Thomas Huth <th...@redhat.com> writes:
> On 03/02/2021 20.32, Wainer dos Santos Moschetta wrote:
>> Hi,
>>
>> On 2/3/21 8:32 AM, Thomas Huth wrote:
>>> From: Philippe Mathieu-Daudé <phi...@redhat.com>
>>>
>>> Similarly to commit 8cdb2cef3f1, move the gprof/gcov test to GitLab.
>>>
>>> The coverage-summary.sh script is not Travis-CI specific, make it
>>> generic.
>>>
>>> Signed-off-by: Philippe Mathieu-Daudé <phi...@redhat.com>
>>> Message-Id: <20201108204535.2319870-10-phi...@redhat.com>
>>> [thuth: Add gcovr and bsdmainutils which are required for the
>>> overage-summary.sh script to the ubuntu docker file]
>> s/overage/coverage/
>>> Signed-off-by: Thomas Huth <th...@redhat.com>
>>> ---
>>> .gitlab-ci.yml | 12 ++++++++++++
>>> .travis.yml | 14 --------------
>>> MAINTAINERS | 2 +-
>>> scripts/{travis => ci}/coverage-summary.sh | 2 +-
>>> tests/docker/dockerfiles/ubuntu2004.docker | 2 ++
>>> 5 files changed, 16 insertions(+), 16 deletions(-)
>>> rename scripts/{travis => ci}/coverage-summary.sh (92%)
>>>
>>> diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
>>> index 7c0db64710..8b97b512bb 100644
>>> --- a/.gitlab-ci.yml
>>> +++ b/.gitlab-ci.yml
>>> @@ -468,6 +468,18 @@ check-deprecated:
>>> MAKE_CHECK_ARGS: check-tcg
>>> allow_failure: true
>>> +# gprof/gcov are GCC features
>>> +build-gprof-gcov:
>>> + <<: *native_build_job_definition
>>> + variables:
>>> + IMAGE: ubuntu2004
>>> + CONFIGURE_ARGS: --enable-gprof --enable-gcov
>>> + MAKE_CHECK_ARGS: build-tcg
>>
>> With build-tcg it generates an empty report, e.g.,
>> https://gitlab.com/wainersm/qemu/-/jobs/1005923421
>>
>> Shouldn't it run `make check`?
>
> D'oh, you're right. I think we need to run at least a "make check-unit"
> here. I'll rework my patch accordingly...
>
> By the way, it's broken on Travis since a long time, e.g. with version 5.0
> there is already only a stack trace:
>
> https://travis-ci.org/github/qemu/qemu/jobs/680661167#L8411
>
> Seems like nobody noticed this for almost a year now...
doh - the check_coverage was an after_success step so never influenced
the result. It was only a band aid really though - it would be better if
we published the html coverage report on gitlab's pages (like we now do
for annotated gtags source: https://qemu-project.gitlab.io/qemu/src/
>
> Thomas
--
Alex Bennée
