bug#70840: 29.3; ERC 5.5.0.29.1: erc-kill-buffer-on-part kills server buffer

[J.P.]

"J.P." <jp@neverwas.me> writes:

> "J.P." <jp@neverwas.me> writes:
>
>> There may well be other issues here I've not yet considered, so if you
>> can think of any, please share. Also, if you're up for taking this on
>> (not necessarily in the way I've described), please feel free to send a
>> patch. The same goes for anyone else out there. I can definitely help
>> with the leg work and test coverage if that makes things easier.
>> Otherwise, I may not be able to get to this in time for the next release
>> (not that there's any rush).
>
> I've gone ahead with a provisional patch (attached) that incorporates
> the points raised so far. Anyone interested, please take a look.

After thinking a bit more on this, I've flip-flopped entirely on the
notion of running `erc-part-hook' in `erc-kill-channel' and inhibiting
it in `erc-server-PART' with a would-be null buffer argument. Basically,
I've come to believe the current faulty behavior has been around too
long to justify such a change in the name of fixing a bug that's only
tangentially related. Instead, I think we should introduce a flag that
`erc-kill-channel-hook' members can use to detect when `erc-server-PART'
is killing a buffer on behalf of the option `erc-kill-buffer-on-part'.
Updated patch attached. Thanks.

0000-v1-v2.diff
Description: Text Data

0001-5.6-Don-t-kill-server-buffer-with-erc-kill-buffer-on.patch
Description: Text Data