Re: sqlite3

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: emacs-devel@gnu.org
> Date: Wed, 29 Dec 2021 19:05:08 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > The "simple" test still fails sometimes.  I even enlarged the
> > sleep-for to 10 sec, and it still fails.
> 
> Huh.  Then it sounds like our analysis of the problem is wrong, and it's
> not failing on the time-less-p thing?

No, it was failing there, allright.

> Or...  can Windows (or our usage of it) cache the modification times
> for a long time?

I found the problem.  In multisession-tests.el we were waiting on the
wrong side of call-process: the time stamp of the value-file doesn't
change once call-process exits, so waiting there is futile.  Waiting
_before_ the call, OTOH, is what the doctor ordered, because it makes
sure the sub-process updates the value-file after enough time has
passed since the previous update to let 1-sec file-times resolution
get out of our way.

The test runs without any failures now.