Austin Acton wrotes:
:: The %post macro is executed *only* when you install the actual
:: i586.rpm. It is not involved in the build process at all.
::
Ok. Thanks. I didn't know this (which just shows how profoundly
ignorant I really am!)
:: > I know how to make sure there are no leftovers. And there are none.
:: > Each build is clean. I have even deleted the the unpacked sources and
:: > the spec file and done rpm -i on the .src.rpm a couple of times
:: > "just in case".
::
:: Then I was wrong, it's not leftover files causing the problem.
::
:: However, considering that nobody else has this problem, and considering
:: that it CAN'T have anything to do with the fact that it's a dual athlon,
:: I can only conclude that you have a non-standard system, and the error
:: is due to having a mix of 9.0 and cooker rpms on the same system.
I agree. And with the cleanup of the "alternatives" links for libgcj-3.2
which I mentioned in my eariler private mail to you, I think this
build is going to work. I just looked in the BUILD/gcc-3.2.1 dir
and I neither see the bogus link nor the root/etc directory
in:
/usr/src/RPM/BUILD/gcc-3.2.1/obj-athlon-mandrake-linux-gnu/gcc/include
But I'll wait to completion of the build before I say any more.
It's currently do the compiler tests.
But I _will_ say that all I did on this machine that was different
than the laptop on which building always has worked was that I built
and installed gcc-3.2-4mdk from cooker in late Nov/early Dec. I'm
guessing that something in _that_ install screwed up the alternatives
links for libgcj and led to my current problems.
So the fact that nobody has seen it only means that very few have
walked the exact path I did. Since I didn't even know about
"alternatives" before two days ago, I doubt I am (directly) to blame :-)
But who knows. Clearly the bug is not in the current package.
Dean
