Austin Acton wrote:
 :: Seems hard to believe that error.

Yeah. I know.

 :: If you're just doing
 :: rpm -ba
 :: it won't make ANY difference what processor you're using.

Unless there's some logic in the .spec file that detects processor
and branches differently. The .spec file for MPlayer certainly does this.

 :: It will build
 :: for i586 (unless you have altered the .rpmrc file).  In fact, from your
 :: post, it looks like you were building i586-mandrake-linux-gnu.

(see below)

 ::  So the
 :: error shouldn't be caused by building on an athlon.  And if it works on
 :: you other system, it makes me think that your dual athlon has a dirty
 :: build environment or something.

(see differences between the two systems in my "other message" dicussed below)

 :: However, I have a dual athlon and I will try to reproduce the problem
 :: right now.
 :: 
 :: Austin


Austin, be sure to read the (long, involved) message I just
wrote to Todd L. on cooker about this.

In fact (as I said in that message) the bug was originally tickled
when I built for the athlon-mp (modifed /etc/rpmrc file) but
I deleted that file on a second attempt to remove one variable
and the build still bombed out.

See that message for the two other differences between the
"good" and the "bad" system.

Now that I "fixed" the problem (see message on how) I will build again
tonight to see if the build goes through.

Problem is that I know nothing at all about "update-alternatives"
which seems to be the fundamental cause of the problem.
For some reason it is creating the error-causing link on one machine
and not in the other.

Dean, starting yet another build of gcc-3.2.1

Reply via email to