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
19 error "Couldn't load 'net/https': #{e.inspect}"
20 error "Secured HTTP connections will fail"
25 # Read chunks from the body until we have at least _size_ bytes, yielding
26 # the partial text at each chunk. Return the partial body.
27 def partial_body(size=0, &block)
31 self.read_body { |chunk|
33 yield partial if block_given?
34 break if size and size > 0 and partial.length >= size
47 # class for making http requests easier (mainly for plugins to use)
48 # this class can check the bot proxy configuration to determine if a proxy
49 # needs to be used, which includes support for per-url proxy configuration.
51 BotConfig.register BotConfigBooleanValue.new('http.use_proxy',
52 :default => false, :desc => "should a proxy be used for HTTP requests?")
53 BotConfig.register BotConfigStringValue.new('http.proxy_uri', :default => false,
54 :desc => "Proxy server to use for HTTP requests (URI, e.g http://proxy.host:port)")
55 BotConfig.register BotConfigStringValue.new('http.proxy_user',
57 :desc => "User for authenticating with the http proxy (if required)")
58 BotConfig.register BotConfigStringValue.new('http.proxy_pass',
60 :desc => "Password for authenticating with the http proxy (if required)")
61 BotConfig.register BotConfigArrayValue.new('http.proxy_include',
63 :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")
64 BotConfig.register BotConfigArrayValue.new('http.proxy_exclude',
66 :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")
67 BotConfig.register BotConfigIntegerValue.new('http.max_redir',
69 :desc => "Maximum number of redirections to be used when getting a document")
70 BotConfig.register BotConfigIntegerValue.new('http.expire_time',
72 :desc => "After how many minutes since last use a cached document is considered to be expired")
73 BotConfig.register BotConfigIntegerValue.new('http.max_cache_time',
75 :desc => "After how many minutes since first use a cached document is considered to be expired")
76 BotConfig.register BotConfigIntegerValue.new('http.no_expire_cache',
78 :desc => "Set this to true if you want the bot to never expire the cached pages")
79 BotConfig.register BotConfigIntegerValue.new('http.info_bytes',
81 :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.")
84 attr_accessor :response, :last_used, :first_used, :count, :expires, :date
86 def self.maybe_new(resp)
87 debug "maybe new #{resp}"
88 return nil unless Net::HTTPOK === resp ||
89 Net::HTTPMovedPermanently === resp ||
90 Net::HTTPFound === resp ||
91 Net::HTTPPartialContent === resp
93 cc = resp['cache-control']
94 return nil if cc && (cc =~ /no-cache/i)
98 date = Time.httpdate(d)
101 return nil if resp['expires'] && (Time.httpdate(resp['expires']) < date)
103 debug "creating cache obj"
110 @first_used = now if @count == 0
116 debug "checking expired?"
117 if cc = self.response['cache-control'] && cc =~ /must-revalidate/
120 return self.expires < Time.now
123 def setup_headers(hdr)
124 hdr['if-modified-since'] = self.date.rfc2822
126 debug "ims == #{hdr['if-modified-since']}"
128 if etag = self.response['etag']
129 hdr['if-none-match'] = etag
130 debug "etag: #{etag}"
134 def revalidate(resp = self.response)
137 self.date = resp.key?('date') ? Time.httpdate(resp['date']) : Time.now
139 cc = resp['cache-control']
140 if cc && (cc =~ /max-age=(\d+)/)
141 self.expires = self.date + $1.to_i
142 elsif resp.key?('expires')
143 self.expires = Time.httpdate(resp['expires'])
144 elsif lm = resp['last-modified']
145 delta = self.date - Time.httpdate(lm)
146 delta = 10 if delta <= 0
148 self.expires = self.date + delta
150 self.expires = self.date + 300
152 # self.expires = Time.now + 10 # DEBUG
153 debug "expires on #{self.expires}"
164 rescue Exception => e
166 error e.backtrace.join("\n")
176 'Accept-Charset' => 'utf-8;q=1.0, *;q=0.8',
178 "rbot http util #{$version} (http://linuxbrit.co.uk/rbot/)"
180 debug "starting http cache cleanup timer"
181 @timer = @bot.timer.add(300) {
182 self.remove_stale_cache unless @bot.config['http.no_expire_cache']
187 debug 'stopping http cache cleanup timer'
188 @bot.timer.remove(@timer)
191 # if http_proxy_include or http_proxy_exclude are set, then examine the
192 # uri to see if this is a proxied uri
193 # the in/excludes are a list of regexps, and each regexp is checked against
194 # the server name, and its IP addresses
195 def proxy_required(uri)
197 if @bot.config["http.proxy_exclude"].empty? && @bot.config["http.proxy_include"].empty?
203 list.concat Resolv.getaddresses(uri.host)
204 rescue StandardError => err
205 warning "couldn't resolve host uri.host"
208 unless @bot.config["http.proxy_exclude"].empty?
209 re = @bot.config["http.proxy_exclude"].collect{|r| Regexp.new(r)}
219 unless @bot.config["http.proxy_include"].empty?
220 re = @bot.config["http.proxy_include"].collect{|r| Regexp.new(r)}
230 debug "using proxy for uri #{uri}?: #{use_proxy}"
234 # uri:: Uri to create a proxy for
236 # return a net/http Proxy object, which is configured correctly for
237 # proxying based on the bot's proxy configuration.
238 # This will include per-url proxy configuration based on the bot config
239 # +http_proxy_include/exclude+ options.
241 def get_proxy(uri, options = {})
253 if @bot.config["http.use_proxy"]
254 if (ENV['http_proxy'])
255 proxy = URI.parse ENV['http_proxy'] rescue nil
257 if (@bot.config["http.proxy_uri"])
258 proxy = URI.parse @bot.config["http.proxy_uri"] rescue nil
261 debug "proxy is set to #{proxy.host} port #{proxy.port}"
262 if proxy_required(uri)
263 proxy_host = proxy.host
264 proxy_port = proxy.port
265 proxy_user = @bot.config["http.proxy_user"]
266 proxy_pass = @bot.config["http.proxy_pass"]
271 h = Net::HTTP.new(uri.host, uri.port, proxy_host, proxy_port, proxy_user, proxy_port)
272 h.use_ssl = true if uri.scheme == "https"
274 h.read_timeout = opts[:read_timeout]
275 h.open_timeout = opts[:open_timeout]
279 def handle_response(uri, resp, opts, &block)
280 if Net::HTTPRedirection === resp && opts[:max_redir] >= 0
281 if resp.key?('location')
282 raise 'Too many redirections' if opts[:max_redir] <= 0
283 yield resp if opts[:yield] == :all && block_given?
284 loc = resp['location']
285 new_loc = URI.join(uri.to_s, loc) rescue URI.parse(loc)
287 new_opts[:max_redir] -= 1
288 case opts[:method].to_s.downcase.intern
289 when :post, :"net::http::post"
290 new_opts[:method] = :get
292 debug "following the redirect to #{new_loc}"
293 return get_response(new_loc, new_opts, &block)
295 warning ":| redirect w/o location?"
307 # uri:: uri to query (Uri object or String)
308 # opts:: options. Currently used:
309 # :method:: request method [:get (default), :post or :head]
310 # :open_timeout:: open timeout for the proxy
311 # :read_timeout:: read timeout for the proxy
312 # :cache:: should we cache results?
313 # :yield:: if :final [default], call &block for the response object
314 # if :all, call &block for all intermediate redirects, too
315 # :max_redir:: how many redirects to follow before raising the exception
316 # if -1, don't follow redirects, just return them
317 # :range:: make a ranged request (usually GET). accepts a string
318 # for HTTP/1.1 "Range:" header (i.e. "bytes=0-1000")
319 # :body:: request body (usually for POST requests)
321 # Generic http transaction method
323 # It will return a HTTP::Response object or raise an exception
325 # If a block is given, it will yield the response (see :yield option)
327 def get_response(uri_or_s, options = {}, &block)
328 uri = uri_or_s.kind_of?(URI) ? uri_or_s : URI.parse(uri_or_s.to_s)
330 :max_redir => @bot.config['http.max_redir'],
339 req_class = case opts[:method].to_s.downcase.intern
340 when :head, :"net::http::head"
341 opts[:max_redir] = -1
343 when :get, :"net::http::get"
345 when :post, :"net::http::post"
347 opts[:body] or raise 'post request w/o a body?'
348 warning "refusing to cache POST request" if options[:cache]
351 warning "unsupported method #{opts[:method]}, doing GET"
355 if req_class != Net::HTTP::Get && opts[:range]
356 warning "can't request ranges for #{req_class}"
360 cache_key = "#{opts[:range]}|#{req_class}|#{uri.to_s}"
362 if req_class != Net::HTTP::Get && req_class != Net::HTTP::Head
364 warning "can't cache #{req_class.inspect} requests, working w/o cache"
369 debug "get_response(#{uri}, #{opts.inspect})"
371 if opts[:cache] && cached = @cache[cache_key]
376 return handle_response(uri, cached.response, opts, &block)
380 headers = @headers.dup.merge(opts[:headers] || {})
381 headers['Range'] = opts[:range] if opts[:range]
383 cached.setup_headers(headers) if cached && (req_class == Net::HTTP::Get)
384 req = req_class.new(uri.request_uri, headers)
385 req.basic_auth(uri.user, uri.password) if uri.user && uri.password
386 req.body = opts[:body] if req_class == Net::HTTP::Post
387 debug "prepared request: #{req.to_hash.inspect}"
389 get_proxy(uri, opts).start do |http|
390 http.request(req) do |resp|
391 if Net::HTTPNotModified === resp
394 cached.revalidate(resp)
395 rescue Exception => e
397 error e.backtrace.join("\n")
399 debug "reusing cached"
400 resp = cached.response
401 elsif Net::HTTPServerError === resp || Net::HTTPClientError === resp
402 debug "http error, deleting cached obj" if cached
403 @cache.delete(cache_key)
404 elsif opts[:cache] && cached = CachedObject.maybe_new(resp) rescue nil
405 debug "storing to cache"
406 @cache[cache_key] = cached
408 return handle_response(uri, resp, opts, &block)
413 # uri:: uri to query (Uri object)
415 # simple get request, returns (if possible) response body following redirs
416 # and caching if requested
417 def get(uri, opts = {}, &block)
419 resp = get_response(uri, opts, &block)
420 raise "http error: #{resp}" unless Net::HTTPOK === resp ||
421 Net::HTTPPartialContent === resp
423 rescue Exception => e
425 error e.backtrace.join("\n")
430 def head(uri, options = {}, &block)
431 opts = {:method => :head}.merge(options)
433 resp = get_response(uri, opts, &block)
434 raise "http error #{resp}" if Net::HTTPClientError === resp ||
435 Net::HTTPServerError == resp
437 rescue Exception => e
439 error e.backtrace.join("\n")
444 def post(uri, data, options = {}, &block)
445 opts = {:method => :post, :body => data, :cache => false}.merge(options)
447 resp = get_response(uri, opts, &block)
448 raise 'http error' unless Net::HTTPOK === resp
450 rescue Exception => e
452 error e.backtrace.join("\n")
457 def get_partial(uri, nbytes = @bot.config['http.info_bytes'], options = {}, &block)
458 opts = {:range => "bytes=0-#{nbytes}"}.merge(options)
459 return get(uri, opts, &block)
462 def remove_stale_cache
463 debug "Removing stale cache"
465 max_last = @bot.config['http.expire_time'] * 60
466 max_first = @bot.config['http.max_cache_time'] * 60
467 debug "#{@cache.size} pages before"
469 @cache.reject! { |k, val|
470 (now - val.last_used > max_last) || (now - val.first_used > max_first)
473 error "Failed to remove stale cache: #{e.inspect}"
475 debug "#{@cache.size} pages after"
482 class HttpUtilPlugin < CoreBotModule
485 debug 'initializing httputil'
486 @bot.httputil = Irc::Utils::HttpUtil.new(@bot)
490 debug 'shutting down httputil'
491 @bot.httputil.cleanup