Hi Francesco,

Thanks for pointing out that the gtk3 support is still maturing.  I built
with gtk2 and indeed that fixes the major glitches.  (For completeness,
though, I'll note that the checker movement is still *slightly* glitchy,
but the effect is ephemeral and you have to be really sensitive to it to
notice the shards of checkers being displayed and then covered up.  Not a
huge deal, but it is there.)

The stock package for gnubg on Linux Mint also uses gtk3 and exhibits
similar problems.   I had built it in order to try out the 3d board, which
is not enabled in that stock build.  But that doesn't work with the mouse
on my setup, and it is not important to me anyway.

What IS important to me, and what gtk3 provides, is respect for desktop
themes, in particular Dark Mode.   The gtk2 build of gnubg insists on
showing the move history in a blindingly-bright white window, which makes
the game basically unplayable at night on a computer set up in dark mode
with an appropriate board.  While this might sound a bit picky, I think it
is important if gnubg doesn't want to fully commit to being a relic.

Anyway, I have attached to this email the match file for the scoring issue
I described.  (I'm not sure if it will make it to the mailing list, but you
said you'd like to see it so here it is.)   As you can see, there are four
games, and the gammon I described was played to completion on game 3, but
the fourth game seems to also be called game 3, and at least on my setup
gnubg is not scoring the gammon.


On Wed, Feb 25, 2026 at 1:00 AM Francesco Ariis <[email protected]> wrote:

> Hello Todd,
>
> Il 24 febbraio 2026 alle 17:04 Todd Doucet ha scritto:
> > PS:  I also find that 2D board animation is glitchy on my Linux Mint
> > system, and 3D board animation is extremely glitchy.  I spent a day
> > debugging it with chat-gpt's help and our conclusion was that it was
> > probably a gnubg problem, not a graphics stack problem.  I disabled the
> 3d
> > board entirely from the build that seemed to make the 2d board more
> > stable.  All told, it is pretty disappointing to have glitchy rendering
> in
> > 2025, although I suppose one could blame it on the display stack.  But
> > honestly, not likely.  Welcome to 1995.
>
> I had a similar problem.  Philippe Michel correctly diagnosed it:
>
>     https://lists.gnu.org/archive/html/bug-gnubg/2025-08/msg00004.html
>
> tl;dr: compile with Gtk2
>
>
> Regarding the main part of the message:
>
> Il 24 febbraio 2026 alle 17:04 Todd Doucet ha scritto:
> > The bug occurred during a 5-point match.  I was trailing 0-3 to 5, and I
> > had cubed to 2 and gnubg accepted.   I gammoned gnubg on that game, the
> > final roll was 2-2 and I had two checkers on each of the ace and deuce.
> So
> > a forced move, a gammon, and it should be 4-1 crawford.
> >
> > At the beginning of the next game, I noticed that gnubg thought the score
> > was still 0-3 to 5, but it also thought that the game was crawford.
> >
> > […]
> >
> > I don't subscribe to this list but if somebody wants a copy of the match
> > file, email me directly and I will send it.
>
> I am sure many of us are interested!  I do not know if you can post
> attachments to the list, but I would like to take a look at it
> —F
>

Attachment: gnubg-buggy.sgf
Description: application/go-sgf

Reply via email to