Re: optflags problem

["Matthew R. MacIntyre" <matt pipfield ca>]

> Yup, CLAGS in the environment is honored.

Ahh, excellent.  That should make life a little easier.  I haven't seen this
documented anywhere though ; might this be a worthy addition to the manual
pages?

> Basically setting up "reproducible builds" is far more complicated then
> just chossing optflags.

Good point.  Related to what I'm trying to do.  What other factors need to be
considered in order to create a reproducible build?

> You still have not described what your build problem is, or even identified
> what package or arch is involved. Without that info, I can't begin to tell you
> what to do.

Well, basically, I'm trying to roll my own distribution.  I'm taking the
redhat distro, merging in the updates, and then rebuilding everything.
I'm building for athlon, i686, and i386 to start, and I'm basically trying
to rebuild everything with -O3 -march=i386 -fno-exceptions -funroll-loops
for CFLAGS.  There have been a few instances where I've had to sed out some
of those flags in a spec file, but it has generally worked ok (from a package
building perspective anyway).

I've made folders under /usr/lib/rpm for i386-pipfield-linux,
i686-pipfield-linux, athlon-pipfield-linux, and noarch-pipfield-linux, with
the macros files in there.  This whole thread started when I couldn't get rpm
to recognize the optflags I had set in $arch-pipfield-linux/macros, and was
trying to break it down a little more, but couldn't figure out the problem.

When I'm building a package for i686-pipfield-linux, I want it to
use the CFLAGS above instead of the ones supplied by default without having to
do much extra work every time.  Is this possible, or should I go back to using
custom rpmrc files with the optflags overriding the defaults?

Thanks for your help so far with this, Jeff.  I appreciate it.

-matt