bug#28596: 26.0.60; Gnus gets into an ambiguous plugged/unplugged state that prevents retrieval of news

[N. Jackson]

At 19:54 +0200 on Thursday 2017-12-28, Eli Zaretskii wrote:
>
>> From: Lars Ingebrigtsen <larsi@gnus.org>
>> Cc: nljlistbox2@gmail.com,  eric@ericabrahamsen.net,  28596@debbugs.gnu.org
>> Date: Thu, 28 Dec 2017 17:29:34 +0100
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> > So what would be the best way of making progress with this bug?
>> 
>> What happened to the patch set to fix various irregularities with
>> accept-process-output that was going around the other month?  Was it
>> merged?  I was kinda waiting for that to go in to see whether that made
>> everything work before poking into this any further...
>
> You mean the thread Re: wait_reading_process_ouput hangs in certain
> cases (w/ patches)?  We are still waiting for the updated version of
> the patch (I've just sent a reminder).
>
> But in any case, that patch was for master, not the release branch.
> We could see if it also fixes this problem, and maybe consider
> backporting if it does.  But I hoped there could be a safer fix for
> this bug.

I do not think that this bug (with the ambiguous Gnus
plugged/unplugged state) is the same as the
wait_reading_process_output bug, and I'm not entirely convinced
that the two bugs are related, but of course I might be missing
something.

I can now reproduce this bug (#28596) at will -- and without any
apparent hang at all. So I think the hang I was experiencing while
trying to get Gnus to behave itself was a red herring -- quite
possibly I was running into the other bug while I was trying to
get myself out of this one.

And IIUC the wait_reading_process_output bug has been in Emacs for
a while (forever?) whereas this bug exists in Emacs 26 but I have
never seen it and cannot reproduce in Emacs 25.

N.