Subject: Re: remaining issues with gnome-2.20
To: Julio M. Merino Vidal <>
From: Matthias Drochner <>
List: tech-pkg
Date: 10/31/2007 20:14:58 said:
> IIRC, I worked on a patch for glib to workaround this last issue.
> Was it removed?

It was not removed, but it suffered from both a change in the
configure script and some botch in a recent update so it didn't
work until yesterday.
btw, what is the global variable good for? It makes the result
depending on whether the test is compiled with -export-dynamic
which is probably not intended.

> maybe that code using dlopen(NULL) (if you  
> really mean the dlopen call) could be migrated to use the module  
> management functions in glib.  The workaround is not perfect but lets  
> many stuff work.

Both glade3 and evolution use the glib module management functions.
The problem is that your patch does only fix the case of symbols
residing in dlopen'd modules but not that of symbols residing in
libraries loaded by dlopen'd modules. I can't imagine a solution
for that which is not a disgusting hack.

> As for all other problems, maybe it is worth to file PRs about them  
> so that they don't get lost?

I might do so about the v6only thing. Looking at that I found
a number of inconsistencies in the NetBSD tree, and there are also
political issues about the v6only sysctl setting... This might
mean some more or less fruitful discussion.
And the fix of course, but that wouldn't help for NetBSD<=4.
For the others - some application problems where we first should
make sure where the problems lies. HAL has been an SOC project
which I hope will lead to results. libgtop is a SMOP, and it already
has been worked on.

best regards

Forschungszentrum Juelich GmbH
52425 Juelich

Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDirig'in Baerbel Brumme-Bothe
Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender), Dr. Ulrich Krafft (stellv.