I've tested this on multiple computers(and heard alot of complaints), it seems -O3 is causing BitchX to seg fault right after you run it, i suggest making the ebuild force -O1 or -O2
will do -- but this is hardly a blocker, in that it does not stop Gentoo Linux from getting installed. anyway, what about compiler flags like -fomit-frame-pointer, etc?
Right. Compiling with -O3 works fine, but the program segfaults. Falling back to -O2 does the trick. Here are my CFLAGS/CXXFLAGS: -march=athlon-tbird -O2 -pipe -m3dnow -mmmx -fforce-addr -fomit-frame-pointer -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -falign-functions=4 -maccumulate-outgoing-args Yeah i know that some of them are redundant. I'll force bitchX to use -O2. Could you do me a favor and try to merge epic4 with -O3? It has the same problem as BX over here (well BX is epic4 based iirc). That way we could fix two applications with one bug. :) -phoen][x-
Ok, i just compiled it on a AMDK6-3 and a Celeron 700 with -O3 On the k6-3 it didn't seg fault but on the Celeron it segfaulted just like with BX (Sorry about the 'blocker'=)
using just: CHOST="i686-pc-linux-gnu" CFLAGS="-mcpu=i686 -O9 -pipe" CXXFLAGS="-mcpu=i686 -O9 -pipe" results in a segfault too (p4 1.8gig)
compiled it with CHOST="i586-pc-linux-gnu" CFLAGS="-march=i586 -O3 -pipe" CXXFLAGS="-march=i586 -O3 -pipe" works fine p166 MMX watch out #6273 there's a ebuild maybe try it
*** Bug 6273 has been marked as a duplicate of this bug. ***
[standard_answer] Sorry for the delay [/standard_answer] I'll force -O2 in the ebuild. From what i've read/seen, this should fix this bug. If the problem still exists, we shall fall back to -O1. Anyways, lets start with small steps. -phoen][x-
Well seems like bcowan fixed the bug in the meantime. He used flag-o-matic to replace -O3 with -O2. I modified it to also take care of -O4++; -O[3-9] will be replaced with -O2, now. Works fine for me. Any issues, reopen this bug. -phoen][x-