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
20 error "Couldn't load 'net/https': #{e.inspect}"
21 error "Secured HTTP connections will fail"
26 attr_accessor :no_cache
27 if !instance_methods.include?('raw_body')
31 def body_charset(str=self.raw_body)
32 ctype = self['content-type'] || 'text/html'
33 return nil unless ctype =~ /^text/i || ctype =~ /x(ht)?ml/i
35 charsets = ['latin1'] # should be in config
37 if self['content-type'].match(/charset=["']?([^\s"']+)["']?/i)
39 debug "charset #{charsets.last} added from header"
43 when /<\?xml\s[^>]*encoding=['"]([^\s"'>]+)["'][^>]*\?>/i
45 debug "xml charset #{charsets.last} added from xml pi"
46 when /<(meta\s[^>]*http-equiv=["']?Content-Type["']?[^>]*)>/i
48 if meta =~ /charset=['"]?([^\s'";]+)['"]?/
50 debug "html charset #{charsets.last} added from meta"
57 charsets = self.body_charset(str) or return str
59 charsets.reverse_each { |charset|
61 return Iconv.iconv('utf-8//ignore', charset, str).first
63 debug "conversion failed for #{charset}"
70 return self.body_to_utf(self.raw_body)
73 # Read chunks from the body until we have at least _size_ bytes, yielding
74 # the partial text at each chunk. Return the partial body.
75 def partial_body(size=0, &block)
80 self.read_body { |chunk|
82 yield self.body_to_utf(partial) if block_given?
83 break if size and size > 0 and partial.length >= size
86 return self.body_to_utf(partial)
96 # class for making http requests easier (mainly for plugins to use)
97 # this class can check the bot proxy configuration to determine if a proxy
98 # needs to be used, which includes support for per-url proxy configuration.
100 BotConfig.register BotConfigBooleanValue.new('http.use_proxy',
101 :default => false, :desc => "should a proxy be used for HTTP requests?")
102 BotConfig.register BotConfigStringValue.new('http.proxy_uri', :default => false,
103 :desc => "Proxy server to use for HTTP requests (URI, e.g http://proxy.host:port)")
104 BotConfig.register BotConfigStringValue.new('http.proxy_user',
106 :desc => "User for authenticating with the http proxy (if required)")
107 BotConfig.register BotConfigStringValue.new('http.proxy_pass',
109 :desc => "Password for authenticating with the http proxy (if required)")
110 BotConfig.register BotConfigArrayValue.new('http.proxy_include',
112 :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")
113 BotConfig.register BotConfigArrayValue.new('http.proxy_exclude',
115 :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")
116 BotConfig.register BotConfigIntegerValue.new('http.max_redir',
118 :desc => "Maximum number of redirections to be used when getting a document")
119 BotConfig.register BotConfigIntegerValue.new('http.expire_time',
121 :desc => "After how many minutes since last use a cached document is considered to be expired")
122 BotConfig.register BotConfigIntegerValue.new('http.max_cache_time',
124 :desc => "After how many minutes since first use a cached document is considered to be expired")
125 BotConfig.register BotConfigIntegerValue.new('http.no_expire_cache',
127 :desc => "Set this to true if you want the bot to never expire the cached pages")
128 BotConfig.register BotConfigIntegerValue.new('http.info_bytes',
130 :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.")
133 attr_accessor :response, :last_used, :first_used, :count, :expires, :date
135 def self.maybe_new(resp)
136 debug "maybe new #{resp}"
137 return nil if resp.no_cache
138 return nil unless Net::HTTPOK === resp ||
139 Net::HTTPMovedPermanently === resp ||
140 Net::HTTPFound === resp ||
141 Net::HTTPPartialContent === resp
143 cc = resp['cache-control']
144 return nil if cc && (cc =~ /no-cache/i)
148 date = Time.httpdate(d)
151 return nil if resp['expires'] && (Time.httpdate(resp['expires']) < date)
153 debug "creating cache obj"
160 @first_used = now if @count == 0
166 debug "checking expired?"
167 if cc = self.response['cache-control'] && cc =~ /must-revalidate/
170 return self.expires < Time.now
173 def setup_headers(hdr)
174 hdr['if-modified-since'] = self.date.rfc2822
176 debug "ims == #{hdr['if-modified-since']}"
178 if etag = self.response['etag']
179 hdr['if-none-match'] = etag
180 debug "etag: #{etag}"
184 def revalidate(resp = self.response)
187 self.date = resp.key?('date') ? Time.httpdate(resp['date']) : Time.now
189 cc = resp['cache-control']
190 if cc && (cc =~ /max-age=(\d+)/)
191 self.expires = self.date + $1.to_i
192 elsif resp.key?('expires')
193 self.expires = Time.httpdate(resp['expires'])
194 elsif lm = resp['last-modified']
195 delta = self.date - Time.httpdate(lm)
196 delta = 10 if delta <= 0
198 self.expires = self.date + delta
200 self.expires = self.date + 300
202 # self.expires = Time.now + 10 # DEBUG
203 debug "expires on #{self.expires}"
213 self.response.raw_body
214 rescue Exception => e
216 error e.backtrace.join("\n")
226 'Accept-Charset' => 'utf-8;q=1.0, *;q=0.8',
228 "rbot http util #{$version} (http://linuxbrit.co.uk/rbot/)"
230 debug "starting http cache cleanup timer"
231 @timer = @bot.timer.add(300) {
232 self.remove_stale_cache unless @bot.config['http.no_expire_cache']
237 debug 'stopping http cache cleanup timer'
238 @bot.timer.remove(@timer)
241 # if http_proxy_include or http_proxy_exclude are set, then examine the
242 # uri to see if this is a proxied uri
243 # the in/excludes are a list of regexps, and each regexp is checked against
244 # the server name, and its IP addresses
245 def proxy_required(uri)
247 if @bot.config["http.proxy_exclude"].empty? && @bot.config["http.proxy_include"].empty?
253 list.concat Resolv.getaddresses(uri.host)
254 rescue StandardError => err
255 warning "couldn't resolve host uri.host"
258 unless @bot.config["http.proxy_exclude"].empty?
259 re = @bot.config["http.proxy_exclude"].collect{|r| Regexp.new(r)}
269 unless @bot.config["http.proxy_include"].empty?
270 re = @bot.config["http.proxy_include"].collect{|r| Regexp.new(r)}
280 debug "using proxy for uri #{uri}?: #{use_proxy}"
284 # uri:: Uri to create a proxy for
286 # return a net/http Proxy object, which is configured correctly for
287 # proxying based on the bot's proxy configuration.
288 # This will include per-url proxy configuration based on the bot config
289 # +http_proxy_include/exclude+ options.
291 def get_proxy(uri, options = {})
303 if @bot.config["http.use_proxy"]
304 if (ENV['http_proxy'])
305 proxy = URI.parse ENV['http_proxy'] rescue nil
307 if (@bot.config["http.proxy_uri"])
308 proxy = URI.parse @bot.config["http.proxy_uri"] rescue nil
311 debug "proxy is set to #{proxy.host} port #{proxy.port}"
312 if proxy_required(uri)
313 proxy_host = proxy.host
314 proxy_port = proxy.port
315 proxy_user = @bot.config["http.proxy_user"]
316 proxy_pass = @bot.config["http.proxy_pass"]
321 h = Net::HTTP.new(uri.host, uri.port, proxy_host, proxy_port, proxy_user, proxy_port)
322 h.use_ssl = true if uri.scheme == "https"
324 h.read_timeout = opts[:read_timeout]
325 h.open_timeout = opts[:open_timeout]
329 def handle_response(uri, resp, opts, &block)
330 if Net::HTTPRedirection === resp && opts[:max_redir] >= 0
331 if resp.key?('location')
332 raise 'Too many redirections' if opts[:max_redir] <= 0
333 yield resp if opts[:yield] == :all && block_given?
334 loc = resp['location']
335 new_loc = URI.join(uri.to_s, loc) rescue URI.parse(loc)
337 new_opts[:max_redir] -= 1
338 case opts[:method].to_s.downcase.intern
339 when :post, :"net::http::post"
340 new_opts[:method] = :get
342 debug "following the redirect to #{new_loc}"
343 return get_response(new_loc, new_opts, &block)
345 warning ":| redirect w/o location?"
351 # Net::HTTP wants us to read the whole body here
357 # uri:: uri to query (Uri object or String)
358 # opts:: options. Currently used:
359 # :method:: request method [:get (default), :post or :head]
360 # :open_timeout:: open timeout for the proxy
361 # :read_timeout:: read timeout for the proxy
362 # :cache:: should we cache results?
363 # :yield:: if :final [default], call &block for the response object
364 # if :all, call &block for all intermediate redirects, too
365 # :max_redir:: how many redirects to follow before raising the exception
366 # if -1, don't follow redirects, just return them
367 # :range:: make a ranged request (usually GET). accepts a string
368 # for HTTP/1.1 "Range:" header (i.e. "bytes=0-1000")
369 # :body:: request body (usually for POST requests)
371 # Generic http transaction method
373 # It will return a HTTP::Response object or raise an exception
375 # If a block is given, it will yield the response (see :yield option)
377 def get_response(uri_or_s, options = {}, &block)
378 uri = uri_or_s.kind_of?(URI) ? uri_or_s : URI.parse(uri_or_s.to_s)
380 :max_redir => @bot.config['http.max_redir'],
389 req_class = case opts[:method].to_s.downcase.intern
390 when :head, :"net::http::head"
391 opts[:max_redir] = -1
393 when :get, :"net::http::get"
395 when :post, :"net::http::post"
397 opts[:body] or raise 'post request w/o a body?'
398 warning "refusing to cache POST request" if options[:cache]
401 warning "unsupported method #{opts[:method]}, doing GET"
405 if req_class != Net::HTTP::Get && opts[:range]
406 warning "can't request ranges for #{req_class}"
410 cache_key = "#{opts[:range]}|#{req_class}|#{uri.to_s}"
412 if req_class != Net::HTTP::Get && req_class != Net::HTTP::Head
414 warning "can't cache #{req_class.inspect} requests, working w/o cache"
419 debug "get_response(#{uri}, #{opts.inspect})"
421 if opts[:cache] && cached = @cache[cache_key]
426 return handle_response(uri, cached.response, opts, &block)
430 headers = @headers.dup.merge(opts[:headers] || {})
431 headers['Range'] = opts[:range] if opts[:range]
433 cached.setup_headers(headers) if cached && (req_class == Net::HTTP::Get)
434 req = req_class.new(uri.request_uri, headers)
435 req.basic_auth(uri.user, uri.password) if uri.user && uri.password
436 req.body = opts[:body] if req_class == Net::HTTP::Post
437 debug "prepared request: #{req.to_hash.inspect}"
439 get_proxy(uri, opts).start do |http|
440 http.request(req) do |resp|
441 resp['x-rbot-location'] = uri.to_s
442 if Net::HTTPNotModified === resp
445 cached.revalidate(resp)
446 rescue Exception => e
448 error e.backtrace.join("\n")
450 debug "reusing cached"
451 resp = cached.response
452 elsif Net::HTTPServerError === resp || Net::HTTPClientError === resp
453 debug "http error, deleting cached obj" if cached
454 @cache.delete(cache_key)
457 return handle_response(uri, resp, opts, &block)
459 if cached = CachedObject.maybe_new(resp) rescue nil
460 debug "storing to cache"
461 @cache[cache_key] = cached
466 return handle_response(uri, resp, opts, &block)
471 # uri:: uri to query (Uri object)
473 # simple get request, returns (if possible) response body following redirs
474 # and caching if requested
475 def get(uri, opts = {}, &block)
477 resp = get_response(uri, opts, &block)
478 raise "http error: #{resp}" unless Net::HTTPOK === resp ||
479 Net::HTTPPartialContent === resp
481 rescue Exception => e
483 error e.backtrace.join("\n")
488 def head(uri, options = {}, &block)
489 opts = {:method => :head}.merge(options)
491 resp = get_response(uri, opts, &block)
492 raise "http error #{resp}" if Net::HTTPClientError === resp ||
493 Net::HTTPServerError == resp
495 rescue Exception => e
497 error e.backtrace.join("\n")
502 def post(uri, data, options = {}, &block)
503 opts = {:method => :post, :body => data, :cache => false}.merge(options)
505 resp = get_response(uri, opts, &block)
506 raise 'http error' unless Net::HTTPOK === resp
508 rescue Exception => e
510 error e.backtrace.join("\n")
515 def get_partial(uri, nbytes = @bot.config['http.info_bytes'], options = {}, &block)
516 opts = {:range => "bytes=0-#{nbytes}"}.merge(options)
517 return get(uri, opts, &block)
520 def remove_stale_cache
521 debug "Removing stale cache"
523 max_last = @bot.config['http.expire_time'] * 60
524 max_first = @bot.config['http.max_cache_time'] * 60
525 debug "#{@cache.size} pages before"
527 @cache.reject! { |k, val|
528 (now - val.last_used > max_last) || (now - val.first_used > max_first)
531 error "Failed to remove stale cache: #{e.inspect}"
533 debug "#{@cache.size} pages after"
540 class HttpUtilPlugin < CoreBotModule
543 debug 'initializing httputil'
544 @bot.httputil = Irc::Utils::HttpUtil.new(@bot)
548 debug 'shutting down httputil'
549 @bot.httputil.cleanup