12.09.2012 22:13, [email protected] kirjoitti: > Log Message > > disable 'fortify' to prevent crashes (mga#6723) [... > Modified: cauldron/wine/current/SPECS/wine.spec > =================================================================== > --- cauldron/wine/current/SPECS/wine.spec 2012-09-12 19:02:56 UTC (rev > 293157) > +++ cauldron/wine/current/SPECS/wine.spec 2012-09-12 19:13:38 UTC (rev > 293158) > @@ -25,7 +25,7 @@ > Name: wine > Version: 1.5.12 > %define pre 0 > -%define rel 1 > +%define rel 2 > %if %pre > Release: %mkrel 0.%pre.%rel > %define o_ver %version-%pre > @@ -245,6 +245,10 @@ > > %build > autoreconf > +# disable fortify as it breaks wine > +# http://bugs.winehq.org/show_bug.cgi?id=24606#c22 > +# http://bugs.winehq.org/show_bug.cgi?id=25073 > +export CFLAGS="`echo $RPM_OPT_FLAGS | sed -e 's/-Wp,-D_FORTIFY_SOURCE=2//'` > -Wno-error" > %configure2_5x --with-x \ > --with-pulse \ > %ifarch x86_64
Strange, wine configure script already disables fortify since 1.3.6 (Oct 2010), so something weird must be going on: http://source.winehq.org/git/wine.git/?a=commitdiff;h=105793bff75f6a869bd9ccfb390b2f88fa861235 ... ok, it is actually because we use -Wp,-D_FORTIFY_SOURCE=2 instead of -D_FORTIFY_SOURCE=2 and -U_FORTIFY_SOURCE=2 fails to override the former. So I sent a patch to wine-patches@, no idea if they agree: http://www.winehq.org/pipermail/wine-patches/2012-September/118257.html Also, a cleaner way to disable fortify is %define _fortify_cflags %nil -- Anssi Hannula
