Re: [PULL 00/18] testing and doc updates

[Thomas Huth <thuth@redhat.com>]

On 02/05/2023 13.46, Peter Maydell wrote:
> On Thu, 27 Apr 2023 at 16:47, Alex Bennée <alex.bennee@linaro.org> wrote:
>>
>> The following changes since commit 1eb95e1baef852d0971a1dd62a3293cd68f1ec35:
>>
>>    Merge tag 'migration-20230426-pull-request' of https://gitlab.com/juan.quintela/qemu into staging (2023-04-27 10:47:14 +0100)
>>
>> are available in the Git repository at:
>>
>>    https://gitlab.com/stsquad/qemu.git tags/pull-testing-docs-270423-1
>>
>> for you to fetch changes up to ef46ae67ba9a785cf0cce58b5fc5a36ed3c6c7b9:
>>
>>    docs/style: call out the use of GUARD macros (2023-04-27 14:58:51 +0100)
>>
>> ----------------------------------------------------------------
>> Testing and documentation updates:
>>
>>    - bump avocado to 101.0
> 
> Did this change the behaviour of 'make check-avocado' ?
> Previously it would continue the whole test run even if one
> test timed out; now it seems to abort the test run when the
> first test times out.
> 
> I don't object to the current behaviour, but is there some
> command line option I can use to force avocado to complete
> the whole test set even if there are failures or timeouts?
> 
> ('make check-avocado' for arm targets has some persistent
> intermittent timeout stuff still, so I generally prefer to
> eyeball the failures and ignore the stuff I know is flaky.)

I think I've seen it in the past, too:

 
https://lore.kernel.org/qemu-devel/7836af42-9969-a01a-3c8c-717987b0b27e@redhat.com/

... but after Cleber said that there is no change in behavior, I did not 
follow up. Anyway, it is likely related to that --failfast switch of 
Avocado, so that's likely the first thing to look for, I guess...

  Thomas