Re: [Devel] [BUG] freetype2 CVS/HEAD: crash in FT_Get_Name_Index (ftobjs.c:2407)

[Melchior FRANZ]

Sorry for the long delay -- I wasn't at my machine for more than two
weeks.


* Werner LEMBERG -- Tuesday 13 April 2004 07:32:
> > The valgrind debugger (http://valgrind.kde.org/) says:
> > 
> > ==25130== Conditional jump or move depends on uninitialised value(s)
> > ==25130==    at 0x3C06BFF2: FT_Get_Name_Index (ftobjs.c:2406)
> > ==25130==    by 0x804B1BD: checkEncoding (mkfontscale.c:1025)
> > ==25130==    by 0x804AA92: doDirectory (mkfontscale.c:912)
> > ==25130==    by 0x8049548: main (mkfontscale.c:264)
> 
> I've just checked FT_FACE_LOOKUP_SERVICE and FT_FACE_FIND_SERVICE --
> both macros properly initialize all variables IMHO.

Valgrind does hardly ever fail. I've accused it myself a couple of
times already, but after searching for a while (which can be *very*
cumbersome) I always found its analysis correct.



> > The funny (and ugly) thing is, that when I add a single line to the
> > macro, the crash doesn't occur any more:
> 
> Maybe a compiler bug?  Please repeat everything with `-g -O0'.  If
> this doesn't help, please convert FT_FACE_LOOKUP_SERVICE and
> FT_FACE_FIND_SERVICE to real functions so that we get better traces.
> valgrind is developing also quite rapidly; maybe a CVS version
> provides better (this is, more meaningful) results.

Yes, it could be a compiler bug. I *am* using both valgrind and freetype2
from cvs/head. I'll try what you suggested soon. I beg for some more days
time. :-)

m.