LWP::UserAgent::Mockable - Permits recording, and later playing back of LWP requests.


In your test code:

    # setup env vars to control behaviour, allowing them to be
    # overridden from command line.  In current case, do before
    # loading module, so will be actioned on.

    BEGIN {
        $ENV{LWP_UA_MOCK}      ||= 'playback';
        $ENV{LWP_UA_MOCK_FILE} ||= "$0-lwp-mock.out";

    use LWP;
    use LWP::UserAgent::Mockable;

    # setup a callback when recording, to allow modifying the response

    LWP::UserAgent::Mockable->set_record_callback( sub {
        my ( $request, $response ) = @_;

        print "GOT REQUEST TO: " . $request->uri;
        $response->content( lc( $response->content ) );

        return $response;
    } );

    # perform LWP request, as normal

    my $ua = LWP::UserAgent->new;
    my $res = $ua->get( "" );
    print $res->content;

    # when the LWP work is done, inform LWP::UserAgent::Mockable
    # that we're finished.  Will trigger any behaviour specific to
    # the action being done, such as saving the recorded session.

    END {
        # END block ensures cleanup if script dies early

To run the tests:

    # Store data
    LWP_UA_MOCK=record prove t/my-test.t

    # Use stored data
    prove t/my-test.t  # playback is default in example
    # or
    LWP_UA_MOCK=playback prove t/my-test.t

    # Re-record stored data
    LWP_UA_MOCK=record prove t/my-test.t

    # Ignore stored data
    LWP_UA_MOCK=passthrough prove t/my-test.t


This module adds session record and playback options for LWP requests, whilst trying to introduce as little clutter as necessary.

When in record mode, all LWP requests and responses will be captured in-memory, until the finished method is called, at which point they will then be written out to a file. In playback mode, LWP responses are short-circuited, to instead return the responses that were previously dumped out. If neither of the above actions are requested, this module does nothing, so LWP requests are handled as normal.

Most of the control of this module is done via environment variables, both to control the action being done (LWP_UA_MOCK env var, allowed values being 'record', 'playback', 'passthrough' (the default) ), and to control the file that is used for storing or replaying the responses (LWP_UA_MOCK_FILE env var, not used for 'passthrough' mode).

The only mandatory change to incorporate this module is to call the 'finished' method, to indicate that LWP processing is completed. Other than that, LWP handling can be done as-normal.

As the initial impetus for this module was to allow mocking of external HTTP calls within unit tests, a couple of optional callback (one for each action of the valid action types), to allow for custom handling of responses, or to modify the response that is returned back to the client (this is useful for simulating the requested system being down, or when playing back, to modify the mocked response to reflect expected dynamic content).


As there is only a singleton instance of LWP::UserAgent::Mockable, all methods are class methods.

finished() - required

Informs LWP::UserAgent::Mockable that no further requests are expected, and allow it to do any post-processing that is required.

When in 'record' mode, this will cause the playback file (controlled by LWP_UA_MOCK_FILE env var) to be created. When in 'playback' mode, this will issue a warning if there is still additional mocked responses that haven't been returned.

set_record_callback( <sub {}> ) - optional
set_playback_callback( <sub {}> ) - optional

These optional methods allow custom callbacks to be inserted, when performing the relevant actions. The callback will be invoked for each LWP request, AFTER the request has been actioned (see set_record_pre_callback for a method o short-circuiting the LWP fetch). They will be passed in 2 parameters, an HTTP::Request and an HTTP::Response object. For the record callback (which is used for both 'record' and 'passthrough' mode) the request will be the HTTP::Request object used to perform the request, and the response the HTTP::Response result from that. In playback mode, the request will be the HTTP::Request object used to perform the request, and the response the mocked response object.

When the callbacks are being used, they're expected to return an HTTP::Response object, which will be treated as the actual reply from the call being made. Failure to do do will result in a fatal error being raised.

To clear a callback, call the relevant method, passing in no argument.

set_record_pre_callback( <sub {}> ) - optional

This callback is similar to set_record_callback, except that it will short-circuit the actual fetching of the remote URL. Only a single parameter is passed through to this callback, that being the HTTP::Request object. It's expected to construct an return an HTTP::Response object (or subclass thereof). Should anything other than an HTTP::Response subclass be returned, a fatal error will be raised.

This callback will be invoked for both 'record' and 'passthrough' modes. Note that there is no analagous callback for 'playback' mode.

To clear the callback, pass in no argument.

set_playback_validation_callback( <sub {}> ) - optional

This callback allows validation of the received request. It receives two parameters, both HTTP::Requests, the first being the actual request made, the second being the mocked request that was received when recording a session. It's up to the callback to do any validation that it wants, and to perform any action that is warranted.

As with other callbacks, to clear, pass in no argument to the method.

reset( <action>, <file> ) - optional

Reset the state of mocker, allowing the action and file operation on to change. Will also reset all callbacks. Note that this will raise an error, if called whilst there are outstanding requests, and the finished method hasn't been called.


The playback file generated by this is not encrypted in any manner. As it's only using Storable to dump the file, it's easy to get at the data contained within, even if the requests are going to HTTPS sites. Treat the playback file as if it were the original data, security-wise.



Mark Morgan, <>


Michael Jemmeson, <michael.jemmeson at>

Kit Peters, <popefelix at>

Mohammad S. Anwar, <mohammad.anwar at>

Slaven Rezić, <SREZIC at>


Bugs / Feature Requests

Please report any bugs or feature requests through the issue tracker at You will be notified automatically of any progress on your issue.

Source Code

This is open source software. The code repository is available for public review and contribution under the terms of the license.

    git clone git://


Copyright 2009 Mark Morgan, all rights reserved.

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