bug#50007: 28.0.50; normalize logic surrounding erc-server-reconnecting

[Amin Bandali <bandali@gnu.org>]

Hi Stefan, all,

Stefan Kangas writes:

> "J.P." <jp@neverwas.me> writes:
>
>> Tags: patch
>>
>> Knowing when we're reconnecting and having that logic be as sane and
>> straightforward as possible is (IMO) a critical component of unraveling
>> the buffer-association game I've been trying to drag everyone into.
>>
>> The existing approach is troublesome because it's basically inscrutable
>> in the sense of being too difficult to sit down and reason through once
>> session identities enter the fray (at least for a simpleton like me).
>> And though the solution I'm proposing may seem underwhelming, it wasn't
>> just intuited out of thin air but rather discovered with the backing of
>> behavioral tests like the ones included in these patches [1]. Thanks.
>>
>>
>> [1] https://debbugs.gnu.org/cgi/bugreport.cgi?bug=48598
>>     https://debbugs.gnu.org/cgi/bugreport.cgi?bug=49860
>
> Copying in the ERC maintainer Amin Bandali here.
>
> Amin, could you help review this patch?

I will review this patch -- and the other two you tagged me on -- but
most likely not before emacs-28 is cut: per J.P.[0] the tests covering
these three patches live in another, rather large, patch and I would
prefer to leave the merging of these to after the other large patch is
reviewed and merged (which may not be until a few more weeks).

[0]: https://lists.gnu.org/r/emacs-erc/2021-09/msg00004.html

Of course if you're volunteering to help test/review the large
patch(es) and these three that would be very welcome. :)