On Tue, May 25, 2010 at 08:33:59PM +0200, Iustin Pop wrote: > On Tue, May 25, 2010 at 08:05:11PM +0200, sean finney wrote: > > i would hazard a guess that the problems are still there but that you > > don't see them because the tests are passing and thus the package builds. > > istr that running with -O0 -ftrapv still produces an abort(), which would > > definitively confirm that the overflow bugs are still there. > > I have just run a test on amd64 with -O0 -ftrapv and the code indeed aborts.
oh, that's interesting... when i tried with -ftrapv way back when i
didn't see this on my amd64 box, but maybe that is something that
on amd64 gets optimized out (don't think i did both -O0 and -ftrapv)
or otherwise not detected due to register voodoo. i guess that should
make debugging a bit easier then...
fwiw, i found the easiest way to debug is to set a break on abort() before
running the app in the debugger (i think i also had to work around libtool
and pass a couple extra things via LD_LIBRARY_PATH). then it became a bit
easier to do the hack/run/debug cycle resulting in the previous patch.
sean
--
signature.asc
Description: Digital signature
