Re: Cygwin cleanup

From: Thomas Munro <thomas(dot)munro(at)gmail(dot)com>
To: Justin Pryzby <pryzby(at)telsasoft(dot)com>
Cc: vignesh C <vignesh21(at)gmail(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org, Andrew Dunstan <andrew(at)dunslane(dot)net>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Andres Freund <andres(at)anarazel(dot)de>
Subject: Re: Cygwin cleanup
Date: 2023-01-05 12:22:52
Message-ID: CA+hUKG+Q6DU4Ov9LrvUyDcF3oHS4KMRVSKmVGaeePq-kOyG9gA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Jan 4, 2023 at 3:25 AM Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
> On Tue, Jan 03, 2023 at 05:54:56PM +0530, vignesh C wrote:
> > Is there still some work pending for this thread as Andres had
> > committed some part, if so, can you post an updated patch for the
> > same.
>
> Thomas, what's your opinion ?

One observation is that your CI patch *nearly* succeeds, even if
hacked to turn on the full TAP tests, if applied on top of the
WaitEventSet-for-postmaster patch:

https://cirrus-ci.com/task/4533371804581888

No cigar though, it still failed a few times for me in the
subscription tests with EAGAIN, when accessing semaphores:

semctl(24576010, 14, SETVAL, 0) failed: Resource temporarily unavailable

That isn't an error I expect from semctl(), but from some cursory
research it seems like that system call is actually talking to the
cygserver process over a pipe (?) to implement SysV semaphores. Maybe
it couldn't keep up, but doesn't like to block? Perhaps we could try
to tune that server, but let's try the POSIX kind of semaphores
instead. From a quick peek at the source, they are implemented some
other way on direct native NT voodoo, no cygserver involved.

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...

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andrew Dunstan 2023-01-05 12:31:15 Re: Using AF_UNIX sockets always for tests on Windows
Previous Message Alexander Korotkov 2023-01-05 12:11:43 Re: POC: Lock updated tuples in tuple_update() and tuple_delete()