Please do not reply directly to this email. All additional comments should be made in the comments box of this bug.
https://bugzilla.redhat.com/show_bug.cgi?id=731191
--- Comment #5 from Richard Shaw hobbes1069@gmail.com 2011-08-17 09:12:15 EDT --- (In reply to comment #4)
- MUST: build doesn't use $RPM_OPT_FLAGS. maybe use something like:
--- configure.opt 2011-08-17 07:47:12.275486930 -0500 +++ configure 2011-08-17 07:52:24.599256027 -0500 @@ -74,7 +74,7 @@ fi
if [ "$OPT" = 'yes' ]; then
echo 'opt = -O3' >>Makefile
echo 'opt = -O3 $(RPM_OPT_FLAGS)' >>Makefilefi
if [ "$X11" = 'yes' ]; then
I'll get this updated and see if a patch wouldn't be better then a bunch of sed hacks. I usually just sed update "CFLAGS =" to "CFLAGS +="
- MUST. static library build/packaged. Please provide
justification/rationale for doing so, or remove it.
Yeah, I was wondering about that. I only packaged it because it built it. I guess I just need to "rm -f" it so I don't get an "installed but unpackaged" error
- SHOULD. In %files, be explicit about what soname to package, so future abi
bumps don't come as a surprise, use something like %files %{_libdir}/libspnav.so.0* instead?
OK
- SHOULD. Given all the configure/makefile hacks (optflags, DESTDIR, lib64)
in the .spec, I'm wondering if it may be more worthwhile to make an upstreamable patch instead? I can help do that, if that's agreeable with you.
I'll ask but these makefiles are VERY simple and the packages have not been updated recently so I wonder how active upstream is.
I'll post a new spec and SRPM shortly.