$summary example: x86_64-pc-linux-gnu-gcc -march=native -O2 -g0 -std=c99 -Wall -fPIC -fsigned-char -O2 -o tcfmgr tcfmgr.o -Wl,-O1 -Wl,--as-needed -Wl,--hash-style=gnu -L. -L/usr/lib64 -L/tmp/portage/dev-db/tokyocabinet-1.4.46/homedir/lib -L/usr/local/lib -ltokyocabinet -lbz2 -lz -lrt -lpthread -lm -lc the last -O2 overwrites my -O?
Yes, that's not so nice. From configure.in: if test "$enable_devel" = "yes" then MYCFLAGS="-std=c99 -Wall -fPIC -pedantic -fsigned-char -g -O2 -pipe" so that should be patched, and in 1.4.47 that has been rewritten again. I think I'll just try to fix up .47 and file a stablerequest, that should make everyone happy :)
Sanitized in .47, that should be a lot nicer now. Arch teams: Please stabilize 1.4.47
-fomit-frame-pointer is added arbitrarily but it works
(In reply to comment #3) > -fomit-frame-pointer is added arbitrarily but it works Aw crummy, I removed that too. So it should make you happy now ...
amd64 done, thanks Agostino
Stable for HPPA.
Looks good here on x86.
x86 stable. Thanks Andreas
ping! other arch does not want stabilize?
It appends -minline-all-stringops to CFLAGS. That's x86/amd64 specific therefore build fails for other arches, please fix it
+ 13 Aug 2011; Patrick Lauer <patrick@gentoo.org> tokyocabinet-1.4.47.ebuild: + Fixing CFLAGS for #361129 That should make it work on non-x86. Maybe making it respect system CFLAGS is a better fix, I'm open for good ideas.
ppc/ppc64 stable
alpha stable, closing