Streamlines setting up and using meck-based mocks
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


This is a (somewhat) hacky attempt at making meck just a bit easier to use. My motivation to streamline and automate meck is based on two recent experiences: mocking out a complex database API to reduce testing dependencies and helping new Erlang programmers come up to speed with Eunit and meck.


All files in the repository are licensed under the Apache 2.0 license. If any file is missing the License header it should assume the following is attached;

Copyright 2014 Chef Software Inc

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
See the License for the specific language governing permissions and
limitations under the License.

automeck has two main features. The first is a data driven format for generating simple mock functions which pattern match on their arguments to select the correct output. The file priv/mocks.config is an example of how to configure automeck to generate a set of mocks.

The second feature allows automeck to record function calls. automeck records the module & function name, the call parameters, and the return value. The file priv/record.config is an example of how to configure automeck to record function calls on a set of modules. After you've collected enough data you can tell automeck to generate a set of mocks from the recorded data. The Erlang shell session listed below illustrates the API for this feature in more detail.

1> {ok, Session} = automeck:record("priv/record.config").
%% Let's generate a bunch of function calls
2> [{foo:test1(X), foo:test1(X, X + random:uniform(100)), bar:test(), bar:test(X, X)} || X <- lists:seq(1, 5)].
%% Let's look at the recorded data
3> {ok, RawData} = file:read_file(Session).
{ok, <<...>>}
4> rp(RawData).
<<"{foo, test1, [{[1,53], -52}]}.
   {bar, test, [{[], bar}]}.
   {bar, test, [{[1,1], 2}]}.
   {foo, test1, [{[2,56], -108}]}.
   {bar, test, [{[], bar}]}.
   {bar, test, [{[2,2], 4}]}.
   {foo, test1, [{[3,34], -93}]}...">>

%% Finish the recording session and generate mock config file
5> {ok, MockConfig} = automeck:finish_recording(Session).
6> MockedModules = automeck:mock(MockConfig).
[bar, foo]
%% Let's call a couple of mocked functions and an unmocked one
7> foo:test1(1, 53).
8> bar:test(2, 2).
%% The process tree crashes because this function wasn't mocked
%% Same crash would occur if a mocked function was called with
%% arguments not listed in mocks.config
9> bar:test(1).
=ERROR REPORT==== 27-Aug-2011::15:07:46 ===
** Generic server foo_meck terminating
** Last message in was {'EXIT',<0.31.0>,

automeck can also combine multiple recording sessions into a single mocks.config via automeck:combine_recordings/2. This should simplify the case where automeck is reconfigured multiple times during a test suite run, e.g. eunit's foreach and foreachx constructs. The merging process is smart enough to find conflicting function calls and abort the merge. A conflicting function call is one where the same function (same module name, function name and arity) is called in different sessions with the same inputs but returns differing results. There's no way for automeck to handle this cleanly so aborting the merge process is the only sane option.

automeck started out as an experimental hack but I think these features are generally useful for testing and understanding non-trivial Erlang systems.

automeck's lack of docs, specs, and tests will be addressed either as I have time or as I receive pull requests. Bug reports and/or suggestions also welcome :-)