bug#39413: 26.2; Emacs gets hung

[Eli Zaretskii]

> Cc: larsi@gnus.org, npostavs@gmail.com, 39413@debbugs.gnu.org,
>  chiaki.ishikawa@ubin.jp
> From: "ISHIKAWA,chiaki" <ishikawa@yk.rim.or.jp>
> Date: Wed, 11 Aug 2021 22:41:56 +0900
> 
> What about the dump below, which was in the *Messages* buffer following 
> dump I sent.
> I am afraid I picked up the wrong dump in my previous e-mail.
> 
> Now, is "gc-elapsed" the elapsed time spent in the GC routine?
> If so, the two samples at the end of the dump here certainly look 
> suspicious.

Yes, they show a couple of GCs that took about 150 to 200 sec.  That
is indeed excessive, but gc-elapsed reports the wall-clock time, not
the CPU time, so if Emacs wasn't getting processing resources to run,
which could happen on a stressed system, that could explain the long
GC times.  Note that other GC cycles, both before and after the 2
problematic ones, took much less time.  And the statistics returned by
memory-use-counts doesn't show any excessive number of Lisp objects;
in fact, my current Emacs session, which runs for 9.5 days, has an
order of magnitude more objects than what you show, and yet GC cycles
here are barely perceptible.