Open duckwork opened 2 years ago
From what I understand, it's supposed to be a combination of these choices:
The naming is a bit unfortunate. It might make more sense to instead expose it as a list of these choices or maybe even two separate variables. Does the current behavior still seem to be a bug to you?
A "bug," I'm not sure -- however, I would like the added choices. My personal want would be to kill the server and all chat buffers without confirmation, so maybe a list (a la modus-themes) of options might be good? Or two options might be easier to do.
Honestly two options, each taking a symbol or list of symbols, might be best. Like, circe-confirm
could take kill
, server-kill
, join
, or w/e, and circe-kill-chat-buffers-with-server
could be t or nil. Or .. something along those lines.
Hi there, and thanks for circe!
I'm writing an issue instead of submitting a PR because I want to make sure I'm not misunderstanding behavior here. According to the docstring of
circe-server-killed-confirmation
, a value ofkill-all
should kill all associated channel buffers without prompting the user. However, the code forcirce-server-killed
asks the user to confirm whenevercirce-server-killed-confirmation
is non-nil, including when it'skill-all
.However, when
circe-server-killed-confirmation
is nil, only the server buffer is killed. The way the options read, I thought nil would only kill the server buffer,kill-all
would kill the server buffer and associated buffers without confirmation, andask-and-kill-all
would kill the server buffer and associated buffers with confirmation.Have I misunderstood the logic of these choices, or is this a bug in
circe-server-killed
? From what I can tell, at this momentkill-all
andask-and-kill-all
are identical except in the question they ask the user, which doesn't really make any sense.Anyway, just let me know. I'm happy to submit a PR if it's a bug.