Dear Urs,
Il 09/02/2015 10:16, Urs Liska ha scritto:
Am 09.02.2015 um 09:56 schrieb Matteo Ceccarello:
Here I see room for starting some discussion on automated testing. I
think that example files can serve also as test files, to check that
code changes or version changes do not break things. Github has an
associated continuous integration service called [Travis
CI](https://travis-ci.org/) that runs test suites (or arbitrary
commands) on each push. There is no explicit support for LilyPond,
but I think that we can make it work anyway.
This is a good idea. But it would also imply that usage example files
should be required to also cover failure conditions.
I opened a pull request
(https://github.com/openlilylib/openlilylib/pull/60) which implements a
very simple testing strategy. Library developers are able to configure
arbitrary lilypond files as test cases, so examples and other more
specific files can be used as test cases. So example files do not
necessarily need to be included in the tests.
This still does not handle failure condition, for which I think we will
need a full fledged testing framework. Anyway, the simple testing
strategy I'm proposing can already catch bugs like the one you pointed
out in this thread (btw, I fixed it, in the next days I'll merge the fix
into master).
For openLilyLib I can also offer the openlilylib.org server for
implementing continuous integration, but I would need help to set
everything up (actually this is something I'd be quite interested in
learning), for example with a Jenkins server.
I have no experience in setting up Jenkins, I think that the tradeoffs
between running on Travis CI vs. a self-hosted Jenkins server need to be
evaluated, so further investigation is needed.
Bests,
Matteo
_______________________________________________
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user