Hello,
Back in June, I have added REL1_18 to the continuous integration test suite. To make it works, I had to disable the 'Database' test group which was severely broken at the time.
Demon fixed the Database group slowness with r88755. I have backported it in REL1_18 together with other test fixes (see commit message).
The backport revision is r92239 [2]. I have run the tests on CC for this revision WITHOUT the Database group. Then I enabled the Database group and triggered a build manually. End result:
The GOOD: CC is now really running tests for 1.18 (including parser)
The BAD: 1.18 is greatly broken (robla will make us fix it)
The Weird: my commit (r92239) is not the root cause :-))
From a quick look, broken tests are related to the Block rewrite, some weird Api breakages and parser tests fix that need backports.
[1] http://ci.tesla.usability.wikimedia.org/cruisecontrol/ [2] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/92239
On Fri, Jul 15, 2011 at 8:42 AM, Ashar Voultoiz hashar+wmf@free.fr wrote:
From a quick look, broken tests are related to the Block rewrite, some weird Api breakages and parser tests fix that need backports.
For more on the "weird API breakage," see r91215 for an incomplete fix that shuts up the error (but as Bryan points out isn't strictly correct).
The block tests are all running fine in trunk, so we probably need to backport the rest of the block cleanup or back it out of REL1_18 entirely. I'm leaning towards the latter as I write this.
The LST failures need to be tracked down and possibly backported if necessary.
Thanks for your work on this, as always!
-Chad
On 15/07/11 14:56, Chad wrote: <snip>
For more on the "weird API breakage," see r91215 for an incomplete fix that shuts up the error (but as Bryan points out isn't strictly correct).
That one would need to be merged so :-)
The block tests are all running fine in trunk, so we probably need to backport the rest of the block cleanup or back it out of REL1_18 entirely. I'm leaning towards the latter as I write this.
The whole block rewrite should probably have been done in a branch. Would have made things easier. If it is stable enough in trunk we might as well backport it. If it requires too much work, back it off and release it with 1.19.
The LST failures need to be tracked down and possibly backported if necessary.
I am pretty sure I have fixed them myself. IIRC, the LST tests are broken following a change in the way <h2> are rendered. The backport should be easy.
Thanks Chad for the review!
wikitech-l@lists.wikimedia.org