Am Mittwoch 14 Mai 2008 18:40:56 schrieb Bryce Harrington:
> Stefan, thanks for the fix. Do you have more background on how the bug
> manifested itself? Was it an issue with something linking against
> libxfont? If so, perhaps we need a bug against that package so we can
> fix it there.
In theory, with -Bsymbolic-functions, you cannot override functions defined in
libxfont1 by programs/libraries linking against libxfont1.
However I'm still trying to bring this theory down to c-code, which will in
fact show different behaviour for different linker options.
I guess it's possible to find out who's the "evildoer" (it might even be done
on purpose) by comparing symbols of libxfont1 with symbols of libs/programs
linking against it, but I'm far from sure there.
Once I know more, I'll post updates to bug #230460.
Hi Bryce,
Am Mittwoch 14 Mai 2008 18:40:56 schrieb Bryce Harrington:
> Stefan, thanks for the fix. Do you have more background on how the bug
> manifested itself? Was it an issue with something linking against
> libxfont? If so, perhaps we need a bug against that package so we can
> fix it there.
In theory, with -Bsymbolic- functions, you cannot override functions defined in
libxfont1 by programs/libraries linking against libxfont1.
However I'm still trying to bring this theory down to c-code, which will in
fact show different behaviour for different linker options.
As a reminder, I've just filed bug #230460.
I guess it's possible to find out who's the "evildoer" (it might even be done
on purpose) by comparing symbols of libxfont1 with symbols of libs/programs
linking against it, but I'm far from sure there.
Once I know more, I'll post updates to bug #230460.
Cheers,
Stefan.