On 2023-01-05 Th 16:39, Thomas Munro wrote:
> On Fri, Jan 6, 2023 at 1:22 AM Thomas Munro <thomas.munro@gmail.com> wrote:
>> https://cirrus-ci.com/task/5142810819559424 [still running at time of writing]
>>
>> Gotta run, but I'll check again in the morning to see if that does better...
> Yes! Two successful runs with all TAP tests so far. So it looks like
> we can probably stop lorikeet's spurious failures, by happy
> coincidence due to other work, and we could seriously consider
> committing this optional CI test for it, much like we have the
> optional MSYS build. Any interest in producing a tidied up version of
> the patch, Justin? Or I can, but I'll go and work on other things
> first.
>
> I pushed a change to switch the semaphore implementation. I haven't
> personally seen that failure mode on lorikeet, but I would guess
> that's because (1) it's only running a tiny subset of the tests, (2)
> it crashes for the other reason with higher likelihood, and/or (3)
> it's not using much concurrency yet because the build farm doesn't use
> meson yet.
OK, should I now try re-enabling TAP tests on lorikeet?
cheers
andrew
--
Andrew Dunstan
EDB: https://www.enterprisedb.com