Doc/lib/libasyncore.tex

   1 \section{\module{asyncore} ---
   2          Asynchronous socket handler}
   3
   4 \declaremodule{builtin}{asyncore}
   5 \modulesynopsis{A base class for developing asynchronous socket
   6                 handling services.}
   7 \moduleauthor{Sam Rushing}{rushing@nightmare.com}
   8 \sectionauthor{Christopher Petrilli}{petrilli@amber.org}
   9 \sectionauthor{Steve Holden}{sholden@holdenweb.com}
  10 % Heavily adapted from original documentation by Sam Rushing.
  11
  12 This module provides the basic infrastructure for writing asynchronous
  13 socket service clients and servers.
  14
  15 There are only two ways to have a program on a single processor do
  16 ``more than one thing at a time.'' Multi-threaded programming is the
  17 simplest and most popular way to do it, but there is another very
  18 different technique, that lets you have nearly all the advantages of
  19 multi-threading, without actually using multiple threads.  It's really
  20 only practical if your program is largely I/O bound.  If your program
  21 is processor bound, then pre-emptive scheduled threads are probably what
  22 you really need. Network servers are rarely processor bound, however.
  23
  24 If your operating system supports the \cfunction{select()} system call
  25 in its I/O library (and nearly all do), then you can use it to juggle
  26 multiple communication channels at once; doing other work while your
  27 I/O is taking place in the ``background.''  Although this strategy can
  28 seem strange and complex, especially at first, it is in many ways
  29 easier to understand and control than multi-threaded programming.
  30 The \module{asyncore} module solves many of the difficult problems for
  31 you, making the task of building sophisticated high-performance
  32 network servers and clients a snap. For ``conversational'' applications
  33 and protocols the companion  \refmodule{asynchat} module is invaluable.
  34
  35 The basic idea behind both modules is to create one or more network
  36 \emph{channels}, instances of class \class{asyncore.dispatcher} and
  37 \class{asynchat.async_chat}. Creating the channels adds them to a global
  38 map, used by the \function{loop()} function if you do not provide it
  39 with your own \var{map}.
  40
  41 Once the initial channel(s) is(are) created, calling the \function{loop()}
  42 function activates channel service, which continues until the last
  43 channel (including any that have been added to the map during asynchronous
  44 service) is closed.
  45
  46 \begin{funcdesc}{loop}{\optional{timeout\optional{, use_poll\optional{,
  47                        map}}}}
  48   Enter a polling loop that only terminates after all open channels
  49   have been closed.  All arguments are optional.  The \var{timeout}
  50   argument sets the timeout parameter for the appropriate
  51   \function{select()} or \function{poll()} call, measured in seconds;
  52   the default is 30 seconds.  The \var{use_poll} parameter, if true,
  53   indicates that \function{poll()} should be used in preference to
  54   \function{select()} (the default is \code{False}).  The \var{map} parameter
  55   is a dictionary whose items are the channels to watch.  As channels
  56   are closed they are deleted from their map.  If \var{map} is
  57   omitted, a global map is used (this map is updated by the default
  58   class \method{__init__()}
  59   -- make sure you extend, rather than override, \method{__init__()}
  60   if you want to retain this behavior).
  61
  62   Channels (instances of \class{asyncore.dispatcher}, \class{asynchat.async_chat}
  63   and subclasses thereof) can freely be mixed in the map.
  64 \end{funcdesc}
  65
  66 \begin{classdesc}{dispatcher}{}
  67   The \class{dispatcher} class is a thin wrapper around a low-level socket object.
  68   To make it more useful, it has a few methods for event-handling  which are called
  69   from the asynchronous loop.
  70   Otherwise, it can be treated as a normal non-blocking socket object.
  71
  72   Two class attributes can be modified, to improve performance,
  73   or possibly even to conserve memory.
  74
  75   \begin{datadesc}{ac_in_buffer_size}
  76   The asynchronous input buffer size (default \code{4096}).
  77   \end{datadesc}
  78
  79   \begin{datadesc}{ac_out_buffer_size}
  80   The asynchronous output buffer size (default \code{4096}).
  81   \end{datadesc}
  82
  83   The firing of low-level events at certain times or in certain connection
  84   states tells the asynchronous loop that certain higher-level events have
  85   taken place. For example, if we have asked for a socket to connect to
  86   another host, we know that the connection has been made when the socket
  87   becomes writable for the first time (at this point you know that you may
  88   write to it with the expectation of success). The implied higher-level
  89   events are:
  90
  91   \begin{tableii}{l|l}{code}{Event}{Description}
  92     \lineii{handle_connect()}{Implied by the first write event}
  93     \lineii{handle_close()}{Implied by a read event with no data available}
  94     \lineii{handle_accept()}{Implied by a read event on a listening socket}
  95   \end{tableii}
  96
  97   During asynchronous processing, each mapped channel's \method{readable()}
  98   and \method{writable()} methods are used to determine whether the channel's
  99   socket should be added to the list of channels \cfunction{select()}ed or
 100   \cfunction{poll()}ed for read and write events.
 101
 102 \end{classdesc}
 103
 104 Thus, the set of channel events is larger than the basic socket events.
 105 The full set of methods that can be overridden in your subclass follows:
 106
 107 \begin{methoddesc}{handle_read}{}
 108   Called when the asynchronous loop detects that a \method{read()}
 109   call on the channel's socket will succeed.
 110 \end{methoddesc}
 111
 112 \begin{methoddesc}{handle_write}{}
 113   Called when the asynchronous loop detects that a writable socket
 114   can be written.
 115   Often this method will implement the necessary buffering for
 116   performance.  For example:
 117
 118 \begin{verbatim}
 119 def handle_write(self):
 120     sent = self.send(self.buffer)
 121     self.buffer = self.buffer[sent:]
 122 \end{verbatim}
 123 \end{methoddesc}
 124
 125 \begin{methoddesc}{handle_expt}{}
 126   Called when there is out of band (OOB) data for a socket
 127   connection.  This will almost never happen, as OOB is
 128   tenuously supported and rarely used.
 129 \end{methoddesc}
 130
 131 \begin{methoddesc}{handle_connect}{}
 132   Called when the active opener's socket actually makes a connection.
 133   Might send a ``welcome'' banner, or initiate a protocol
 134   negotiation with the remote endpoint, for example.
 135 \end{methoddesc}
 136
 137 \begin{methoddesc}{handle_close}{}
 138   Called when the socket is closed.
 139 \end{methoddesc}
 140
 141 \begin{methoddesc}{handle_error}{}
 142   Called when an exception is raised and not otherwise handled.  The default
 143   version prints a condensed traceback.
 144 \end{methoddesc}
 145
 146 \begin{methoddesc}{handle_accept}{}
 147   Called on listening channels (passive openers) when a
 148   connection can be established with a new remote endpoint that
 149   has issued a \method{connect()} call for the local endpoint.
 150 \end{methoddesc}
 151
 152 \begin{methoddesc}{readable}{}
 153   Called each time around the asynchronous loop to determine whether a
 154   channel's socket should be added to the list on which read events can
 155   occur.  The default method simply returns \code{True},
 156   indicating that by default, all channels will be interested in
 157   read events.
 158 \end{methoddesc}
 159
 160 \begin{methoddesc}{writable}{}
 161   Called each time around the asynchronous loop to determine whether a
 162   channel's socket should be added to the list on which write events can
 163   occur.  The default method simply returns \code{True},
 164   indicating that by default, all channels will be interested in
 165   write events.
 166 \end{methoddesc}
 167
 168 In addition, each channel delegates or extends many of the socket methods.
 169 Most of these are nearly identical to their socket partners.
 170
 171 \begin{methoddesc}{create_socket}{family, type}
 172   This is identical to the creation of a normal socket, and
 173   will use the same options for creation.  Refer to the
 174   \refmodule{socket} documentation for information on creating
 175   sockets.
 176 \end{methoddesc}
 177
 178 \begin{methoddesc}{connect}{address}
 179   As with the normal socket object, \var{address} is a
 180   tuple with the first element the host to connect to, and the
 181   second the port number.
 182 \end{methoddesc}
 183
 184 \begin{methoddesc}{send}{data}
 185   Send \var{data} to the remote end-point of the socket.
 186 \end{methoddesc}
 187
 188 \begin{methoddesc}{recv}{buffer_size}
 189   Read at most \var{buffer_size} bytes from the socket's remote end-point.
 190   An empty string implies that the channel has been closed from the other
 191   end.
 192 \end{methoddesc}
 193
 194 \begin{methoddesc}{listen}{backlog}
 195   Listen for connections made to the socket.  The \var{backlog}
 196   argument specifies the maximum number of queued connections
 197   and should be at least 1; the maximum value is
 198   system-dependent (usually 5).
 199 \end{methoddesc}
 200
 201 \begin{methoddesc}{bind}{address}
 202   Bind the socket to \var{address}.  The socket must not already
 203   be bound.  (The format of \var{address} depends on the address
 204   family --- see above.)
 205 \end{methoddesc}
 206
 207 \begin{methoddesc}{accept}{}
 208   Accept a connection.  The socket must be bound to an address
 209   and listening for connections.  The return value is a pair
 210   \code{(\var{conn}, \var{address})} where \var{conn} is a
 211   \emph{new} socket object usable to send and receive data on
 212   the connection, and \var{address} is the address bound to the
 213   socket on the other end of the connection.
 214 \end{methoddesc}
 215
 216 \begin{methoddesc}{close}{}
 217   Close the socket.  All future operations on the socket object
 218   will fail.  The remote end-point will receive no more data (after
 219   queued data is flushed).  Sockets are automatically closed
 220   when they are garbage-collected.
 221 \end{methoddesc}
 222
 223
 224 \subsection{asyncore Example basic HTTP client \label{asyncore-example}}
 225
 226 As a basic example, below is a very basic HTTP client that uses the
 227 \class{dispatcher} class to implement its socket handling:
 228
 229 \begin{verbatim}
 230 class http_client(asyncore.dispatcher):
 231     def __init__(self, host,path):
 232         asyncore.dispatcher.__init__(self)
 233         self.path = path
 234         self.create_socket(socket.AF_INET, socket.SOCK_STREAM)
 235         self.connect( (host, 80) )
 236         self.buffer = 'GET %s HTTP/1.0\r\n\r\n' % self.path
 237
 238     def handle_connect(self):
 239         pass
 240
 241     def handle_read(self):
 242         data = self.recv(8192)
 243         print data
 244
 245     def writable(self):
 246         return (len(self.buffer) > 0)
 247
 248     def handle_write(self):
 249         sent = self.send(self.buffer)
 250         self.buffer = self.buffer[sent:]
 251 \end{verbatim}