07:21 <dnkl> ifreund: thanks!

13:29 <vyivel> seems like foot changes cursor (outline<->solid) depending on wl_keyboard focus rather than xdg_toplevel activated status, why? the latter seems more fitting to me

13:42 <ifreund> hmm, yeah doing this based on xdg_toplevel activated state instead would completely avoid the issue

14:18 <dnkl> ifreund: vyivel: design decision from long, long ago - I wanted the hollow vs non-hollow cursor to represent whether input was possible or not (i.e keyboard focus, rather than activated). Compositors usually have their own way of indicating "visual" focus anyway, no need to duplicate that with the cursor.

14:21 <vyivel> hm

14:23 <vyivel> weston sends leave+enter to update client's key state without sending a wl_keyboard.key event when handling keybindings, and i'm planning to do the same in wlroots (until wl_keyboard.keys arrives)

14:23 <vyivel> this might cause the cursor to flicker

14:25 <jpp1> https://manpages.debian.org/bookworm/foot/foot.ini.5.en.html

14:26 <dnkl> it's not super important to me, and in most cases, activated == keyboard focus. let me know if it does cause issues, and I'll change it. might change it anyway, but would still like to know if current behavior really does cause issues or not

15:00 <ifreund> dnkl: in any case, I confirmed that your patch fixes the segfault, thanks for the quick response!

21:48 <ashfield> if I hold down alt+tab I can make foot have focus but still have a hollow cursor, at least

21:50 <ashfield> foot 1.20.1 on kwin 6.2.4

21:51 <vyivel> okay seems like alactitty and kitty and weston-terminal and gnome-terminal behave the same way actually

21:51 <vyivel> oh well then

21:51 <vyivel> nvm scratch weston-terminal