oracular (3) HTTP::Async.3pm.gz

NAME

       HTTP::Async - process multiple HTTP requests in parallel without blocking.

SYNOPSIS

       Create an object and add some requests to it:

           use HTTP::Async;
           my $async = HTTP::Async->new;

           # create some requests and add them to the queue.
           $async->add( HTTP::Request->new( GET => 'http://www.perl.org/'         ) );
           $async->add( HTTP::Request->new( GET => 'http://www.ecclestoad.co.uk/' ) );

       and then EITHER process the responses as they come back:

           while ( my $response = $async->wait_for_next_response ) {
               # Do some processing with $response
           }

       OR do something else if there is no response ready:

           while ( $async->not_empty ) {
               if ( my $response = $async->next_response ) {
                   # deal with $response
               } else {
                   # do something else
               }
           }

       OR just use the async object to fetch stuff in the background and deal with the responses at the end.

           # Do some long code...
           for ( 1 .. 100 ) {
             some_function();
             $async->poke;            # lets it check for incoming data.
           }

           while ( my $response = $async->wait_for_next_response ) {
               # Do some processing with $response
           }

DESCRIPTION

       Although using the conventional "LWP::UserAgent" is fast and easy it does have some drawbacks - the code
       execution blocks until the request has been completed and it is only possible to process one request at a
       time.  "HTTP::Async" attempts to address these limitations.

       It gives you a 'Async' object that you can add requests to, and then get the requests off as they finish.
       The actual sending and receiving of the requests is abstracted. As soon as you add a request it is
       transmitted, if there are too many requests in progress at the moment they are queued. There is no
       concept of starting or stopping - it runs continuously.

       Whilst it is waiting to receive data it returns control to the code that called it meaning that you can
       carry out processing whilst fetching data from the network. All without forking or threading - it is
       actually done using "select" lists.

Default settings:

       There are a number of default settings that should be suitable for most uses.  However in some
       circumstances you might wish to change these.

                   slots:  20
                 timeout:  180 (seconds)
        max_request_time:  300 (seconds)
           max_redirect:   7
           poll_interval:  0.05 (seconds)
              proxy_host:  ''
              proxy_port:  ''
              local_addr:  ''
              local_port:  ''
              ssl_options: {}
              cookie_jar:  undef
              peer_addr:   ''

       If defined, is expected to be similar to "HTTP::Cookies", with extract_cookies and add_cookie_header
       methods.

       The option max_redirects has been renamed to max_redirect to be consistent with LWP::UserAgent, although
       max_redirects still works.

METHODS

   new
           my $async = HTTP::Async->new( %args );

       Creates a new HTTP::Async object and sets it up. Variations from the default can be set by passing them
       in as %args.

   slots, timeout, max_request_time, poll_interval, max_redirect, proxy_host, proxy_port, local_addr,
       local_port, ssl_options, cookie_jar, peer_addr
           $old_value = $async->slots;
           $new_value = $async->slots( $new_value );

       Get/setters for the $async objects config settings. Timeout is for inactivity and is in seconds.

       Slots is the maximum number of parallel requests to make.

   add
           my @ids      = $async->add(@requests);
           my $first_id = $async->add(@requests);

       Adds requests to the queues. Each request is given an unique integer id (for this $async) that can be
       used to track the requests if needed. If called in list context an array of ids is returned, in scalar
       context the id of the first request added is returned.

   add_with_opts
           my $id = $async->add_with_opts( $request, \%opts );

       This method lets you add a single request to the queue with options that differ from the defaults. For
       example you might wish to set a longer timeout or to use a specific proxy. Returns the id of the request.

       The method croaks when passed an invalid option.

   poke
           $async->poke;

       At fairly frequent intervals some housekeeping needs to performed - such as reading received data and
       starting new requests. Calling "poke" lets the object do this and then return quickly. Usually you will
       not need to use this as most other methods do it for you.

       You should use "poke" if your code is spending time elsewhere (ie not using the async object) to allow it
       to keep the data flowing over the network. If it is not used then the buffers may fill up and completed
       responses will not be replaced with new requests.

   next_response
           my $response          = $async->next_response;
           my ( $response, $id ) = $async->next_response;

       Returns the next response (as a HTTP::Response object) that is waiting, or returns undef if there is
       none. In list context it returns a (response, id) pair, or an empty list if none. Does not wait for a
       response so returns very quickly.

   wait_for_next_response
           my $response          = $async->wait_for_next_response( 3.5 );
           my ( $response, $id ) = $async->wait_for_next_response( 3.5 );

       As "next_response" but only returns if there is a next response or the time in seconds passed in has
       elapsed. If no time is given then it blocks. Whilst waiting it checks the queues every c<poll_interval>
       seconds. The times can be fractional seconds.

   to_send_count
           my $pending = $async->to_send_count;

       Returns the number of items which have been added but have not yet started being processed.

   to_return_count
           my $completed = $async->to_return_count;

       Returns the number of items which have completed transferring, and are waiting to be returned by
       next_response().

   in_progress_count
           my $running = $async->in_progress_count;

       Returns the number of items which are currently being processed asynchronously.

   total_count
           my $total = $async->total_count;

       Returns the sum of the to_send_count, in_progress_count and to_return_count.

       This should be the total number of items which have been added that have not yet been returned by
       next_response().

   info
           print $async->info;

       Prints a line describing what the current state is.

   remove
           $async->remove($id);
           my $success = $async->remove($id);

       Removes the item with the given id no matter which state it is currently in. Returns true if an item is
       removed, and false otherwise.

   remove_all
           $async->remove_all;
           my $success = $async->remove_all;

       Removes all items no matter what states they are currently in. Returns true if any items are removed, and
       false otherwise.

   empty, not_empty
           while ( $async->not_empty ) { ...; }
           while (1) { ...; last if $async->empty; }

       Returns true or false depending on whether there are request or responses still on the object.

   DESTROY
       The destroy method croaks if an object is destroyed but is not empty. This is to help with debugging.

SEE ALSO

       HTTP::Async::Polite - a polite form of this module. Slows the scraping down by domain so that the remote
       server is not overloaded.

GOTCHAS

       The responses may not come back in the same order as the requests were made.  For https requests to work,
       you must have Net::HTTPS::NB installed.

THANKS

       Egor Egorov contributed patches for proxies, catching connections that die before headers sent and more.

       Tomohiro Ikebe from livedoor.jp submitted patches (and a test) to properly handle 304 responses.

       Naveed Massjouni for adding the https handling code.

       Alex Balhatchet for adding the https + proxy handling code, and for making the tests run ok in parallel.

       Josef Toman for fixing two bugs, one related to header handling and another related to producing an
       absolute URL correctly.

       Github user 'c00ler-' for adding LocalAddr and LocalPort support.

       rt.cpan.org user 'Florian (fschlich)' for typo in documentation.

       Heikki Vatiainen for the ssl_options support patch.

       Daniel Lintott of the Debian Perl Group for pointing out a test failure when using a very recent version
       of HTTP::Server::Simple to implement t/TestServer.pm

BUGS AND REPO

       Please submit all bugs, patches etc on github

       <https://github.com/evdb/HTTP-Async>

AUTHOR

       Edmund von der Burg "<evdb@ecclestoad.co.uk>".

       <http://www.ecclestoad.co.uk/>

LICENCE AND COPYRIGHT

       Copyright (c) 2006, Edmund von der Burg "<evdb@ecclestoad.co.uk>".  All rights reserved.

       This module is free software; you can redistribute it and/or modify it under the same terms as Perl
       itself.

DISCLAIMER OF WARRANTY

       BECAUSE THIS SOFTWARE IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE SOFTWARE, TO THE EXTENT
       PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER
       PARTIES PROVIDE THE SOFTWARE "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED,
       INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
       PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE SOFTWARE IS WITH YOU. SHOULD THE
       SOFTWARE PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR, OR CORRECTION.

       IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY
       OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE SOFTWARE AS PERMITTED BY THE ABOVE LICENCE, BE LIABLE
       TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES ARISING OUT OF
       THE USE OR INABILITY TO USE THE SOFTWARE (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING
       RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE SOFTWARE TO OPERATE
       WITH ANY OTHER SOFTWARE), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
       DAMAGES.