4 # :title: rbot HTTP provider
6 # Author:: Tom Gilbert <tom@linuxbrit.co.uk>
7 # Author:: Giuseppe "Oblomov" Bilotta <giuseppe.bilotta@gmail.com>
8 # Author:: Dmitry "jsn" Kim <dmitry point kim at gmail point com>
10 # Copyright:: (C) 2002-2005 Tom Gilbert
11 # Copyright:: (C) 2006 Tom Gilbert, Giuseppe Bilotta
12 # Copyright:: (C) 2007 Giuseppe Bilotta, Dmitry Kim
21 error "Couldn't load 'net/https': #{e.pretty_inspect}"
22 error "Secured HTTP connections will fail"
25 # To handle Gzipped pages
31 attr_accessor :no_cache
32 if !instance_methods.include?('raw_body')
36 def body_charset(str=self.raw_body)
37 ctype = self['content-type'] || 'text/html'
38 return nil unless ctype =~ /^text/i || ctype =~ /x(ht)?ml/i
40 charsets = ['latin1'] # should be in config
42 if self['content-type'].match(/charset=["']?([^\s"']+)["']?/i)
44 debug "charset #{charsets.last} added from header"
48 when /<\?xml\s[^>]*encoding=['"]([^\s"'>]+)["'][^>]*\?>/i
50 debug "xml charset #{charsets.last} added from xml pi"
51 when /<(meta\s[^>]*http-equiv=["']?Content-Type["']?[^>]*)>/i
53 if meta =~ /charset=['"]?([^\s'";]+)['"]?/
55 debug "html charset #{charsets.last} added from meta"
62 charsets = self.body_charset(str) or return str
64 charsets.reverse_each { |charset|
66 return Iconv.iconv('utf-8//ignore', charset, str).first
68 debug "conversion failed for #{charset}"
74 def decompress_body(str)
75 method = self['content-encoding']
80 debug "gunzipping body"
82 return Zlib::GzipReader.new(StringIO.new(str)).read
83 rescue Zlib::Error => e
84 # If we can't unpack the whole stream (e.g. because we're doing a
86 debug "full gunzipping failed (#{e}), trying to recover as much as possible"
89 Zlib::GzipReader.new(StringIO.new(str)).each_byte { |byte|
97 raise "Unhandled content encoding #{method}"
102 return self.body_to_utf(self.decompress_body(self.raw_body))
105 # Read chunks from the body until we have at least _size_ bytes, yielding
106 # the partial text at each chunk. Return the partial body.
107 def partial_body(size=0, &block)
112 self.read_body { |chunk|
114 yield self.body_to_utf(self.decompress_body(partial)) if block_given?
115 break if size and size > 0 and partial.length >= size
118 return self.body_to_utf(self.decompress_body(partial))
123 Net::HTTP.version_1_2
128 # class for making http requests easier (mainly for plugins to use)
129 # this class can check the bot proxy configuration to determine if a proxy
130 # needs to be used, which includes support for per-url proxy configuration.
132 BotConfig.register BotConfigBooleanValue.new('http.use_proxy',
133 :default => false, :desc => "should a proxy be used for HTTP requests?")
134 BotConfig.register BotConfigStringValue.new('http.proxy_uri', :default => false,
135 :desc => "Proxy server to use for HTTP requests (URI, e.g http://proxy.host:port)")
136 BotConfig.register BotConfigStringValue.new('http.proxy_user',
138 :desc => "User for authenticating with the http proxy (if required)")
139 BotConfig.register BotConfigStringValue.new('http.proxy_pass',
141 :desc => "Password for authenticating with the http proxy (if required)")
142 BotConfig.register BotConfigArrayValue.new('http.proxy_include',
144 :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")
145 BotConfig.register BotConfigArrayValue.new('http.proxy_exclude',
147 :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")
148 BotConfig.register BotConfigIntegerValue.new('http.max_redir',
150 :desc => "Maximum number of redirections to be used when getting a document")
151 BotConfig.register BotConfigIntegerValue.new('http.expire_time',
153 :desc => "After how many minutes since last use a cached document is considered to be expired")
154 BotConfig.register BotConfigIntegerValue.new('http.max_cache_time',
156 :desc => "After how many minutes since first use a cached document is considered to be expired")
157 BotConfig.register BotConfigIntegerValue.new('http.no_expire_cache',
159 :desc => "Set this to true if you want the bot to never expire the cached pages")
160 BotConfig.register BotConfigIntegerValue.new('http.info_bytes',
162 :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.")
165 attr_accessor :response, :last_used, :first_used, :count, :expires, :date
167 def self.maybe_new(resp)
168 debug "maybe new #{resp}"
169 return nil if resp.no_cache
170 return nil unless Net::HTTPOK === resp ||
171 Net::HTTPMovedPermanently === resp ||
172 Net::HTTPFound === resp ||
173 Net::HTTPPartialContent === resp
175 cc = resp['cache-control']
176 return nil if cc && (cc =~ /no-cache/i)
180 date = Time.httpdate(d)
183 return nil if resp['expires'] && (Time.httpdate(resp['expires']) < date)
185 debug "creating cache obj"
192 @first_used = now if @count == 0
198 debug "checking expired?"
199 if cc = self.response['cache-control'] && cc =~ /must-revalidate/
202 return self.expires < Time.now
205 def setup_headers(hdr)
206 hdr['if-modified-since'] = self.date.rfc2822
208 debug "ims == #{hdr['if-modified-since']}"
210 if etag = self.response['etag']
211 hdr['if-none-match'] = etag
212 debug "etag: #{etag}"
216 def revalidate(resp = self.response)
219 self.date = resp.key?('date') ? Time.httpdate(resp['date']) : Time.now
221 cc = resp['cache-control']
222 if cc && (cc =~ /max-age=(\d+)/)
223 self.expires = self.date + $1.to_i
224 elsif resp.key?('expires')
225 self.expires = Time.httpdate(resp['expires'])
226 elsif lm = resp['last-modified']
227 delta = self.date - Time.httpdate(lm)
228 delta = 10 if delta <= 0
230 self.expires = self.date + delta
232 self.expires = self.date + 300
234 # self.expires = Time.now + 10 # DEBUG
235 debug "expires on #{self.expires}"
245 self.response.raw_body
246 rescue Exception => e
257 'Accept-Charset' => 'utf-8;q=1.0, *;q=0.8',
258 'Accept-Encoding' => 'gzip;q=1, identity;q=0.8, *;q=0.2',
260 "rbot http util #{$version} (http://linuxbrit.co.uk/rbot/)"
262 debug "starting http cache cleanup timer"
263 @timer = @bot.timer.add(300) {
264 self.remove_stale_cache unless @bot.config['http.no_expire_cache']
269 debug 'stopping http cache cleanup timer'
270 @bot.timer.remove(@timer)
273 # if http_proxy_include or http_proxy_exclude are set, then examine the
274 # uri to see if this is a proxied uri
275 # the in/excludes are a list of regexps, and each regexp is checked against
276 # the server name, and its IP addresses
277 def proxy_required(uri)
279 if @bot.config["http.proxy_exclude"].empty? && @bot.config["http.proxy_include"].empty?
285 list.concat Resolv.getaddresses(uri.host)
286 rescue StandardError => err
287 warning "couldn't resolve host uri.host"
290 unless @bot.config["http.proxy_exclude"].empty?
291 re = @bot.config["http.proxy_exclude"].collect{|r| Regexp.new(r)}
301 unless @bot.config["http.proxy_include"].empty?
302 re = @bot.config["http.proxy_include"].collect{|r| Regexp.new(r)}
312 debug "using proxy for uri #{uri}?: #{use_proxy}"
316 # uri:: Uri to create a proxy for
318 # return a net/http Proxy object, which is configured correctly for
319 # proxying based on the bot's proxy configuration.
320 # This will include per-url proxy configuration based on the bot config
321 # +http_proxy_include/exclude+ options.
323 def get_proxy(uri, options = {})
335 if @bot.config["http.use_proxy"]
336 if (ENV['http_proxy'])
337 proxy = URI.parse ENV['http_proxy'] rescue nil
339 if (@bot.config["http.proxy_uri"])
340 proxy = URI.parse @bot.config["http.proxy_uri"] rescue nil
343 debug "proxy is set to #{proxy.host} port #{proxy.port}"
344 if proxy_required(uri)
345 proxy_host = proxy.host
346 proxy_port = proxy.port
347 proxy_user = @bot.config["http.proxy_user"]
348 proxy_pass = @bot.config["http.proxy_pass"]
353 h = Net::HTTP.new(uri.host, uri.port, proxy_host, proxy_port, proxy_user, proxy_port)
354 h.use_ssl = true if uri.scheme == "https"
356 h.read_timeout = opts[:read_timeout]
357 h.open_timeout = opts[:open_timeout]
361 def handle_response(uri, resp, opts, &block)
362 if Net::HTTPRedirection === resp && opts[:max_redir] >= 0
363 if resp.key?('location')
364 raise 'Too many redirections' if opts[:max_redir] <= 0
365 yield resp if opts[:yield] == :all && block_given?
366 loc = resp['location']
367 new_loc = URI.join(uri.to_s, loc) rescue URI.parse(loc)
369 new_opts[:max_redir] -= 1
370 case opts[:method].to_s.downcase.intern
371 when :post, :"net::http::post"
372 new_opts[:method] = :get
374 debug "following the redirect to #{new_loc}"
375 return get_response(new_loc, new_opts, &block)
377 warning ":| redirect w/o location?"
383 # Net::HTTP wants us to read the whole body here
389 # uri:: uri to query (Uri object or String)
390 # opts:: options. Currently used:
391 # :method:: request method [:get (default), :post or :head]
392 # :open_timeout:: open timeout for the proxy
393 # :read_timeout:: read timeout for the proxy
394 # :cache:: should we cache results?
395 # :yield:: if :final [default], call &block for the response object
396 # if :all, call &block for all intermediate redirects, too
397 # :max_redir:: how many redirects to follow before raising the exception
398 # if -1, don't follow redirects, just return them
399 # :range:: make a ranged request (usually GET). accepts a string
400 # for HTTP/1.1 "Range:" header (i.e. "bytes=0-1000")
401 # :body:: request body (usually for POST requests)
403 # Generic http transaction method
405 # It will return a HTTP::Response object or raise an exception
407 # If a block is given, it will yield the response (see :yield option)
409 def get_response(uri_or_s, options = {}, &block)
410 uri = uri_or_s.kind_of?(URI) ? uri_or_s : URI.parse(uri_or_s.to_s)
412 :max_redir => @bot.config['http.max_redir'],
421 req_class = case opts[:method].to_s.downcase.intern
422 when :head, :"net::http::head"
423 opts[:max_redir] = -1
425 when :get, :"net::http::get"
427 when :post, :"net::http::post"
429 opts[:body] or raise 'post request w/o a body?'
430 warning "refusing to cache POST request" if options[:cache]
433 warning "unsupported method #{opts[:method]}, doing GET"
437 if req_class != Net::HTTP::Get && opts[:range]
438 warning "can't request ranges for #{req_class}"
442 cache_key = "#{opts[:range]}|#{req_class}|#{uri.to_s}"
444 if req_class != Net::HTTP::Get && req_class != Net::HTTP::Head
446 warning "can't cache #{req_class.inspect} requests, working w/o cache"
451 debug "get_response(#{uri}, #{opts.inspect})"
453 if opts[:cache] && cached = @cache[cache_key]
458 return handle_response(uri, cached.response, opts, &block)
462 headers = @headers.dup.merge(opts[:headers] || {})
463 headers['Range'] = opts[:range] if opts[:range]
465 cached.setup_headers(headers) if cached && (req_class == Net::HTTP::Get)
466 req = req_class.new(uri.request_uri, headers)
467 req.basic_auth(uri.user, uri.password) if uri.user && uri.password
468 req.body = opts[:body] if req_class == Net::HTTP::Post
469 debug "prepared request: #{req.to_hash.inspect}"
471 get_proxy(uri, opts).start do |http|
472 http.request(req) do |resp|
473 resp['x-rbot-location'] = uri.to_s
474 if Net::HTTPNotModified === resp
477 cached.revalidate(resp)
478 rescue Exception => e
481 debug "reusing cached"
482 resp = cached.response
483 elsif Net::HTTPServerError === resp || Net::HTTPClientError === resp
484 debug "http error, deleting cached obj" if cached
485 @cache.delete(cache_key)
488 return handle_response(uri, resp, opts, &block)
490 if cached = CachedObject.maybe_new(resp) rescue nil
491 debug "storing to cache"
492 @cache[cache_key] = cached
497 return handle_response(uri, resp, opts, &block)
502 # uri:: uri to query (Uri object)
504 # simple get request, returns (if possible) response body following redirs
505 # and caching if requested
506 def get(uri, opts = {}, &block)
508 resp = get_response(uri, opts, &block)
509 raise "http error: #{resp}" unless Net::HTTPOK === resp ||
510 Net::HTTPPartialContent === resp
512 rescue Exception => e
518 def head(uri, options = {}, &block)
519 opts = {:method => :head}.merge(options)
521 resp = get_response(uri, opts, &block)
522 raise "http error #{resp}" if Net::HTTPClientError === resp ||
523 Net::HTTPServerError == resp
525 rescue Exception => e
531 def post(uri, data, options = {}, &block)
532 opts = {:method => :post, :body => data, :cache => false}.merge(options)
534 resp = get_response(uri, opts, &block)
535 raise 'http error' unless Net::HTTPOK === resp
537 rescue Exception => e
543 def get_partial(uri, nbytes = @bot.config['http.info_bytes'], options = {}, &block)
544 opts = {:range => "bytes=0-#{nbytes}"}.merge(options)
545 return get(uri, opts, &block)
548 def remove_stale_cache
549 debug "Removing stale cache"
551 max_last = @bot.config['http.expire_time'] * 60
552 max_first = @bot.config['http.max_cache_time'] * 60
553 debug "#{@cache.size} pages before"
555 @cache.reject! { |k, val|
556 (now - val.last_used > max_last) || (now - val.first_used > max_first)
559 error "Failed to remove stale cache: #{e.pretty_inspect}"
561 debug "#{@cache.size} pages after"
568 class HttpUtilPlugin < CoreBotModule
571 debug 'initializing httputil'
572 @bot.httputil = Irc::Utils::HttpUtil.new(@bot)
576 debug 'shutting down httputil'
577 @bot.httputil.cleanup