bug#62694: 30.0.50; eglot-tests fails with recent pylsp

[Eli Zaretskii]

> From: Michael Albinus <michael.albinus@gmx.de>
> Cc: 62694@debbugs.gnu.org,  João Távora
>  <joaotavora@gmail.com>
> Date: Fri, 07 Apr 2023 14:13:32 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> Hi Eli,
> 
> >> Alternatively, if helping me with that check or adding pip install
> >> or not installing debian's pylsp on EMBA is too much to ask, then
> >> just delete all the pylsp tests.
> >
> > Michael, does EMBA run all the tests, or does it skip, say, the
> > 'unstable' ones?  If the latter, how about marking these specific
> > tests as 'unstable'? would that solve the problem?
> 
> Sure, the :unstable tag suppresses a test for all "make test" invocation
> variants. I've proposed this a while ago for some eglot tests, but João
> didn't agree because "the tests are stable".

Given João's reluctance to help you to find a better solution, I'm
afraid we will have to overrule his disagreement.  And I'm not sure he
is still in disagreement, given that he just proposed removing them
completely as a possible solution.  Running these tests is easy, even
if they are marked unstable, so the problem is not a serious one for
those who do want to run them.

> > Alternatively, we could mark these tests as known to fail, right?
> 
> This would be unfortune if there is a pylsp version on the machine which
> let the tests succeed.

Why unfortunate? what am I missing?