Re: bug#70127: 29.3; ERC 5.6-git: /ignore timeouts are not cancelled/reset

[J.P.]

Alcor via General discussion about ERC <emacs-erc@gnu.org> writes:

> Severity: wishlist
>
> On Erc (5.6snapshot0.20240323.151032):
>
> 1. /IGNORE a specific nick with an explicit timeout (e.g. 5m)
> 2. /IGNORE the same nick again with a different timeout (e.g. 10m)
>
> Expected behavior:
>
> The /IGNORE'ed nick is unignored after 10 minutes.
>
> Observed behavior:
>
> The /IGNORE'ed nick is unignored after 5 minutes, afterwhich it is
> unignored again after 10 minutes. The latter command is a no-op though,
> as the nick would be already unignored at that point.

I'm taking the liberty of "extending" this bug to include a somewhat
related issue. I suppose that's bad practice, but it'll allow us to roll
the original request into the fix without much hassle.

>From Emacs -Q:

1. M-x erc RET localhost RET RET RET RET
2. /ignore foo RET
3. Hit RET to forgo adding a timer (doesn't impact result)
4. /unignore fool RET ; note the trailing "l"
5. Answer "y" when asked whether to remove "foo"
6. /ignore RET
7. Notice "foo" is still listed and still appears in `erc-ignore-list'

The attached patch tries to address this issue as well as the original
UX problem regarding rustic timeout handling, with ideas borrowed from
Alcor's gist shared on the emacs-erc list.

Not to keep harping on this, but I'd much rather we rewrite the ignore
and lurker stuff completely and move it to a new module in a new file.
Anyone feeling up to the challenge, please step forward. Thanks.

0001-5.6-Allow-updating-of-IGNORE-timeouts-in-ERC.patch
Description: Text Data