Re: The real reason why TAP testing isn't ready for prime time

On 07/29/2015 11:17 PM, Tom Lane wrote:
> Michael Paquier <michael(dot)paquier(at)gmail(dot)com> writes:
>> hamster has not complained for a couple of weeks now, and the issue
>> was reproducible every 4~6 days:
>> http://buildfarm.postgresql.org/cgi-bin/show_history.pl?nm=hamster&br=HEAD
>> Hence let's consider the issue as resolved.
> Nah, I'm afraid not. We are definitely still seeing a lot of unexplained
> pg_ctl failures; as a recent example,
>
> http://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=axolotl&dt=2015-07-29%2018%3A00%3A19
>
> Now maybe that's not the same thing as hamster's issue, but there's still
> something funny going on. That's why I've been bitching about the lack of
> debug information. I hope that we can make progress on understanding this
> once Andrew pushes out the new buildfarm release (especially since some of
> his machines are particularly prone to this type of failure, so I trust
> he'll update those critters promptly).
>
>

I want to deal with running the tests on MSVC first. Meanwhile I'll drop
the latest run_build.pl on axolotl.

Anyone who wants to be on the bleeding edge can grab it from
<https://raw.githubusercontent.com/PGBuildFarm/client-code/master/run_build.pl>

cheers

andrew