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"
28 # class for making http requests easier (mainly for plugins to use)
29 # this class can check the bot proxy configuration to determine if a proxy
30 # needs to be used, which includes support for per-url proxy configuration.
32 BotConfig.register BotConfigBooleanValue.new('http.use_proxy',
33 :default => false, :desc => "should a proxy be used for HTTP requests?")
34 BotConfig.register BotConfigStringValue.new('http.proxy_uri', :default => false,
35 :desc => "Proxy server to use for HTTP requests (URI, e.g http://proxy.host:port)")
36 BotConfig.register BotConfigStringValue.new('http.proxy_user',
38 :desc => "User for authenticating with the http proxy (if required)")
39 BotConfig.register BotConfigStringValue.new('http.proxy_pass',
41 :desc => "Password for authenticating with the http proxy (if required)")
42 BotConfig.register BotConfigArrayValue.new('http.proxy_include',
44 :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")
45 BotConfig.register BotConfigArrayValue.new('http.proxy_exclude',
47 :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")
48 BotConfig.register BotConfigIntegerValue.new('http.max_redir',
50 :desc => "Maximum number of redirections to be used when getting a document")
51 BotConfig.register BotConfigIntegerValue.new('http.expire_time',
53 :desc => "After how many minutes since last use a cached document is considered to be expired")
54 BotConfig.register BotConfigIntegerValue.new('http.max_cache_time',
56 :desc => "After how many minutes since first use a cached document is considered to be expired")
57 BotConfig.register BotConfigIntegerValue.new('http.no_expire_cache',
59 :desc => "Set this to true if you want the bot to never expire the cached pages")
60 BotConfig.register BotConfigIntegerValue.new('http.info_bytes',
62 :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.")
65 attr_accessor :response, :last_used, :first_used, :count, :expires, :date
67 def self.maybe_new(resp)
68 debug "maybe new #{resp}"
69 return nil unless Net::HTTPOK === resp ||
70 Net::HTTPMovedPermanently === resp ||
71 Net::HTTPFound === resp ||
72 Net::HTTPPartialContent === resp
74 cc = resp['cache-control']
75 return nil if cc && (cc =~ /no-cache/i)
79 date = Time.httpdate(d)
82 return nil if resp['expires'] && (Time.httpdate(resp['expires']) < date)
84 debug "creating cache obj"
91 @first_used = now if @count == 0
97 debug "checking expired?"
98 if cc = self.response['cache-control'] && cc =~ /must-revalidate/
101 return self.expires < Time.now
104 def setup_headers(hdr)
105 hdr['if-modified-since'] = self.date.rfc2822
107 debug "ims == #{hdr['if-modified-since']}"
109 if etag = self.response['etag']
110 hdr['if-none-match'] = etag
111 debug "etag: #{etag}"
115 def revalidate(resp = self.response)
118 self.date = resp.key?('date') ? Time.httpdate(resp['date']) : Time.now
120 cc = resp['cache-control']
121 if cc && (cc =~ /max-age=(\d+)/)
122 self.expires = self.date + $1.to_i
123 elsif resp.key?('expires')
124 self.expires = Time.httpdate(resp['expires'])
125 elsif lm = resp['last-modified']
126 delta = self.date - Time.httpdate(lm)
127 delta = 10 if delta <= 0
129 self.expires = self.date + delta
131 self.expires = self.date + 300
133 # self.expires = Time.now + 10 # DEBUG
134 debug "expires on #{self.expires}"
145 rescue Exception => e
147 error e.backtrace.join("\n")
157 'Accept-Charset' => 'utf-8;q=1.0, *;q=0.8',
159 "rbot http util #{$version} (http://linuxbrit.co.uk/rbot/)"
161 debug "starting http cache cleanup timer"
162 @timer = @bot.timer.add(300) {
163 self.remove_stale_cache unless @bot.config['http.no_expire_cache']
168 debug 'stopping http cache cleanup timer'
169 @bot.timer.remove(@timer)
172 # if http_proxy_include or http_proxy_exclude are set, then examine the
173 # uri to see if this is a proxied uri
174 # the in/excludes are a list of regexps, and each regexp is checked against
175 # the server name, and its IP addresses
176 def proxy_required(uri)
178 if @bot.config["http.proxy_exclude"].empty? && @bot.config["http.proxy_include"].empty?
184 list.concat Resolv.getaddresses(uri.host)
185 rescue StandardError => err
186 warning "couldn't resolve host uri.host"
189 unless @bot.config["http.proxy_exclude"].empty?
190 re = @bot.config["http.proxy_exclude"].collect{|r| Regexp.new(r)}
200 unless @bot.config["http.proxy_include"].empty?
201 re = @bot.config["http.proxy_include"].collect{|r| Regexp.new(r)}
211 debug "using proxy for uri #{uri}?: #{use_proxy}"
215 # uri:: Uri to create a proxy for
217 # return a net/http Proxy object, which is configured correctly for
218 # proxying based on the bot's proxy configuration.
219 # This will include per-url proxy configuration based on the bot config
220 # +http_proxy_include/exclude+ options.
222 def get_proxy(uri, options = {})
234 if @bot.config["http.use_proxy"]
235 if (ENV['http_proxy'])
236 proxy = URI.parse ENV['http_proxy'] rescue nil
238 if (@bot.config["http.proxy_uri"])
239 proxy = URI.parse @bot.config["http.proxy_uri"] rescue nil
242 debug "proxy is set to #{proxy.host} port #{proxy.port}"
243 if proxy_required(uri)
244 proxy_host = proxy.host
245 proxy_port = proxy.port
246 proxy_user = @bot.config["http.proxy_user"]
247 proxy_pass = @bot.config["http.proxy_pass"]
252 h = Net::HTTP.new(uri.host, uri.port, proxy_host, proxy_port, proxy_user, proxy_port)
253 h.use_ssl = true if uri.scheme == "https"
255 h.read_timeout = opts[:read_timeout]
256 h.open_timeout = opts[:open_timeout]
260 def handle_response(uri, resp, opts, &block)
261 if Net::HTTPRedirection === resp && opts[:max_redir] >= 0
262 if resp.key?('location')
263 raise 'Too many redirections' if opts[:max_redir] <= 0
264 yield resp if opts[:yield] == :all && block_given?
265 loc = resp['location']
266 new_loc = URI.join(uri.to_s, loc) rescue URI.parse(loc)
268 new_opts[:max_redir] -= 1
269 case opts[:method].to_s.downcase.intern
270 when :post, :"net::http::post"
271 new_opts[:method] = :get
273 debug "following the redirect to #{new_loc}"
274 return get_response(new_loc, new_opts, &block)
276 warning ":| redirect w/o location?"
288 # uri:: uri to query (Uri object or String)
289 # opts:: options. Currently used:
290 # :method:: request method [:get (default), :post or :head]
291 # :open_timeout:: open timeout for the proxy
292 # :read_timeout:: read timeout for the proxy
293 # :cache:: should we cache results?
294 # :yield:: if :final [default], call &block for the response object
295 # if :all, call &block for all intermediate redirects, too
296 # :max_redir:: how many redirects to follow before raising the exception
297 # if -1, don't follow redirects, just return them
298 # :range:: make a ranged request (usually GET). accepts a string
299 # for HTTP/1.1 "Range:" header (i.e. "bytes=0-1000")
300 # :body:: request body (usually for POST requests)
302 # Generic http transaction method
304 # It will return a HTTP::Response object or raise an exception
306 # If a block is given, it will yield the response (see :yield option)
308 def get_response(uri_or_s, options = {}, &block)
309 uri = uri_or_s.kind_of?(URI) ? uri_or_s : URI.parse(uri_or_s.to_s)
311 :max_redir => @bot.config['http.max_redir'],
320 req_class = case opts[:method].to_s.downcase.intern
321 when :head, :"net::http::head"
322 opts[:max_redir] = -1
324 when :get, :"net::http::get"
326 when :post, :"net::http::post"
328 opts[:body] or raise 'post request w/o a body?'
329 warning "refusing to cache POST request" if options[:cache]
332 warning "unsupported method #{opts[:method]}, doing GET"
336 if req_class != Net::HTTP::Get && opts[:range]
337 warning "can't request ranges for #{req_class}"
341 cache_key = "#{opts[:range]}|#{req_class}|#{uri.to_s}"
343 if req_class != Net::HTTP::Get && req_class != Net::HTTP::Head
345 warning "can't cache #{req_class.inspect} requests, working w/o cache"
350 debug "get_response(#{uri}, #{opts.inspect})"
352 if opts[:cache] && cached = @cache[cache_key]
357 return handle_response(uri, cached.response, opts, &block)
361 headers = @headers.dup.merge(opts[:headers] || {})
362 headers['Range'] = opts[:range] if opts[:range]
364 cached.setup_headers(headers) if cached && (req_class == Net::HTTP::Get)
365 req = req_class.new(uri.request_uri, headers)
366 req.basic_auth(uri.user, uri.password) if uri.user && uri.password
367 req.body = opts[:body] if req_class == Net::HTTP::Post
368 debug "prepared request: #{req.to_hash.inspect}"
370 get_proxy(uri, opts).start do |http|
371 http.request(req) do |resp|
372 if Net::HTTPNotModified === resp
375 cached.revalidate(resp)
376 rescue Exception => e
378 error e.backtrace.join("\n")
380 debug "reusing cached"
381 resp = cached.response
382 elsif Net::HTTPServerError === resp || Net::HTTPClientError === resp
383 debug "http error, deleting cached obj" if cached
384 @cache.delete(cache_key)
385 elsif opts[:cache] && cached = CachedObject.maybe_new(resp) rescue nil
386 debug "storing to cache"
387 @cache[cache_key] = cached
389 return handle_response(uri, resp, opts, &block)
394 # uri:: uri to query (Uri object)
396 # simple get request, returns (if possible) response body following redirs
397 # and caching if requested
398 def get(uri, opts = {}, &block)
400 resp = get_response(uri, opts, &block)
401 raise "http error: #{resp}" unless Net::HTTPOK === resp ||
402 Net::HTTPPartialContent === resp
404 rescue Exception => e
406 error e.backtrace.join("\n")
411 def head(uri, options = {}, &block)
412 opts = {:method => :head}.merge(options)
414 resp = get_response(uri, opts, &block)
415 raise "http error #{resp}" if Net::HTTPClientError === resp ||
416 Net::HTTPServerError == resp
418 rescue Exception => e
420 error e.backtrace.join("\n")
425 def post(uri, data, options = {}, &block)
426 opts = {:method => :post, :body => data, :cache => false}.merge(options)
428 resp = get_response(uri, opts, &block)
429 raise 'http error' unless Net::HTTPOK === resp
431 rescue Exception => e
433 error e.backtrace.join("\n")
438 def get_partial(uri, nbytes = @bot.config['http.info_bytes'], options = {}, &block)
439 opts = {:range => "bytes=0-#{nbytes}"}.merge(options)
440 return get(uri, opts, &block)
443 def remove_stale_cache
444 debug "Removing stale cache"
446 max_last = @bot.config['http.expire_time'] * 60
447 max_first = @bot.config['http.max_cache_time'] * 60
448 debug "#{@cache.size} pages before"
450 @cache.reject! { |k, val|
451 (now - val.last_used > max_last) || (now - val.first_used > max_first)
454 error "Failed to remove stale cache: #{e.inspect}"
456 debug "#{@cache.size} pages after"
463 class HttpUtilPlugin < CoreBotModule
466 debug 'initializing httputil'
467 @bot.httputil = Irc::Utils::HttpUtil.new(@bot)
471 debug 'shutting down httputil'
472 @bot.httputil.cleanup