From: Thomas Huth <thuth@redhat.com>
To: Paolo Bonzini <pbonzini@redhat.com>, qemu-devel@nongnu.org
Cc: "Philippe Mathieu-Daudé" <philmd@redhat.com>,
"Daniel P . Berrangé" <berrange@redhat.com>,
"Wainer dos Santos Moschetta" <wainersm@redhat.com>
Subject: Re: [PATCH v1] gitlab-ci.yml: Speed up CI by using "meson test --no-rebuild"
Date: Tue, 26 Jan 2021 08:03:42 +0100 [thread overview]
Message-ID: <b6be5064-c3aa-8666-126d-e48e47392108@redhat.com> (raw)
In-Reply-To: <e8699ffe-3260-3033-cc13-ae9c3f86cbf1@redhat.com>
On 25/01/2021 10.51, Paolo Bonzini wrote:
> On 25/01/21 10:03, Thomas Huth wrote:
>> Currently, our check-system-* jobs are recompiling the whole sources
>> again. This happens due to the fact that the jobs are checking out
>> the whole source tree and required submodules again, and only try
>> to use the "build" directory with the binaries and object files
>> as an artifact from the previous stage - which simply does not work
>> anymore (with the current version of meson). Due to some changed
>> time stamps, meson is always trying to rebuild the whole tree.
>>
>> To fix this problem, we can use "meson test --no-rebuild" instead of
>> make check" to avoid rebuilding all binaries every time. However, the
>> iotests ("make check-block") are not run by "meson test", so we have
>> to execute these manually now. But instead of adding them to the same
>> job as "meson test", it's better to put them into a separate new job
>> instead, to keep things nicely running in parallel in the CI.
>> This saves ca. 15 - 20 minutes of precious CI cycles in each run.
>
> The reason why we're not using "meson test" is that the reporting is (still)
> inferior to what you get from "make check", especially with respect to which
> tests are failing. This is being tracked at
> https://github.com/mesonbuild/meson/issues/7830 and the last missing bits
> are at https://github.com/mesonbuild/meson/issues/8200 (after which we'll
> change the "meson test" command line to also include "meson test --verbose").
>
> However, for CI this is a minor issue because we can let GitLab parse the
> XML testing logs. Can you add something like this to the test jobs for v2?
>
> + artifacts:
> + when: always
> + paths:
> + - build/meson-logs/
> + reports:
> + junit: build/meson-logs/testlog.junit.xml
Ok, I've tried that but it also worked not quite as expected:
https://gitlab.com/huth/qemu/-/pipelines/246463068/test_report
The "check-*" jobs now show up in the test report page, but even though I've
made some tests failing (e.g. the check-system-centos job), the failing jobs
are showing up with "Failed: 0" there. Also the duration is always marked
with 0.00ms.
> Another possibility could be to use "make check NINJA=:". I am not sure if
> that works, but if it does it would be the smallest possible workaround.
It works! So since all the other options seem to have other drawbacks right
now, I guess that's the best way to go until we completely switch to "meson
test". Thus I've just sent a patch for that.
Thomas
next prev parent reply other threads:[~2021-01-26 7:06 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-01-25 9:03 [PATCH v1] gitlab-ci.yml: Speed up CI by using "meson test --no-rebuild" Thomas Huth
2021-01-25 9:51 ` Paolo Bonzini
2021-01-25 9:56 ` Daniel P. Berrangé
2021-01-25 10:36 ` Paolo Bonzini
2021-01-26 7:03 ` Thomas Huth [this message]
2021-01-26 8:48 ` Paolo Bonzini
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=b6be5064-c3aa-8666-126d-e48e47392108@redhat.com \
--to=thuth@redhat.com \
--cc=berrange@redhat.com \
--cc=pbonzini@redhat.com \
--cc=philmd@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=wainersm@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).