Re: Some experience with the igc branch

[Eli Zaretskii <eliz@gnu.org>]

> Date: Sun, 29 Dec 2024 11:09:47 -0800
> Cc: eller.helmut@gmail.com, gerd.moellmann@gmail.com, pipcet@protonmail.com,
>  emacs-devel@gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> 
> On 12/28/24 21:47, Eli Zaretskii wrote:
> > Yes, that's the purpose of the global lock: a Lisp thread can only run
> > if it acquires the lock.
> 
> OS signals are therefore ignored if a non-main Lisp thread is buggy and 
> holds that global lock indefinitely. I assume this is the expected 
> fallout from such a bug.

"Ignored" in what sense?  AFAIU, we've just established that a signal
delivered to the main thread will cause the signal handler to run, and
after the handler returns, the main thread will return to waiting for
the lock.  Is that what you mean by "ignored"?  If so, why is this
"ignoring"?

And yes, if a non-main thread's Lisp function keeps running forever,
never calling any primitives that release the lock, that thread will
run forever, and the main thread will never get a chance.  Lisp
programs that cause this are considered buggy, of course.

> I don't see this issue documented explicitly in doc/elisp/threads.texi. 
> Should it be?

I will have an opinion when I better understand what "it" is in this
case.  I hope this is just a misunderstanding.