Re: collate.linux.utf8 test coverage

On 04/27/2015 02:39 AM, Michael Paquier wrote:
> On Sat, Apr 25, 2015 at 11:36 PM, Michael Paquier
> <michael.paquier@gmail.com> wrote:
>> On Sat, Apr 25, 2015 at 4:51 AM, Andrew Dunstan <andrew@dunslane.net> wrote:
>>> The optional buildfarm module that runs this test was broken by commit
>>> dcae5faccab64776376d354decda0017c648bb53
>>>
>>> Since nobody has responded to my complaint about this, I have disabled it on
>>> crake, the only buildfarm machine that has actually been running the test,
>>> so we now have no buildfarm coverage for it.
>> Commit dcae5fac has switched in pg_regress --temp-install to
>> --temp-instance, --psqldir to --bindir, and has removed --top-builddir
>> and --extra-install. After looking at TestCollateLinuxUTF8.pm, I think
>> that you need to use the one-liner attached.
> I am giving a shot at this test in dangomushi by doing the following:
> - Add TestCollateLinuxUTF8 in the list of modules
> - Add 'en_US.utf8' in the list of locales
> - Enable tr_TR.utf8, sv_SE.utf8 and de_DE.utf8.
> - Patching the module with the patch upthread.
> As far as I can see the patch is working, I just missed to install a
> couple of locales on this machine, making the test fail for the first
> time, but things will hopefully stabilize soon.
> Regards,


Yeah, thanks. crake has been back to running this check for a few days - 
see for example 

<http://www.pgbuildfarm.org/cgi-bin/show_stage_log.pl?nm=crake&dt=2015-04-26%2021%3A09%3A47&stg=install-check-collate-en_US.utf8>

The patch needed a little adjustment so it runs with the right switches 
in the right branch, I think. See 
<https://github.com/PGBuildFarm/client-code/commit/2176bba516f9effd17fceec45e93218ef8a21e8d>

cheers

andrew