>
> Patches in svn tickets sometimes got old too. When no one cares (or has
> time), then it doesn't matter if we use svn or github. I would say PRs on
> github are easier to find and track than what we used before,
> although seem
> harder to test, but that's what we signed up for.
Given a pull request of number ID, that's just a 2-liner:
git fetch origin pull/{ID}/head:pr-{ID}
git checkout pr-{ID}
(more verbose explanations at
https://help.github.com/en/articles/checking-out-pull-requests-locally )
But actually, you should rarely need to test a pull request locally if you
take care of checking that the pull request which adds new functionality comes
with associated tests. And if it is a bugfix, hopefully your test suite should
already test them modified code. If not, then that's the opportunity for the
pull request to add that new test.
Pull request do work well with a strong test suite.
Looks like https://github.com/OSGeo/grass/blob/master/.travis/linux.script.sh
is missing a "make check" target :-)
Even
--
Spatialys - Geospatial professional services
http://www.spatialys.com
_______________________________________________
grass-dev mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/grass-dev
