How to do a "Profile-guided optimization" build of the #TerminalEmulator foot.
A PGO build is ~30% faster than a "normal" release build. The exact numbers depend on what you're running, CPU etc etc. Point is: it is much faster.
Turns out to be a regression specific to the O2+PGO combo.
The work around was simple enough: use -O3 instead of -O2.
Seeing a ~20-40% performance drop in PGO builds with #gcc 10.1, compared to gcc 9.3 :/
Nicer looking buffer dividers?
(set-display-table-slot standard-display-table 'vertical-border ?│)
Don't like the blinking cursor?
(setq visible-cursor nil)
Should the auto-detection not work, you can also set 'xterm-extra-capabilities' to force-enable them.
For example 'modifyOtherKeys', to map even more keys.
Or 'setSelection' and/or 'getSelection' to use the terminal's OSC52 support to directly integrate with the system clipboard without using e.g. xclip or wl-copy)
Better Emacs terminal experience in non-xterm (but xterm compatible) terminals, using 'foot' as an example:
(add-to-list 'term-file-aliases '("foot" . "xterm"))
This will make Emacs define key sequences for all function keys, and modifier combinations (M-..., C-...).
It will also try to query the terminal for version info to enable more key sequences and features.
Only sway itself has been "fixed". Swaybar and swaynag etc don't even compile.
I got fed up with pango's (not) rendering of bitmap fonts.
Since Sway is just about the only thing I use which depends on pango, I created a Sway fork where I render the text in the title bars myself.
Top it off with a misspelled commit message 😂
Well this was embarrassing...
@kde looks like kwin/Wayland clears (forgets) surface damage when a buffer is a attached.
I.e you have to *first* attach a buffer, *then* apply damage (and finally commit).
@Linuxkompis från hemsidan: "IRC må vara gammalt, men det är ett väl fungerande alterantiv för..."
Would you like a super fast, lightweight, Wayland native, no fuzz no frills terminal emulator?
Linux Geeks doing what Linux Geeks do..