with -O3 in your cflags and compiling using gcc 3.3, emacs will die with an error about the wrong data type being passed, type listp is expected. switching to -O2 solved this problem. I'd very much like to see a 'export CFLAGS="${CFLAGS//-O?} -O2"' whenever the next release ebuild is written. I spent 2 hours slogging though emacs trying to figure out what was causing the problem, and i expected i was going to need to patch something.... no wonder i couldnt figure it out, it was a cflag thing. *hangs head* I dont even use emacs either. :/ it doesnt seem that gcc 3.2.2 shares this fragility in cflag choosing, but this could just be for my choice of cflags and not have much to do with gcc 3.2.2 being more flag friendly than gcc 3.3.
I forgot I had even filed this bug about two and a half weeks ago, but after weeks of attacking gcc 3.3 bugs and issues I have come to the conclusion that most of them are caused by the hammer update patch. I re-merged gcc without this evil patch and then attempted to compile emacs with -O3. Suprise suprise, the end result was a perfect errorless merge and as such this bug should be closed. This isnt an emacs related bug in any way shape or form. This is a hammer update bug. The fix is simple and solves most of the issues i've seen that arent true legitimate C99 changes or other such issues that arent really bugs. ayanami gcc # diff ./gcc-3.3.ebuild ./gcc-3.3-r1.ebuild 191c191 < epatch ${DISTDIR}/${P}-athlon-hammer-branch-20030515.patch.bz2 --- > use athlon64 && epatch ${DISTDIR}/${P}-athlon-hammer-branch-20030515.patch.bz2
ok thats an interesting find. i've been suspect of the hammer patch for another (not yet in portage) package called LUSH (the lisp universal shell), which also works without the hammer patch applied.
not emacs related
please use a newer version of gcc