Re: Cask for testing

[Sebastian Wiesner]

Am 14.08.2014 um 12:31 schrieb Phillip Lord <phillip.lord@newcastle.ac.uk>:

> 
> 
> I have been investigating Cask for use with testing my Emacs packages,
> but have been struggling with it. The documentation is not very rich as
> it stands. I was wondering if anyone has a good workflow?
> 
> I thought I would start of with m-buffer.el which has few dependencies.
> 
> https://github.com/phillord/m-buffer-el
> 
> Currently, I have this testing in batch with a shell script. The
> dependencies have been included and downloaded into the test directory
> (which is obviously not ideal). I am using travis for CI. Testing also
> works interactively from within Emacs. My hope was to use Cask for batch
> dependency resolution.
> 
> My Cask file looks like this:
> 
> (source gnu)
> (source melpa)
> 
> (package-file "m-buffer.el")
> (depends-on "dash")
> 
> (development
> (depends-on "ert-runner"))
> 
> Cask correctly downloads the dependencies. But it doesn't put
> m-buffer.el onto the load path. So any attempt to use ert-runner fails.
> What am I doing wrong?


Nothing, Cask simply doesn’t do that.  It doesn’t make assumption about
the local layout of the source code tree, since that’d be too flaky for
more complex packages, e.g. consisting of different files in different 
directories.

However, you can tell erg-runner to additional files, e.g.

cask exec ert-runner -l ./m-buffer.el 

To make this permanent, add an „.ert-runner“ file to the root of your
source code tree, with the following contents:

-l ./m-buffer.el 

Please note that Cask is not part of GNU Emacs, nor in any way affiliated 
with it.  As such, this list is not the best place to ask questions about
Cask, since it’s not regularly monitored by us.  Please refer to the Github
issue tracker or to Cask’s own mailing list for questions.