# Copyright (c) 2005 Zed A. Shaw # You can redistribute it and/or modify it under the same terms as Ruby. # # Additional work donated by contributors. See http://mongrel.rubyforge.org/attributions.html # for more information. $mongrel_debug_client = false require 'rubygems' require 'socket' require 'http11' require 'tempfile' begin require 'fastthread' rescue RuntimeError => e warn "fastthread not loaded: #{ e.message }" rescue LoadError ensure require 'thread' end require 'stringio' require 'mongrel/cgi' require 'mongrel/handlers' require 'mongrel/command' require 'mongrel/tcphack' require 'yaml' require 'mongrel/configurator' require 'time' require 'etc' require 'uri' # Mongrel module containing all of the classes (include C extensions) for running # a Mongrel web server. It contains a minimalist HTTP server with just enough # functionality to service web application requests fast as possible. module Mongrel class URIClassifier attr_reader :handler_map # Returns the URIs that have been registered with this classifier so far. # The URIs returned should not be modified as this will cause a memory leak. # You can use this to inspect the contents of the URIClassifier. def uris @handler_map.keys end # Simply does an inspect that looks like a Hash inspect. def inspect @handler_map.inspect end end # Used to stop the HttpServer via Thread.raise. class StopServer < Exception; end # Thrown at a thread when it is timed out. class TimeoutError < Exception; end # Every standard HTTP code mapped to the appropriate message. These are # used so frequently that they are placed directly in Mongrel for easy # access rather than Mongrel::Const. HTTP_STATUS_CODES = { 100 => 'Continue', 101 => 'Switching Protocols', 200 => 'OK', 201 => 'Created', 202 => 'Accepted', 203 => 'Non-Authoritative Information', 204 => 'No Content', 205 => 'Reset Content', 206 => 'Partial Content', 300 => 'Multiple Choices', 301 => 'Moved Permanently', 302 => 'Moved Temporarily', 303 => 'See Other', 304 => 'Not Modified', 305 => 'Use Proxy', 400 => 'Bad Request', 401 => 'Unauthorized', 402 => 'Payment Required', 403 => 'Forbidden', 404 => 'Not Found', 405 => 'Method Not Allowed', 406 => 'Not Acceptable', 407 => 'Proxy Authentication Required', 408 => 'Request Time-out', 409 => 'Conflict', 410 => 'Gone', 411 => 'Length Required', 412 => 'Precondition Failed', 413 => 'Request Entity Too Large', 414 => 'Request-URI Too Large', 415 => 'Unsupported Media Type', 500 => 'Internal Server Error', 501 => 'Not Implemented', 502 => 'Bad Gateway', 503 => 'Service Unavailable', 504 => 'Gateway Time-out', 505 => 'HTTP Version not supported' } # Frequently used constants when constructing requests or responses. Many times # the constant just refers to a string with the same contents. Using these constants # gave about a 3% to 10% performance improvement over using the strings directly. # Symbols did not really improve things much compared to constants. # # While Mongrel does try to emulate the CGI/1.2 protocol, it does not use the REMOTE_IDENT, # REMOTE_USER, or REMOTE_HOST parameters since those are either a security problem or # too taxing on performance. module Const DATE = "Date".freeze # This is the part of the path after the SCRIPT_NAME. URIClassifier will determine this. PATH_INFO="PATH_INFO".freeze # This is the initial part that your handler is identified as by URIClassifier. SCRIPT_NAME="SCRIPT_NAME".freeze # The original URI requested by the client. Passed to URIClassifier to build PATH_INFO and SCRIPT_NAME. REQUEST_URI='REQUEST_URI'.freeze REQUEST_PATH='REQUEST_PATH'.freeze MONGREL_VERSION="1.0.1".freeze MONGREL_TMP_BASE="mongrel".freeze # The standard empty 404 response for bad requests. Use Error4040Handler for custom stuff. ERROR_404_RESPONSE="HTTP/1.1 404 Not Found\r\nConnection: close\r\nServer: Mongrel #{MONGREL_VERSION}\r\n\r\nNOT FOUND".freeze CONTENT_LENGTH="CONTENT_LENGTH".freeze # A common header for indicating the server is too busy. Not used yet. ERROR_503_RESPONSE="HTTP/1.1 503 Service Unavailable\r\n\r\nBUSY".freeze # The basic max request size we'll try to read. CHUNK_SIZE=(16 * 1024) # This is the maximum header that is allowed before a client is booted. The parser detects # this, but we'd also like to do this as well. MAX_HEADER=1024 * (80 + 32) # Maximum request body size before it is moved out of memory and into a tempfile for reading. MAX_BODY=MAX_HEADER # A frozen format for this is about 15% faster STATUS_FORMAT = "HTTP/1.1 %d %s\r\nConnection: close\r\n".freeze CONTENT_TYPE = "Content-Type".freeze LAST_MODIFIED = "Last-Modified".freeze ETAG = "ETag".freeze SLASH = "/".freeze REQUEST_METHOD="REQUEST_METHOD".freeze GET="GET".freeze HEAD="HEAD".freeze # ETag is based on the apache standard of hex mtime-size-inode (inode is 0 on win32) ETAG_FORMAT="\"%x-%x-%x\"".freeze HEADER_FORMAT="%s: %s\r\n".freeze LINE_END="\r\n".freeze REMOTE_ADDR="REMOTE_ADDR".freeze HTTP_X_FORWARDED_FOR="HTTP_X_FORWARDED_FOR".freeze HTTP_IF_MODIFIED_SINCE="HTTP_IF_MODIFIED_SINCE".freeze HTTP_IF_NONE_MATCH="HTTP_IF_NONE_MATCH".freeze REDIRECT = "HTTP/1.1 302 Found\r\nLocation: %s\r\nConnection: close\r\n\r\n".freeze HOST = "HOST".freeze end # Basically a Hash with one extra parameter for the HTTP body, mostly used internally. class HttpParams < Hash attr_accessor :http_body end # When a handler is found for a registered URI then this class is constructed # and passed to your HttpHandler::process method. You should assume that # *one* handler processes all requests. Included in the HttpRequest is a # HttpRequest.params Hash that matches common CGI params, and a HttpRequest.body # which is a string containing the request body (raw for now). # # The HttpRequest.initialize method will convert any request that is larger than # Const::MAX_BODY into a Tempfile and use that as the body. Otherwise it uses # a StringIO object. To be safe, you should assume it works like a file. # # The HttpHandler.request_notify system is implemented by having HttpRequest call # HttpHandler.request_begins, HttpHandler.request_progress, HttpHandler.process during # the IO processing. This adds a small amount of overhead but lets you implement # finer controlled handlers and filters. class HttpRequest attr_reader :body, :params # You don't really call this. It's made for you. # Main thing it does is hook up the params, and store any remaining # body data into the HttpRequest.body attribute. def initialize(params, socket, dispatchers) @params = params @socket = socket @dispatchers = dispatchers content_length = @params[Const::CONTENT_LENGTH].to_i remain = content_length - @params.http_body.length # tell all dispatchers the request has begun @dispatchers.each do |dispatcher| dispatcher.request_begins(@params) end unless @dispatchers.nil? || @dispatchers.empty? # Some clients (like FF1.0) report 0 for body and then send a body. This will probably truncate them but at least the request goes through usually. if remain <= 0 # we've got everything, pack it up @body = StringIO.new @body.write @params.http_body update_request_progress(0, content_length) elsif remain > 0 # must read more data to complete body if remain > Const::MAX_BODY # huge body, put it in a tempfile @body = Tempfile.new(Const::MONGREL_TMP_BASE) @body.binmode else # small body, just use that @body = StringIO.new end @body.write @params.http_body read_body(remain, content_length) end @body.rewind if @body end # updates all dispatchers about our progress def update_request_progress(clen, total) return if @dispatchers.nil? || @dispatchers.empty? @dispatchers.each do |dispatcher| dispatcher.request_progress(@params, clen, total) end end private :update_request_progress # Does the heavy lifting of properly reading the larger body requests in # small chunks. It expects @body to be an IO object, @socket to be valid, # and will set @body = nil if the request fails. It also expects any initial # part of the body that has been read to be in the @body already. def read_body(remain, total) begin # write the odd sized chunk first @params.http_body = read_socket(remain % Const::CHUNK_SIZE) remain -= @body.write(@params.http_body) update_request_progress(remain, total) # then stream out nothing but perfectly sized chunks until remain <= 0 or @socket.closed? # ASSUME: we are writing to a disk and these writes always write the requested amount @params.http_body = read_socket(Const::CHUNK_SIZE) remain -= @body.write(@params.http_body) update_request_progress(remain, total) end rescue Object STDERR.puts "ERROR reading http body: #$!" $!.backtrace.join("\n") # any errors means we should delete the file, including if the file is dumped @socket.close rescue Object @body.delete if @body.class == Tempfile @body = nil # Q_SIGNALS that there was a problem end end def read_socket(len) if !@socket.closed? data = @socket.read(len) if !data raise "Socket read return nil" elsif data.length != len raise "Socket read returned insufficient data: #{data.length}" else data end else raise "Socket already closed when reading." end end # Performs URI escaping so that you can construct proper # query strings faster. Use this rather than the cgi.rb # version since it's faster. (Stolen from Camping). def self.escape(s) s.to_s.gsub(/([^ a-zA-Z0-9_.-]+)/n) { '%'+$1.unpack('H2'*$1.size).join('%').upcase }.tr(' ', '+') end # Unescapes a URI escaped string. (Stolen from Camping). def self.unescape(s) s.tr('+', ' ').gsub(/((?:%[0-9a-fA-F]{2})+)/n){ [$1.delete('%')].pack('H*') } end # Parses a query string by breaking it up at the '&' # and ';' characters. You can also use this to parse # cookies by changing the characters used in the second # parameter (which defaults to '&;'. def self.query_parse(qs, d = '&;') params = {} (qs||'').split(/[#{d}] */n).inject(params) { |h,p| k, v=unescape(p).split('=',2) if cur = params[k] if cur.class == Array params[k] << v else params[k] = [cur, v] end else params[k] = v end } return params end end # This class implements a simple way of constructing the HTTP headers dynamically # via a Hash syntax. Think of it as a write-only Hash. Refer to HttpResponse for # information on how this is used. # # One consequence of this write-only nature is that you can write multiple headers # by just doing them twice (which is sometimes needed in HTTP), but that the normal # semantics for Hash (where doing an insert replaces) is not there. class HeaderOut attr_reader :out attr_accessor :allowed_duplicates def initialize(out) @sent = {} @allowed_duplicates = {"Set-Cookie" => true, "Set-Cookie2" => true, "Warning" => true, "WWW-Authenticate" => true} @out = out end # Simply writes "#{key}: #{value}" to an output buffer. def[]=(key,value) if not @sent.has_key?(key) or @allowed_duplicates.has_key?(key) @sent[key] = true @out.write(Const::HEADER_FORMAT % [key, value]) end end end # Writes and controls your response to the client using the HTTP/1.1 specification. # You use it by simply doing: # # response.start(200) do |head,out| # head['Content-Type'] = 'text/plain' # out.write("hello\n") # end # # The parameter to start is the response code--which Mongrel will translate for you # based on HTTP_STATUS_CODES. The head parameter is how you write custom headers. # The out parameter is where you write your body. The default status code for # HttpResponse.start is 200 so the above example is redundant. # # As you can see, it's just like using a Hash and as you do this it writes the proper # header to the output on the fly. You can even intermix specifying headers and # writing content. The HttpResponse class with write the things in the proper order # once the HttpResponse.block is ended. # # You may also work the HttpResponse object directly using the various attributes available # for the raw socket, body, header, and status codes. If you do this you're on your own. # A design decision was made to force the client to not pipeline requests. HTTP/1.1 # pipelining really kills the performance due to how it has to be handled and how # unclear the standard is. To fix this the HttpResponse gives a "Connection: close" # header which forces the client to close right away. The bonus for this is that it # gives a pretty nice speed boost to most clients since they can close their connection # immediately. # # One additional caveat is that you don't have to specify the Content-length header # as the HttpResponse will write this for you based on the out length. class HttpResponse attr_reader :socket attr_reader :body attr_writer :body attr_reader :header attr_reader :status attr_writer :status attr_reader :body_sent attr_reader :header_sent attr_reader :status_sent def initialize(socket) @socket = socket @body = StringIO.new @status = 404 @header = HeaderOut.new(StringIO.new) @header[Const::DATE] = Time.now.httpdate @body_sent = false @header_sent = false @status_sent = false end # Receives a block passing it the header and body for you to work with. # When the block is finished it writes everything you've done to # the socket in the proper order. This lets you intermix header and # body content as needed. Handlers are able to modify pretty much # any part of the request in the chain, and can stop further processing # by simple passing "finalize=true" to the start method. By default # all handlers run and then mongrel finalizes the request when they're # all done. def start(status=200, finalize=false) @status = status.to_i yield @header, @body finished if finalize end # Primarily used in exception handling to reset the response output in order to write # an alternative response. It will abort with an exception if you have already # sent the header or the body. This is pretty catastrophic actually. def reset if @body_sent raise "You have already sent the request body." elsif @header_sent raise "You have already sent the request headers." else @header.out.truncate(0) @body.close @body = StringIO.new end end def send_status(content_length=@body.length) if not @status_sent @header['Content-Length'] = content_length unless @status == 304 write(Const::STATUS_FORMAT % [@status, HTTP_STATUS_CODES[@status]]) @status_sent = true end end def send_header if not @header_sent @header.out.rewind write(@header.out.read + Const::LINE_END) @header_sent = true end end def send_body if not @body_sent @body.rewind write(@body.read) @body_sent = true end end # Appends the contents of +path+ to the response stream. The file is opened for binary # reading and written in chunks to the socket. # # Sendfile API support has been removed in 0.3.13.4 due to stability problems. def send_file(path, small_file = false) if small_file File.open(path, "rb") {|f| @socket << f.read } else File.open(path, "rb") do |f| while chunk = f.read(Const::CHUNK_SIZE) and chunk.length > 0 begin write(chunk) rescue Object => exc break end end end end @body_sent = true end def socket_error(details) # ignore these since it means the client closed off early @socket.close rescue Object done = true raise details end def write(data) @socket.write(data) rescue => details socket_error(details) end # This takes whatever has been done to header and body and then writes it in the # proper format to make an HTTP/1.1 response. def finished send_status send_header send_body end # Used during error conditions to mark the response as "done" so there isn't any more processing # sent to the client. def done=(val) @status_sent = true @header_sent = true @body_sent = true end def done (@status_sent and @header_sent and @body_sent) end end # This is the main driver of Mongrel, while the Mongrel::HttpParser and Mongrel::URIClassifier # make up the majority of how the server functions. It's a very simple class that just # has a thread accepting connections and a simple HttpServer.process_client function # to do the heavy lifting with the IO and Ruby. # # You use it by doing the following: # # server = HttpServer.new("0.0.0.0", 3000) # server.register("/stuff", MyNiftyHandler.new) # server.run.join # # The last line can be just server.run if you don't want to join the thread used. # If you don't though Ruby will mysteriously just exit on you. # # Ruby's thread implementation is "interesting" to say the least. Experiments with # *many* different types of IO processing simply cannot make a dent in it. Future # releases of Mongrel will find other creative ways to make threads faster, but don't # hold your breath until Ruby 1.9 is actually finally useful. class HttpServer attr_reader :acceptor attr_reader :workers attr_reader :classifier attr_reader :host attr_reader :port attr_reader :timeout attr_reader :num_processors # Creates a working server on host:port (strange things happen if port isn't a Number). # Use HttpServer::run to start the server and HttpServer.acceptor.join to # join the thread that's processing incoming requests on the socket. # # The num_processors optional argument is the maximum number of concurrent # processors to accept, anything over this is closed immediately to maintain # server processing performance. This may seem mean but it is the most efficient # way to deal with overload. Other schemes involve still parsing the client's request # which defeats the point of an overload handling system. # # The timeout parameter is a sleep timeout (in hundredths of a second) that is placed between # socket.accept calls in order to give the server a cheap throttle time. It defaults to 0 and # actually if it is 0 then the sleep is not done at all. def initialize(host, port, num_processors=(2**30-1), timeout=0) @socket = TCPServer.new(host, port) @classifier = URIClassifier.new @host = host @port = port @workers = ThreadGroup.new @timeout = timeout @num_processors = num_processors @death_time = 60 end # Does the majority of the IO processing. It has been written in Ruby using # about 7 different IO processing strategies and no matter how it's done # the performance just does not improve. It is currently carefully constructed # to make sure that it gets the best possible performance, but anyone who # thinks they can make it faster is more than welcome to take a crack at it. def process_client(client) begin parser = HttpParser.new params = HttpParams.new request = nil data = client.readpartial(Const::CHUNK_SIZE) nparsed = 0 # Assumption: nparsed will always be less since data will get filled with more # after each parsing. If it doesn't get more then there was a problem # with the read operation on the client socket. Effect is to stop processing when the # socket can't fill the buffer for further parsing. while nparsed < data.length nparsed = parser.execute(params, data, nparsed) if parser.finished? if not params[Const::REQUEST_PATH] # it might be a dumbass full host request header uri = URI.parse(params[Const::REQUEST_URI]) params[Const::REQUEST_PATH] = uri.request_uri end raise "No REQUEST PATH" if not params[Const::REQUEST_PATH] script_name, path_info, handlers = @classifier.resolve(params[Const::REQUEST_PATH]) if handlers params[Const::PATH_INFO] = path_info params[Const::SCRIPT_NAME] = script_name params[Const::REMOTE_ADDR] = params[Const::HTTP_X_FORWARDED_FOR] || client.peeraddr.last # select handlers that want more detailed request notification notifiers = handlers.select { |h| h.request_notify } request = HttpRequest.new(params, client, notifiers) # in the case of large file uploads the user could close the socket, so skip those requests break if request.body == nil # nil Q_SIGNALS from HttpRequest::initialize that the request was aborted # request is good so far, continue processing the response response = HttpResponse.new(client) # Process each handler in registered order until we run out or one finalizes the response. handlers.each do |handler| handler.process(request, response) break if response.done or client.closed? end # And finally, if nobody closed the response off, we finalize it. unless response.done or client.closed? response.finished end else # Didn't find it, return a stock 404 response. client.write(Const::ERROR_404_RESPONSE) end break #done else # Parser is not done, queue up more data to read and continue parsing chunk = client.readpartial(Const::CHUNK_SIZE) break if !chunk or chunk.length == 0 # read failed, stop processing data << chunk if data.length >= Const::MAX_HEADER raise HttpParserError.new("HEADER is longer than allowed, aborting client early.") end end end rescue EOFError,Errno::ECONNRESET,Errno::EPIPE,Errno::EINVAL,Errno::EBADF client.close rescue Object rescue HttpParserError if $mongrel_debug_client STDERR.puts "#{Time.now}: BAD CLIENT (#{params[Const::HTTP_X_FORWARDED_FOR] || client.peeraddr.last}): #$!" STDERR.puts "#{Time.now}: REQUEST DATA: #{data.inspect}\n---\nPARAMS: #{params.inspect}\n---\n" end rescue Errno::EMFILE reap_dead_workers('too many files') rescue Object STDERR.puts "#{Time.now}: ERROR: #$!" STDERR.puts $!.backtrace.join("\n") if $mongrel_debug_client ensure client.close rescue Object request.body.delete if request and request.body.class == Tempfile end end # Used internally to kill off any worker threads that have taken too long # to complete processing. Only called if there are too many processors # currently servicing. It returns the count of workers still active # after the reap is done. It only runs if there are workers to reap. def reap_dead_workers(reason='unknown') if @workers.list.length > 0 STDERR.puts "#{Time.now}: Reaping #{@workers.list.length} threads for slow workers because of '#{reason}'" error_msg = "Mongrel timed out this thread: #{reason}" mark = Time.now @workers.list.each do |w| w[:started_on] = Time.now if not w[:started_on] if mark - w[:started_on] > @death_time + @timeout STDERR.puts "Thread #{w.inspect} is too old, killing." w.raise(TimeoutError.new(error_msg)) end end end return @workers.list.length end # Performs a wait on all the currently running threads and kills any that take # too long. Right now it just waits 60 seconds, but will expand this to # allow setting. The @timeout setting does extend this waiting period by # that much longer. def graceful_shutdown while reap_dead_workers("shutdown") > 0 STDERR.print "Waiting for #{@workers.list.length} requests to finish, could take #{@death_time + @timeout} seconds." sleep @death_time / 10 end end def configure_socket_options case RUBY_PLATFORM when /linux/ # 9 is currently TCP_DEFER_ACCEPT $tcp_defer_accept_opts = [Socket::SOL_TCP, 9, 1] $tcp_cork_opts = [Socket::SOL_TCP, 3, 1] when /freebsd/ # Use the HTTP accept filter if available. # The struct made by pack() is defined in /usr/include/sys/socket.h as accept_filter_arg unless `/sbin/sysctl -nq net.inet.accf.http`.empty? $tcp_defer_accept_opts = [Socket::SOL_SOCKET, Socket::SO_ACCEPTFILTER, ['httpready', nil].pack('a16a240')] end end end # Runs the thing. It returns the thread used so you can "join" it. You can also # access the HttpServer::acceptor attribute to get the thread later. def run BasicSocket.do_not_reverse_lookup=true configure_socket_options if $tcp_defer_accept_opts @socket.setsockopt(*$tcp_defer_accept_opts) rescue nil end @acceptor = Thread.new do while true begin client = @socket.accept if $tcp_cork_opts client.setsockopt(*$tcp_cork_opts) rescue nil end worker_list = @workers.list if worker_list.length >= @num_processors STDERR.puts "Server overloaded with #{worker_list.length} processors (#@num_processors max). Dropping connection." client.close rescue Object reap_dead_workers("max processors") else thread = Thread.new(client) {|c| process_client(c) } thread[:started_on] = Time.now @workers.add(thread) sleep @timeout/100 if @timeout > 0 end rescue StopServer @socket.close rescue Object break rescue Errno::EMFILE reap_dead_workers("too many open files") sleep 0.5 rescue Errno::ECONNABORTED # client closed the socket even before accept client.close rescue Object rescue Object => exc STDERR.puts "!!!!!! UNHANDLED EXCEPTION! #{exc}. TELL ZED HE'S A MORON." STDERR.puts $!.backtrace.join("\n") if $mongrel_debug_client end end graceful_shutdown end return @acceptor end # Simply registers a handler with the internal URIClassifier. When the URI is # found in the prefix of a request then your handler's HttpHandler::process method # is called. See Mongrel::URIClassifier#register for more information. # # If you set in_front=true then the passed in handler will be put in front in the list. # Otherwise it's placed at the end of the list. def register(uri, handler, in_front=false) script_name, path_info, handlers = @classifier.resolve(uri) if not handlers @classifier.register(uri, [handler]) else if path_info.length == 0 or (script_name == Const::SLASH and path_info == Const::SLASH) if in_front handlers.unshift(handler) else handlers << handler end else @classifier.register(uri, [handler]) end end handler.listener = self end # Removes any handlers registered at the given URI. See Mongrel::URIClassifier#unregister # for more information. Remember this removes them *all* so the entire # processing chain goes away. def unregister(uri) @classifier.unregister(uri) end # Stops the acceptor thread and then causes the worker threads to finish # off the request queue before finally exiting. def stop stopper = Thread.new do exc = StopServer.new @acceptor.raise(exc) end stopper.priority = 10 end end end