Le mer. 20 déc. 2023 à 21:18, Mark Brown a écrit :
> clone 1059165 -1
> reassign -1 nodejs
> retitle -1 autopkgtest failures on i386
> found -1 18.19.0+dfsg-6
> block 1059165 by -1
> kthxbye
>
> On Wed, Dec 20, 2023 at 08:15:31PM +0100, Paul Gevers wrote:
>
> > The Release Team considers packages that are out-of-sync between testing
> and
> > unstable for more than 30 days as having a Release Critical bug in
> testing
> > [1]. Your package src:zlib has been trying to migrate for 32 days [2].
> > Hence, I am filing this bug. The version in unstable triggers autopkgtest
> > failures in multiple packages (although I suspect that the current dolfin
> > issues are due to it being flaky). The failure for burp has already a bug
> > report against that package, which leaves nodejs on i386.
>
> ...
>
> > This bug will trigger auto-removal when appropriate. As with all new
> bugs,
> > there will be at least 30 days before the package is auto-removed.
>
> Not sure that's likely in the case of zlib?
>
> > If you believe your package is unable to migrate to testing due to issues
> > beyond your control, don't hesitate to contact the Release Team.
>
> There are non-technical issues with me doing active work on nodejs
> package but from a quick glance the log does not seem particularly
> plausibly related to zlib, and I note that the failures are
>
>not ok 498 parallel/test-debugger-heap-profiler
>not ok 962 parallel/test-fs-utimes-y2K38 # TODO : Fix flaky test
>
> the second of which especially doesn't inspire confidence that this is
> due to zlib rather than general updates to unstable setting off an
> already flaky test (eg, the kernel changed timing?). Full log is:
>
>https://ci.debian.net/packages/n/nodejs/testing/i386/41176091/
>
> and looking at:
>
>https://ci.debian.net/packages/n/nodejs/testing/i386/
>
> there seem to be a number of packages triggering what from spot checks
> look to be the same or similar issues in nodejs in testing.
>
> I frankly don't really know what I'm supposed to do with this, the test
> results look like noise as far as zlib is concerned so I don't see
> anything to fix or investigate in the package itself. AFAICT bugs don't
> get filed for autopkgtest failures like they do for build failures so
> perhaps this was just missed up until now?
NODEJS will probably pass with a retry request
The only test that failed here was parallel/test-debugger-heap-profiler
and it doesn't depend on zlib. It is somewhat flaky but usually not on i386.
I sent a retry request for nodejs/i386 testing to debci. There is a good
chance it'll pass this time.
Meanwhile, next nodejs upload will mark that test as flaky on all platforms.
BURP wrong zlib version check in the failing test - this could be NMUed
./test-suite.log
99%: Checks: 252, Failures: 1, Errors: 0
utest/test_fzp.c:95:F:Core:test_fzp_gzseek:0: Assertion 'fzp_seek(fzp,
d->pos, SEEK_SET)==-1' failed
It fails because the condition is still fulfilled, but it shouldn't be:
START_TEST(test_fzp_gzseek)
{
if(version_to_long(ZLIB_VERSION) <= version_to_long("1.2.3"))
fzp_gzopen_old_zlib_seek_hack=1;
do_seek_tests(fzp_gzopen);
}
END_TEST
DOLFIN has a single test failure, that is odd and unrelated as well - this
could be NMUed
224s with Timer(task) as t:
4608
5224s sleep(0.05)
4609
5224s > assert t.elapsed()[0] >= 0.05
4610
5224s E assert 0.04 >= 0.05
Jérémy
