Trygve Aaberge

This is more apparent if you enable `weechat.look.read_marker_always_show`. Then the read marker will be below the first line when you switch to the buffer, even though you've never read the...

We discussed changing `first_line_not_read` on IRC. After some back and forth regarding having one or two variables to track the read state, we realized that in order to support the...

> Note the two spaces before the closing bracket: they should not be displayed. This is not the fault of WeeChat, but a bug in kitty (https://github.com/kovidgoyal/kitty/issues/1978). If you try...

I bisected this issue to this commit in cpython: https://github.com/python/cpython/commit/de64e7561680fdc5358001e9488091e75d4174a3

I posted an issue to CPython for this: https://github.com/python/cpython/issues/116510 I also noticed that it's fixed in version v3.13.0a4.

This still reproduces for me on Arch Linux with current master (commit 2092715ac). Here's a backtrace for that version with line numbers: ``` (written by WeeChat 4.1.0-dev (git: v4.0.0-164-g2092715ac), compiled...

> Interesting, in fact it crashes with `/script load crash.pl` but not with `/perl load crash.pl` (with the same script being executed). Hm, for me it crashes with both.

> @trygveaa: and what's the content of your test script? I used exactly the scripts posted in the first message here. I see that the crash only occurs if an...

Merged buffers makes this even more complicated, as when `buffer_switched` or one of the commands are triggered, you have to check if there are any buffers merged with the buffer...

Yes, correct. It would be practical to get those values in the signal callback as well, so I don't have to look them up. And I guess it might be...