GNU bug report logs - #72004
30.0.50, master: 'erc--check-prompt-input-for-multiline-blanks' test fail

Previous Next

Full log

Message #17 received at 72004 <at> debbugs.gnu.org (full text, mbox):

From: "J.P." <jp <at> neverwas.me>
To: Andrea Corallo <acorallo <at> gnu.org>
Cc: Eli Zaretskii <eliz <at> gnu.org>, emacs-erc <at> gnu.org, 72004 <at> debbugs.gnu.org
Subject: Re: bug#72004: 30.0.50, master:
 'erc--check-prompt-input-for-multiline-blanks' test fail
Date: Thu, 11 Jul 2024 11:26:15 -0700

Andrea Corallo <acorallo <at> gnu.org> writes:

> I buy the theory of this being related to trampoline generation under
> high system load, looking at my logs I see this test failing only with
> native compilation an passing without.
>
> I'm wondering if we could work around the issue somehow 🤔

I'm afraid I can offer next to nothing when it comes to serious insights
about Emacs internals and related magic 😞

>
> Also I'm not sure the right tag is :expensive, shouldn't be :unstable?

If we want to prevent the test from running completely (at least on
EMBA), then :unstable would indeed make sense. However, if we'd like the
test to run on the 3x daily expensive pipelines, I'm fairly confident my
latest change sidesteps the issue. It now waits to start the subprocess,
which was previously too short-lived, until after the macro has done its
(potentially time-consuming) advising. Thus, the liveliness check that
was signaling the error should now always find a running process.
(Although, for good measure, I also lengthened the timeout from 10s to
5m.) All this said, I'm happy to change it to :unstable or try a safer
approach, such as mocking `process-status'. Thanks.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.