: This reproduces constantly for me, I even rolls back to some (random) old
: commit: 9acc6539959 (in last September) and it still reproduces.
FWIW: The oldest known failure of TestGroupFacetCollector.testRandom in my
archive is the week of 2021-05-31
: Found another similar failure:
:
FWIW: This same exact failure (same assertion msg) happened on main
recently as well.
I've included some log snippets shouwing randomization details of both
failures below.
Historically this test (with this name) has existed unmodified since long
before i started keeping records of failure
>
> We should sometimes run the OpenJ9 to make it explicit that it doesn't
> work.
>
It doesn't - there are occasional hiccups that repeat from time to time and
cannot be reproduced with hotspot. We can create an umbrella issue for
these errors and just append logs there, perhaps it'll be of some
If I have some time, I will update to newer version.
We should sometimes run the OpenJ9 to make it explicit that it doesn't
work. Maybe it helps to improve their code. Unfortunately I don't know
how to best report issues like this.
Uwe
Am 01.06.2023 um 12:37 schrieb Dawid Weiss:
There are
There are frequent odd errors on J9. I think we should turn it off and not
recommend it to people, something is wrong with it.
On Thu, Jun 1, 2023 at 12:00 PM Michael McCandless <
luc...@mikemccandless.com> wrote:
> Does not reproduce for me, but, it looks like it's IBM's J9: Java:
>
Does not reproduce for me, but, it looks like it's IBM's J9: Java:
64bit/openj9/jdk-17.0.5 -XX:+UseCompressedOops -Xgcpolicy:balanced
Maybe a J9 bug...
Mike McCandless
http://blog.mikemccandless.com
On Thu, Jun 1, 2023 at 5:54 AM Michael McCandless
wrote:
> Hmm, maybe it's a stdout/stderr
Hmm, maybe it's a stdout/stderr issue -- the juicy failure details come out
in stderr:
https://jenkins.thetaphi.de/job/Lucene-9.x-Linux/10811/testReport/junit/org.apache.lucene.index/TestIndexFileDeleter/testExcInDecRef/
Mike McCandless
http://blog.mikemccandless.com
On Thu, Jun 1, 2023 at
It used to be that these build failure emails included the "Reproduce with:
...", thanks to Steve Rowe's incredible regular expression (the largest
real-world regexp I have ever seen!!!).
But I don't see it here anymore. Does anyone know why/when it broke?
I know I can click through to the
This reproduces constantly for me, I even rolls back to some (random) old
commit: 9acc6539959 (in last September) and it still reproduces.
Found another similar failure:
https://ci-builds.apache.org/job/Lucene/job/Lucene-Check-main/8835/
The error seems to be a more common one but somehow (so far)