Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 31 Jan 2020 17:32:11 +0000
From:      bugzilla-noreply@freebsd.org
To:        x11@FreeBSD.org
Subject:   [Bug 243649] [patch] devel/libepoll-shim: Update to 0.0.20191117 version to unbreak re-plugging input devices
Message-ID:  <bug-243649-7141-OjtRS0rHfr@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-243649-7141@https.bugs.freebsd.org/bugzilla/>
References:  <bug-243649-7141@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D243649

Michael Gmelin <grembo@FreeBSD.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #211221|0                           |1
        is obsolete|                            |

--- Comment #10 from Michael Gmelin <grembo@FreeBSD.org> ---
Created attachment 211223
  --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=3D211223&action=
=3Dedit
Patch to change libepoll-shim to use actual upstream, test targets, update =
to
20191117, more cosmetic changes

(In reply to Jan Beich from comment #9)

> Sort USES alphabetically. If unsure use ports-mgmt/portfmt then
> check the diff.
That's what I tried :p. TBH, I hoped that portlint would catch that.
Wasn't aware of portfmt, thanks!

>> +
>> +USE_LDCONFIG=3D	yes
>>  USE_GITHUB=3D	yes


> Maybe drop newline after USES per bug 231422. However, if you do then a=20
> newline before TEST_TARGET is also unnecessary.

Ok, why not :)


> At least running 32-bit Wayland clients on 64-bit compositor works fine.=
=20
> Testing runtime on the oldest supported/alternative architecture
> (like 11.3 i386 atm) is part of my pre-commit menu e.g., for www/firefox.

The first test on i386 fails on purpose, the second one is a sizeof
mismatch of some sort (confirmed on i386 on bhyve) - might very well
be a problem with the unit test itself (it also fails without
a message, which should be fixed upstream - I would leave that
to someone else or at least a different point in time though)

> Some tests seem to fail even on amd64 if run inside poudriere jail.
Which ones though and why/with which error message? I tested about
20 times on 11.3 and 12.1 and had no problems. Two things I noticed:

- The unit tests can't run in parallel (e.g. ctest -j4
  makes some tests fail).
- When testing i386 running on bhyve I've seen a few random failures
  on timing sensitive tests that didn't happen when running natively/inside
  a jail.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-243649-7141-OjtRS0rHfr>