'make check' fails when building with gcov code coverage

[Maarten Hoes]

On Wed, Jun 1, 2022 at 5:22 PM Stephan Bergmann <sbergman at redhat.com> wrote:

>
> I agree with you that the status quo would not be useful (and don't
> think you are annoying in any way).
>
> I still strongly (but naively, without digging into the code in any way)
> assume that those tests are not failing for you all the time because
> they would deterministically fail for gcov builds in general, but "just"
> because your gcov builds are sufficiently slow.
>
> Ideally, the authors of those tests could rework them so that they don't
> depend on performance characteristics that certain builds cannot meet.
>


Okay.

Well, I guess that then means that I should let this go for now, and I'm
now off to the next step of my Grand Masterplan (tm) of trying to get
regular automated lcov reports reinstated: Seeing if I can rework the
existing script [1] so that you only have to run it once (with supplying
obvious flags of where the sourcecode lives, and where you want the
generated lcov 'tracefiles' and html report to go), instead of having to
break things up in multiple 'steps' as it is now written.

If/when I have something that does what it should as far as I can
test/tell, I'll put the modified version of the script (and README) up for
review in Gerrit, and ask around on this list if/who is interested in
reviewing the modified version.

Thanks.



[1]
https://git.libreoffice.org/buildbot/+/refs/heads/master/lcov-report/


- Maarten
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice/attachments/20220601/f2b91f61/attachment.htm>