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

Re: Neat gtk/gdk-imlib pain


>>>>> Jason Gunthorpe writes:

 >> We could change the SONAME to designate the break in
 >> compatibility.  But then it wouldn't match upstream.  It really
 >> isn't the upstream maintainer's fault that we released one version
 >> compiled against libglib 1.0, and one against libglib 1.1.

 JG> Exactly what I mentioned during the libtool thread, however that
 JG> doesn't seem to be important. Ideally we could have a
 JG> /usr/lib/gnome-compat dir <lol>

Bwahahahahah! :D

 JG> but it seems our dynamic linker isn't smart enough to handle that
 JG> for anything but the libc case (right??)

Yes, exactly.

All Alexandre, Marcus, and I have been saying for the last while is
that by choosing to implement the ld.so hack for the libc5->libc6
transition, we expended a huge amount of effort to solve only one out
of a very broad class of problems, and thereby set a very bad example
for all the other problems of this class.

If we made every library automatically bump its soname when it was
linked against different dependents, and we had managed to convince
our Chapeau Rouge compatriots to do the same, we would all be running
nice, compatible systems without any weird linker magic.

There is apparantly an EGCS patch called libapi, available in the
Debian egcs package, which is supposed to implement the above.
Adopting and improving this patch would definitely solve your GNOME
problems, Jim.

 Gordon Matzigkeit <gord@fig.org> //\ I'm a FIG (http://www.fig.org/)
    Lovers of freedom, unite!     \// I use GNU (http://www.gnu.org/)
[Unfortunately, www.fig.org is broken.  Please stay tuned for details.]

Reply to: