On 10/15/07, Mark Martin <storycrafter at gmail.com> wrote: > > On 10/15/07, Brian D. Horn <Brian.Horn at sun.com> wrote: > > > > > > Although I haven't looked at the details here, I note the fact that > > there is a -U__i386 which is probably needed > > due to the fact that under some circumstances __i386 is being predefined > > by the compiler, but you may > > need -U__sparc (or something similar). > > > > Or worse, a C standards issue with gcc-4.0.0... > > http://gcc.gnu.org/ml/gcc/2005-02/msg00053.html > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19333 > > using -save-temps and inspecting the .i file reveals it is the case > described in the email ref and bug..
So I've setup a local test harness which confirms that this is an issue for gcc 4.0x and later builds. It doesn't appear to be an issue for gcc 3.4.3and SUNWspro CC, of course. Digging further confirms this: http://www.opensolaris.org/os/project/gccfss-on/ Looking at one of the webrev fixes shows a fix for sys/exec.h (which I was buttoning up the courage to request a change to anyways) http://cr.grommit.com/~alex/webrev/usr/src/uts/common/sys/exec.h.cdiff.html Which moves the empty type array decl to later (after the struct definition). Rest of webrev: http://cr.grommit.com/~alex/webrev/ Where does this leave me? A diff on gcc from 4.0.0 to latest shows no improvement on the regression: http://gcc.gnu.org/viewcvs/trunk/gcc/c-decl.c?r1=94505&r2=128642&diff_format=h So later builds of gcc 4 aren't likely to be helpful. Looks like my only options are to either backtrack to gcc 3.4.3 (and hope that the regression "fix" wasn't really a mask for segfaulting code as suggested in the gcc bug report), or look at the webrev fixes and incorporate. I'm leaning towards backtracking to 3.4.3. Mark -- ------------------------------------------------------ Born to the false world, the wanderer, Storyteller, The Pied Piper On a quest for immortality Gathering a troop to find the fantasy -- Nightwish -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/powerpc-discuss/attachments/20071015/93891364/attachment.html>
