04:27 <dnkl> ls-alh (for the logs): foot propagates the exit value of the shell. When you close the window via a compositor shortcut, or the CSD, foot sends SIGHUP to the shell. Bash, zsh and sh seem to respond by exiting with value 1. Fish exits with value 0

06:48 <dnkl> I suppose we could special case SIGHUP and just exit with 0 in that case

06:57 <dnkl> not sure I think it's a good idea, but made a patch and posted a PR for discussion: https://codeberg.org/dnkl/foot/pulls/1941

10:39 <Guest77> I need some help with `show-url-launch`. The urls are highlighted correctly but whenever I press the corresponding buttons, nothing happens. Running foot with log-level info also shows no relevant info. Are there any other ways of figuring out why this does not work?

10:58 <dnkl> Guest77: by default, foot runs xdg-open <url>. try that manually and see what happens

11:01 <Guest77> running xdg-open manually works as expected

11:27 <dnkl> which is?

11:28 <dnkl> you could also try strace:ing foot, maybe with follow-forks, to see what happens

12:57 <Guest77> dnkl basically firefox opens the file (which is not the correct program, but that's unrelated to foot).

12:57 <Guest77> I use foot in server mode, could this be related or cause this?

12:58 <dnkl> Guest77: yes, --server is likely the cause here. Try in a standalone foot instance to verify.

12:58 <dnkl> if that works, it's hard to say exactly what's wrong with your server setup. *something* in the environment variables perhaps

12:59 <dnkl> fwiw, I open links in firefox, with foot --server, without any issues

13:01 <Guest77> Hmm, I see. Would it me sufficient to execute `foot` in a foot client, and test url open in the newly spawned terminal?

13:05 <dnkl> if 'xdg-open' works from the foot client, I would assume running 'foot' from there, and then launching URLs would work

13:07 <dnkl> other things to try is setting "url.launch=firefox ${url}"

13:07 <dnkl> that is, skip xdg-open

13:09 <Guest77> actually setting the launcher to firefox did something: I now get the error `firefox <full path>: failed to spawn: no such file or directory

13:11 <Guest77> manually executing the same command works as the path is correct

13:12 <dnkl> sounds like Firefox isn't in your PATH variable, as seen by foot

13:23 <Guest77> hmmmmm, I think there is something else going on, I'll have to debug further. For now, I use the copy url command instead. Thanks a lot!

16:50 <ovf> so if i (by accident) am running under rlwrap, there's no single control sequence to say "disable fancy features i don't support", i have to explicitly disable e.g. bracketed paste mode that readline has enabled?

16:51 <ovf> (or perhaps i'm missing something about terminal replies that will allow me to swallow ones i don't recognise)

16:59 <dnkl> ovf: no, there's no such thing

17:00 <dnkl> an application can implement the VT parsing state machine. That would allow it to handle replies it recognizes and ignore ("swallow") all others

17:10 <ovf> thanks! i guess i was missing https://vt100.net/emu/dec_ansi_parser (for some reason that's much more difficult to read than a normal state transition diagram, but at least i can see that there is a grammar of csi replies that i can implement)

20:34 <rockorager> dnkl: any performance difference between using mode 2027 or explicity width? I am defaulting to use explicit width if it's available, so terminals (like foot) that support both will use explicit width

20:35 <rockorager> But since you would be the first to support both - I'd prefer to set a precedence if there is one that is a clear winner

20:35 <rockorager> s/explicity/explicit