Updating from x11-terms/rxvt-unicode-9.26 to x11-terms/rxvt-unicode-9.26-r1 broke reverse character display for me. This is somewhat critical, as rxvt uses reverse colors to display the cursor position. As the only difference I see is the 24-bit color patch, I wonder if this was tested with the following USE flags: font-styles mousewheel utmp wtmp -256-color -blink -fading-colors -gdk-pixbuf -iso14755 -perl -startup-notification -unicode3 -xft (and consequently, -24-bit-color) Downgrading back to 9.26 from git brought reverse character display back.
The 24-bit color patch breaks foreground colors when 256-color is enabled (+256-color -24-bit-color). I believe the root cause might be the same for both issues. Everything works fine with x11-terms/rxvt-unicode-9.26, but with 9.26-r1 only a few foreground colors are usable (looks like it's the basic 8 colors) and everything else is just printed in color 0, background colors work fine in both releases.
Well THAT is annoying. Could I possibly bother one of you to check the USE=-256-color +24-bit-color" combination as well?
Foreground colors seem to work OK with -256-color +24-bit-color, although colors are of course completely wrong (both foreground and background) due to different codes. The basic set of 8 colors works fine.
Just for completeness, +256-color +24-bit-color seems to be working fine. A wild guess (I don't know anything about rxvt-unicode internals): the #ifdef block for foreground color seems a bit too large to me as it results in noop without 24-bit color enabled. BTW, could the experimental patch be completely ignored without +24-bit-color? It seems wrong to apply the patch for everyone unconditionally.
(In reply to Jiri Denemark from comment #4) > BTW, could the experimental patch be completely ignored without > +24-bit-color? > It seems wrong to apply the patch for everyone unconditionally. Actually, the Gentoo policy is to apply patches unconditionally unless it can't be helped because it makes debugging easier. That said, applying this particular patch conditionally is exactly what I am aiming at; I just wanted to confirm 24-bit colour actually works for someone other than me in this version (otherwise I would have yanked this patch altogether).
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=bd5359cbda10c6ba292526e84f09f65dbd6b8bd1 commit bd5359cbda10c6ba292526e84f09f65dbd6b8bd1 Author: Marek Szuba <marecki@gentoo.org> AuthorDate: 2021-07-12 14:24:23 +0000 Commit: Marek Szuba <marecki@gentoo.org> CommitDate: 2021-07-12 14:32:55 +0000 x11-terms/rxvt-unicode-9.26: only apply 24-bit-colour patch when needed It's too aggressive, i.e. USE=-24-bit-color doesn't fully disable this feature. Closes: https://bugs.gentoo.org/801571 Signed-off-by: Marek Szuba <marecki@gentoo.org> ...-9.26-r1.ebuild => rxvt-unicode-9.26-r2.ebuild} | 39 +++++++++++++--------- 1 file changed, 24 insertions(+), 15 deletions(-)