xemacs-21.4.15-r2 ebuild segfaults on amd64

[glassbear]

The ebuild for app-editors/xemacs-21.4.15-r2 has been marked as stable, but segfaults during compilation on my amd64 system.

The ebuild terminates part way through with the following message:

[glassbear]

filed bug 63358

[Tsonn]

It worked on my system; the only obvious difference I can see is CFLAGS. Mine are:

[glassbear]

Changing the CFLAGS did not change the results. Compilation always fails in the file update-elc.el, with the same line # and error message quoted in my original posting.

I tried changing the CFLAGS one at a time to see whether I could isolate the effect on compilation of any one of them. Since compilation never succeeded, I can't point to any individual flag as the culprit, but just for the sake of completeness, here are the different combinations I tried:

• -O2 -pipe -fno-stack-protector (original settings)
  
• -O2 -march=k8 -pipe -fno-stack-protector
  
• -O2 -pipe -fno-stack-protector -fomit-frame-pointer
  
• -O2 -march=k8 -pipe -fno-stack-protector -fomit-frame-pointer
  
• -O2 -march=k8 -pipe -fomit-frame-pointer
  

G

[trantorvega]

I had the same kind of problem during the compilation. Since i had a previous verion of xemacs installed (actually the same version, just -r2 instead of -r3), i thought that the problem might have had something to do with gcc or glibc, since i had previously recompiled them with the hardened flag on. Once recompiled both without the hardened flag (or maybe just one of them, i'm not sure), xemacs merged perfectly. Perhaps the same thing happened to you.

[apryan]

Confirmed that this problem still exists with xemacs-21.4.20-r5. I have tried emerging with -berkdb and -postgres. I have not tried the one other suggestion of reemerging gcc and glibc with -harden as I do not want to screw up gcc. If this problem can be resolved with out going that way, let me know.

thanks