Timer related tests with huge build are failing in Travis CI

[Yegappan Lakshmanan]

Hi all,

It looks like after patch 8.2.0014, the timer related tests
(Test_timer_oneshot(),
Test_timer_repeat_three(), Test_timer_repeat_many(), etc.) in test_alot.vim are
failing in Travis CI. I am not able to reproduce these failures locally in
my setup. The Appveyor and Cirrus CI builds are passing. I am not sure
which change is causing these test failures.

- Yegappan

[Dominique Pellé]

I'm not able to reproduce it either on Linux. It only failed on macOs huge.
Hopefully someone with a mac can reproduce and debug it.

I don't see anything related to timers in 8.2.0014.
Perhaps this is wrong in Test_whichwrap but it looks
unrelated to the failure:

set whichwrap+=h
normal dh
set whichwrap-=h

'whichwrap' should probably be restored with:

set whichwrap&

Regard
Dominique

[Bram Moolenaar]

Dominique wrote:

> Yegappan Lakshmanan <yega...@gmail.com> wrote:
>
> > Hi all,
> >
> > It looks like after patch 8.2.0014, the timer related tests
> > (Test_timer_oneshot(),
> > Test_timer_repeat_three(), Test_timer_repeat_many(), etc.) in test_alot.vim are
> > failing in Travis CI. I am not able to reproduce these failures locally in
> > my setup. The Appveyor and Cirrus CI builds are passing. I am not sure
> > which change is causing these test failures.
> >
> > - Yegappan
>
> I'm not able to reproduce it either on Linux. It only failed on macOs huge.
> Hopefully someone with a mac can reproduce and debug it.
>
> I don't see anything related to timers in 8.2.0014.

This appears to happen on the Mac tests once in a while. I suspect it
just happens because these machines are very busy, the test job doesn't
run fast enough.

These tests are intentionally timing specific. We could fake the time,
but then what are we testing?

> Perhaps this is wrong in Test_whichwrap but it looks
> unrelated to the failure:
>
> set whichwrap+=h
> normal dh
> set whichwrap-=h
>
> 'whichwrap' should probably be restored with:
>
> set whichwrap&

I'll correct that. I also notice this test function name is used twice,
I'll rename one.

--
FIRST HEAD: All right! All right! We'll kill him first and then have tea and
biscuits.
"Monty Python and the Holy Grail" PYTHON (MONTY) PICTURES LTD

/// Bram Moolenaar -- Br...@Moolenaar.net -- http://www.Moolenaar.net \\\
/// sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\ an exciting new programming language -- http://www.Zimbu.org ///
\\\ help me help AIDS victims -- http://ICCF-Holland.org ///

[Yegappan Lakshmanan]

Hi,

On Tue, Dec 17, 2019 at 10:15 AM Dominique Pellé
<dominiq...@gmail.com> wrote:
>
> Yegappan Lakshmanan <yega...@gmail.com> wrote:
>
> > Hi all,
> >
> > It looks like after patch 8.2.0014, the timer related tests
> > (Test_timer_oneshot(),
> > Test_timer_repeat_three(), Test_timer_repeat_many(), etc.) in test_alot.vim are
> > failing in Travis CI. I am not able to reproduce these failures locally in
> > my setup. The Appveyor and Cirrus CI builds are passing. I am not sure
> > which change is causing these test failures.
> >
> > - Yegappan
>
> I'm not able to reproduce it either on Linux. It only failed on macOs huge.
> Hopefully someone with a mac can reproduce and debug it.
>

Thanks for the reply. I do my development on a MacOS system and
I couldn't reproduce the test failure on this system with a huge build.

>
> I don't see anything related to timers in 8.2.0014.
> Perhaps this is wrong in Test_whichwrap but it looks
> unrelated to the failure:
>
> set whichwrap+=h
> normal dh
> set whichwrap-=h
>

This should undo the first change to 'whichwrap'. This line was copied
from the original test.

Regards,
Yegappan