Hi all,
concerning the selenium part of the tests, I suggest to treat functional tests and regression tests differently: * The framework and test runner should remain in maintenance/tests, since it provides a testing infrastructure for MW * We are working on a MediaWiki smoke test which makes sure the wiki is basically up and running. I think, that kind of tests makes perfect sense within phase3, since they might also be run by someone who sets up a wiki and wants to make sure everything in the installation went ok. Also, extension developers could use them to ensure their work didn't break any basic functionality. * Regression tests for bugfixes go to some separate place, since there might be a large number of them. I can see Ashar's point, though. Maybe a solution could be to put them in a separate tests folder and ignore this one when bundling a tarball? A similar system might also make sense for unit tests.
Cheers, Markus (mglaser)
On 6 December 2010 12:08, Petr Kadlec petr.kadlec@gmail.com wrote:
On 6 December 2010 08:47, Ashar Voultoiz hashar+wmf@free.fr wrote:
- I want to be able to commit a bug fix + associated tests in one revision without having to fetch the whole phase3 tree.
You don't have to. Subversion supports (since 1.5?) sparse checkouts http://svnbook.red-bean.com/en/1.5/svn.advanced.sparsedirs.html, allowing you to have only selected parts of a directory checked out in your working copy, which is a feature suitable exactly for scenarios like this one.
Except it is still too inconvenient to use for most cases (yeah I was going to suggest it too): "Subversion 1.5's implementation of shallow checkouts is good but does not support a couple of interesting behaviors. First, you cannot de-telescope a working copy item. Running svn update --set-depth empty in an infinite-depth working copy will not have the effect of discarding everything but the topmost directory-it will simply error out. Second, there is no depth value to indicate that you wish an item to be explicitly excluded. You have to do implicit exclusion of an item by including everything else."
-Niklas
-- Niklas Laxström
_______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l