy to access this URI and access it directly")
196 Bot::Config.register Bot::Config::IntegerValue.new('http.max_redir',
198 :desc => "Maximum number of redirections to be used when getting a document")
199 Bot::Config.register Bot::Config::IntegerValue.new('http.expire_time',
201 :desc => "After how many minutes since last use a cached document is considered to be expired")
202 Bot::Config.register Bot::Config::IntegerValue.new('http.max_cache_time',
204 :desc => "After how many minutes since first use a cached document is considered to be expired")
205 Bot::Config.register Bot::Config::BooleanValue.new('http.no_expire_cache',
207 :desc => "Set this to true if you want the bot to never expire the cached pages")
208 Bot::Config.register Bot::Config::IntegerValue.new('http.info_bytes',
210 :desc => "How many bytes to download from a web page to find some information. Set to 0 to let the bot download the whole page.")
213 attr_accessor :response, :last_used, :first_used, :count, :expires, :date
215 def self.maybe_new(resp)
216 debug "maybe new #{resp}"
217 return nil if resp.no_cache
218 return nil unless Net::HTTPOK === resp ||
219 Net::HTTPMovedPermanently === resp ||
220 Net::HTTPFound === resp ||
221 Net::HTTPPartialContent === resp
223 cc = resp['cache-control']
224 return nil if cc && (cc =~ /no-cache/i)
228 date = Time.httpdate(d)
231 return nil if resp['expires'] && (Time.httpdate(resp['expires']) < date)
233 debug "creating cache obj"
240 @first_used = now if @count == 0
246 debug "checking expired?"
247 if cc = self.response['cache-control'] && cc =~ /must-revalidate/
250 return self.expires < Time.now
253 def setup_headers(hdr)
254 hdr['if-modified-since'] = self.date.rfc2822
256 debug "ims == #{hdr['if-modified-since']}"
258 if etag = self.response['etag']
259 hdr['if-none-match'] = etag
260 debug "etag: #{etag}"
264 def revalidate(resp = self.response)
267 self.date = resp.key?('date') ? Time.httpdate(resp['date']) : Time.now
269 cc = resp['cache-control']
270 if cc && (cc =~ /max-age=(\d+)/)
271 self.expires = self.date + $1.to_i
272 elsif resp.key?('expires')
273 self.expires = Time.httpdate(resp['expires'])
274 elsif lm = resp['last-modified']
275 delta = self.date - Time.httpdate(lm)
276 delta = 10 if delta <= 0
278 self.expires = self.date + delta
280 self.expires = self.date + 300
282 # self.expires = Time.now + 10 # DEBUG
283 debug "expires on #{self.expires}"
293 self.response.raw_body
294 rescue Exception => e
301 # Create the HttpUtil instance, associating it with Bot _bot_
307 'Accept-Charset' => 'utf-8;q=1.0, *;q=0.8',
308 'Accept-Encoding' => 'gzip;q=1, deflate;q=1, identity;q=0.8, *;q=0.2',
310 "rbot http util #{$version} (#{Irc::Bot::SOURCE_URL})"
312 debug "starting http cache cleanup timer"
313 @timer = @bot.timer.add(300) {
314 self.remove_stale_cache unless @bot.config['http.no_expire_cache']
318 # Clean up on HttpUtil unloading, by stopping the cache cleanup timer.
320 debug 'stopping http cache cleanup timer'
321 @bot.timer.remove(@timer)
324 # This method checks if a proxy is required to access _uri_, by looking at
325 # the values of config values +http.proxy_include+ and +http.proxy_exclude+.
327 # Each of these config values, if set, should be a Regexp the server name and
328 # IP address should be checked against.
330 def proxy_required(uri)
332 if @bot.config["http.proxy_exclude"].empty? && @bot.config["http.proxy_include"].empty?
338 list.concat Resolv.getaddresses(uri.host)
339 rescue StandardError => err
340 warning "couldn't resolve host uri.host"
343 unless @bot.config["http.proxy_exclude"].empty?
344 re = @bot.config["http.proxy_exclude"].collect{|r| Regexp.new(r)}
354 unless @bot.config["http.proxy_include"].empty?
355 re = @bot.config["http.proxy_include"].collect{|r| Regexp.new(r)}
365 debug "using proxy for uri #{uri}?: #{use_proxy}"
369 # _uri_:: URI to create a proxy for
371 # Return a net/http Proxy object, configured for proxying based on the
372 # bot's proxy configuration. See proxy_required for more details on this.
374 def get_proxy(uri, options = {})
376 :read_timeout => @bot.config["http.read_timeout"],
377 :open_timeout => @bot.config["http.open_timeout"]
386 if @bot.config["http.use_proxy"]
387 if (ENV['http_proxy'])
388 proxy = URI.parse ENV['http_proxy'] rescue nil
390 if (@bot.config["http.proxy_uri"])
391 proxy = URI.parse @bot.config["http.proxy_uri"] rescue nil
394 debug "proxy is set to #{proxy.host} port #{proxy.port}"
395 if proxy_required(uri)
396 proxy_host = proxy.host
397 proxy_port = proxy.port
398 proxy_user = @bot.config["http.proxy_user"]
399 proxy_pass = @bot.config["http.proxy_pass"]
404 h = Net::HTTP.new(uri.host, uri.port, proxy_host, proxy_port, proxy_user, proxy_pass)
405 h.use_ssl = true if uri.scheme == "https"
407 h.read_timeout = opts[:read_timeout]
408 h.open_timeout = opts[:open_timeout]
412 # Internal method used to hanlde response _resp_ received when making a
413 # request for URI _uri_.
415 # It follows redirects, optionally yielding them if option :yield is :all.
417 # Also yields and returns the final _resp_.
419 def handle_response(uri, resp, opts, &block) # :yields: resp
420 if Net::HTTPRedirection === resp && opts[:max_redir] >= 0
421 if resp.key?('location')
422 raise 'Too many redirections' if opts[:max_redir] <= 0
423 yield resp if opts[:yield] == :all && block_given?
424 # some servers actually provide unescaped location, e.g.
425 # http://ulysses.soup.io/post/60734021/Image%20curve%20ball
426 # rediects to something like
427 # http://ulysses.soup.io/post/60734021/Image curve ball?sessid=8457b2a3752085cca3fb1d79b9965446
428 # causing the URI parser to (obviously) complain. We cannot just
429 # escape blindly, as this would make a mess of already-escaped
430 # locations, so we only do it if the URI.parse fails
431 loc = resp['location']
433 debug "redirect location: #{loc.inspect}"
435 new_loc = URI.join(uri.to_s, loc) rescue URI.parse(loc)
440 loc = URI.escape(loc)
442 debug "escaped redirect location: #{loc.inspect}"
447 new_opts[:max_redir] -= 1
448 case opts[:method].to_s.downcase.intern
449 when :post, :"net::http::post"
450 new_opts[:method] = :get
452 if resp['set-cookie']
453 debug "set cookie request for #{resp['set-cookie']}"
454 cookie, cookie_flags = (resp['set-cookie']+'; ').split('; ', 2)
456 cookie_flags.scan(/(\S+)=(\S+);/) { |key, val|
457 if key.intern == :domain
462 debug "cookie domain #{domain} / #{new_loc.host}"
463 if new_loc.host.rindex(domain) == new_loc.host.length - domain.length
464 debug "setting cookie"
465 new_opts[:headers] ||= Hash.new
466 new_opts[:headers]['Cookie'] = cookie
468 debug "cookie is for another domain, ignoring"
471 debug "following the redirect to #{new_loc}"
472 return get_response(new_loc, new_opts, &block)
474 warning ":| redirect w/o location?"
479 alias :body :cooked_body
481 unless resp['content-type']
482 debug "No content type, guessing"
483 resp['content-type'] =
484 case resp['x-rbot-location']
490 'application/xml+xhtml'
491 when /.(gif|png|jpe?g|jp2|tiff?)$/i
492 "image/#{$1.sub(/^jpg$/,'jpeg').sub(/^tif$/,'tiff')}"
494 'application/octetstream'
500 # Net::HTTP wants us to read the whole body here
506 # _uri_:: uri to query (URI object or String)
508 # Generic http transaction method. It will return a Net::HTTPResponse
509 # object or raise an exception
511 # If a block is given, it will yield the response (see :yield option)
513 # Currently supported _options_:
515 # method:: request method [:get (default), :post or :head]
516 # open_timeout:: open timeout for the proxy
517 # read_timeout:: read timeout for the proxy
518 # cache:: should we cache results?
519 # yield:: if :final [default], calls the block for the response object;
520 # if :all, call the block for all intermediate redirects, too
521 # max_redir:: how many redirects to follow before raising the exception
522 # if -1, don't follow redirects, just return them
523 # range:: make a ranged request (usually GET). accepts a string
524 # for HTTP/1.1 "Range:" header (i.e. "bytes=0-1000")
525 # body:: request body (usually for POST requests)
526 # headers:: additional headers to be set for the request. Its value must
527 # be a Hash in the form { 'Header' => 'value' }
529 def get_response(uri_or_s, options = {}, &block) # :yields: resp
530 uri = uri_or_s.kind_of?(URI) ? uri_or_s : URI.parse(uri_or_s.to_s)
531 unless URI::HTTP === uri
533 raise "#{uri.scheme.inspect} URI scheme is not supported"
535 raise "don't know what to do with #{uri.to_s.inspect}"
540 :max_redir => @bot.config['http.max_redir'],
546 req_class = case opts[:method].to_s.downcase.intern
547 when :head, :"net::http::head"
548 opts[:max_redir] = -1
550 when :get, :"net::http::get"
552 when :post, :"net::http::post"
554 opts[:body] or raise 'post request w/o a body?'
555 warning "refusing to cache POST request" if options[:cache]
558 warning "unsupported method #{opts[:method]}, doing GET"
562 if req_class != Net::HTTP::Get && opts[:range]
563 warning "can't request ranges for #{req_class}"
567 cache_key = "#{opts[:range]}|#{req_class}|#{uri.to_s}"
569 if req_class != Net::HTTP::Get && req_class != Net::HTTP::Head
571 warning "can't cache #{req_class.inspect} requests, working w/o cache"
576 debug "get_response(#{uri}, #{opts.inspect})"
578 cached = @cache[cache_key]
580 if opts[:cache] && cached
585 return handle_response(uri, cached.response, opts, &block)
589 headers = @headers.dup.merge(opts[:headers] || {})
590 headers['Range'] = opts[:range] if opts[:range]
591 headers['Authorization'] = opts[:auth_head] if opts[:auth_head]
593 if opts[:cache] && cached && (req_class == Net::HTTP::Get)
594 cached.setup_headers headers
597 req = req_class.new(uri.request_uri, headers)
598 if uri.user && uri.password
599 req.basic_auth(uri.user, uri.password)
600 opts[:auth_head] = req['Authorization']
602 req.body = opts[:body] if req_class == Net::HTTP::Post
603 debug "prepared request: #{req.to_hash.inspect}"
606 get_proxy(uri, opts).start do |http|
607 http.request(req) do |resp|
608 resp['x-rbot-location'] = uri.to_s
609 if Net::HTTPNotModified === resp
612 cached.revalidate(resp)
613 rescue Exception => e
616 debug "reusing cached"
617 resp = cached.response
618 elsif Net::HTTPServerError === resp || Net::HTTPClientError === resp
619 debug "http error, deleting cached obj" if cached
620 @cache.delete(cache_key)
624 return handle_response(uri, resp, opts, &block)
626 if cached = CachedObject.maybe_new(resp) rescue nil
627 debug "storing to cache"
628 @cache[cache_key] = cached
633 rescue Exception => e
639 # _uri_:: uri to query (URI object or String)
641 # Simple GET request, returns (if possible) response body following redirs
642 # and caching if requested, yielding the actual response(s) to the optional
643 # block. See get_response for details on the supported _options_
645 def get(uri, options = {}, &block) # :yields: resp
647 resp = get_response(uri, options, &block)
648 raise "http error: #{resp}" unless Net::HTTPOK === resp ||
649 Net::HTTPPartialContent === resp
651 rescue Exception => e
657 # _uri_:: uri to query (URI object or String)
659 # Simple HEAD request, returns (if possible) response head following redirs
660 # and caching if requested, yielding the actual response(s) to the optional
661 # block. See get_response for details on the supported _options_
663 def head(uri, options = {}, &block) # :yields: resp
664 opts = {:method => :head}.merge(options)
666 resp = get_response(uri, opts, &block)
667 # raise "http error #{resp}" if Net::HTTPClientError === resp ||
668 # Net::HTTPServerError == resp
670 rescue Exception => e
676 # _uri_:: uri to query (URI object or String)
677 # _data_:: body of the POST
679 # Simple POST request, returns (if possible) response following redirs and
680 # caching if requested, yielding the response(s) to the optional block. See
681 # get_response for details on the supported _options_
683 def post(uri, data, options = {}, &block) # :yields: resp
684 opts = {:method => :post, :body => data, :cache => false}.merge(options)
686 resp = get_response(uri, opts, &block)
687 raise 'http error' unless Net::HTTPOK === resp or Net::HTTPCreated === resp
689 rescue Exception => e
695 # _uri_:: uri to query (URI object or String)
696 # _nbytes_:: number of bytes to get
698 # Partial GET request, returns (if possible) the first _nbytes_ bytes of the
699 # response body, following redirs and caching if requested, yielding the
700 # actual response(s) to the optional block. See get_response for details on
701 # the supported _options_
703 def get_partial(uri, nbytes = @bot.config['http.info_bytes'], options = {}, &block) # :yields: resp
704 opts = {:range => "bytes=0-#{nbytes}"}.merge(options)
705 return get(uri, opts, &block)
708 def remove_stale_cache
709 debug "Removing stale cache"
711 max_last = @bot.config['http.expire_time'] * 60
712 max_first = @bot.config['http.max_cache_time'] * 60
713 debug "#{@cache.size} pages before"
715 @cache.reject! { |k, val|
716 (now - val.last_used > max_last) || (now - val.first_used > max_first)
719 error "Failed to remove stale cache: #{e.pretty_inspect}"
721 debug "#{@cache.size} pages after"
728 class HttpUtilPlugin < CoreBotModule
731 debug 'initializing httputil'
732 @bot.httputil = Irc::Utils::HttpUtil.new(@bot)
736 debug 'shutting down httputil'
737 @bot.httputil.cleanup