Re: master 262d0c6: Mark some tests as expensive

[Michael Albinus]

Stefan Kangas <stefan@marxist.se> writes:

> Hi Michael,

Hi Stefan,

>> Would be helpful. However, there have been discussions (and changes) of
>> the timeouts in the past, so I'm not sure we can reduce them w/o causing
>> damage. Please check the git history of autorevert-tests.el for such
>> changes, hopefully there are bug numbers in the git log you could consult
>> for the respective discussions. Otherwise, a search on the emacs-devel
>> and emacs-bugs MLs for the given timeframe of such a change might be helpful.
>
> So I have attached here a patch that employs the above mentioned
> techniques somewhat aggressively to reduce the running time as follows:
>
> Ran 7 tests, 7 results as expected, 0 unexpected (2020-10-18
> 19:38:29+0200, 2.323729 sec)

Thanks!

> Everything works fine here, but I don't know if any of this would cause
> any problems elsewhere.  One idea is to just push it and see what
> breaks, and then adapt accordingly.
>
> Please let me know what you think.

Yes, let's go this way. Maybe there will be problems in the remote case,
but this we could adapt later on.

> Also, I've had some trouble finding the past discussions about this.
>
> But I'm not sure if any of that is even relevant here.  Let me know if
> you have any ideas for what I could look for.

No, my history isn't organized as good as it should :-(

> * test/lisp/autorevert-tests.el (auto-revert--timeout): Make into
> defun and shorten timeout by a factor 10.

Why is it a defun now? Likely, I have overseen the reason ...

Best regards, Michael.

> Best regards,
> Stefan Kangas