Mailing List Archive

If any of you happen to USE=hardened to build gcc, please review Bug
89685 and provide feedback, either on the bug report or to me directly.
I believe that I am seeing a regression from gcc-3.3.5-r1, in that so
far as I know, g++-3.3.5-r1 was good, but g++-3.3.5.20050130-r1 is not.
My evidence is the x11-libs/qt-3.3.4-r3 graphics package. Specifically,
g++-3.3.5...-r1(hardened) builds a version of qmake which always
generates a Bus error; g++-3.3.5...-r1(not hardened) builds a working
qmake, and g++-3.3.5-r1(hardened) also built a working qmake (albeit,
from qt-3.3.4-r2).

I do not know if the problem is 20050130-r1+hardened, 20050130-r1 and
hardening exposes the problem, or g++-3.3.5 and (20050130+hardened)
exposes the problem, or what. I do know that for me, the problem is
easily duplicated across systems, and USE='hardened'/USE='-hardened' is
the discriminator between failure or success in upgrading qt.

Consequently, any feedback, confirmation, or refutation is actively
solicited.

Regards,
Ferris
--
Ferris McCormick (P44646, MI) <fmccor@gentoo.org>
Developer, Gentoo Linux (Sparc)