guile-user
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: GNU G-Golf 0.8.0-rc-3 available for testing


From: David Pirotte
Subject: Re: GNU G-Golf 0.8.0-rc-3 available for testing
Date: Sat, 4 May 2024 18:39:44 -0300

Hi Florian,

> Hi David, thank you for your continuing responses.

No problem. I am also interested in fixing this bug

> I will try to get a backtrace with GTK debug symbols by patching
> Guix to add a debug output to GTK, in the hope that I patch the right
> GTK.

It is definitely not a GLib/GObject nor a Gtk/Gdk/Gsk bug - i need the
full backtrace to look at the calls, from the one that segfault, a LIFO
backtrace of all calls, so i can try to see where it goes wrong ...

> However, my believe is that the error is in G-Golf

Yes, we are trying to find it, it's actually annoying that it works
fine in debian :), but we'll get there ...

> closures/callbacks/vfuncs (even though the hl-api tests do pass

there is no vfunc test in the test-suite - the tests are all those
examples that use them ... which is fine, but once i find the source of
this bug, i might add a more specific test to the test-suite.

> when run by Guix build or manually), specifically when I run valgrind
> on guile where I load and run drawing-widget.scm ...

I don't think it's a guile error either - definitely a g-golf error
that appears in guix and unfortunately not in debian ...

> it only shows errors from Boehm GC (at least mostly false-positive)

I'll double check that those snapshot vfunc (closure pointers) do not
become unreachable, i think i did check that but ... that would
definitely segfault the example if not ... it could be that i miss
checked how those are cached.

> I cannot tell you my host distro because there is none (there could
> be a foreign distro, but I have none).  It is Guix System, that is
> Guix all the way down.

Oh, ok -

> To build G-Golf manually, I can and had patched the its
> g-golf/init.scm to use /gnu/store/xxxxxxx/lib/libgirepository-1.0.so
> file names like the guile-g-golf package at
> <https://git.savannah.gnu.org/cgit/guix.git/tree/gnu/packages/guile-xyz.scm>
> does, but it will not be useful.

Imo, you should locally patch your guix image so it has one gdk-pixbuf
lib installed  keep the one that has the debug symbols - then comment
guile-g-golf in the guile-xyz module, create a g-golf module that you
can 'play with', and 'at will.constantly' make it point to the latest as
we a re working on tracking those bugs -

> Setting the GTK_A11Y env var does not make a difference (except the
> warning is gone) and on my x86_64 GNOME setup there never is a warning
> (although GTK_A11Y is unset, but d-bus is there).

I never thought in any moment that this warning was in any way 'linked'
to the problem, i did say and re-state here, that guix should be fixed
so it does not complain it can't find d-bus - terrible imo, it can lead
to other problems - till then, locally define the env var so we get
this annoying warning out of the way

David

Attachment: pgpPVQQjcnDiD.pgp
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]