ARCHITECTURE AND OS: build arch : x86_64 compatible build archs: ia32e x86_64 noarch build os : Linux compatible build os's : linux install arch : ia32e install os : Linux compatible archs : ia32e x86_64 athlon noarch amd64 i686 i586 i486 i386 fat compatible os's : linux RPMRC VALUES: macrofiles : /usr/lib/rpm/macros:/usr/lib/rpm/ia32e-linux/macros:/etc/rpm/macros.*:/etc/rpm/macros:/etc/rpm/ia32e-linux/macros:~/.rpmmacros If I change /etc/rpm/platform to read x86_64-linux rather than ia32e-linux, it will then look in the right place: ARCHITECTURE AND OS: build arch : x86_64 compatible build archs: x86_64 noarch build os : Linux compatible build os's : linux install arch : x86_64 install os : Linux compatible archs : x86_64 amd64 athlon noarch i686 i586 i486 i386 fat compatible os's : linux RPMRC VALUES: macrofiles : /usr/lib/rpm/macros:/usr/lib/rpm/x86_64-linux/macros:/etc/rpm/macros.*:/etc/rpm/macros:/etc/rpm/x86_64-linux/macros:~/.rpmmacros Jeremy thinks that ia32e-linux platform should still be looking in x86_64-linux/macros for stuff. We don't seem to provide an ia32e-linux/macros file.
This seems to be unchanged in current rawhide - do we need this for release? What happens if we fail to fix this?
Will, I may be missing something but it seems to me that this bug is not fixed in recent rawhide. Though I find it a bit odd that my macrofiles setting now references x86_64-linux directories instead of ia32e-linux: ARCHITECTURE AND OS: build arch : x86_64 compatible build archs: ia32e x86_64 noarch build os : Linux compatible build os's : linux install arch : ia32e install os : Linux compatible archs : ia32e x86_64 athlon noarch amd64 i686 i586 i486 i386 fat compatible os's : linux RPMRC VALUES: macrofiles : /usr/lib/rpm/macros:/usr/lib/rpm/x86_64-linux/macros:/us r/lib/rpm/redhat/macros:/etc/rpm/macros.*:/etc/rpm/macros:/etc/rpm/x86_64-linux/ macros:~/.rpmmacros I have not altered /etc/rpm/platform from its default and the file still contains a single line, reading "ia32e-redhat-linux".
Same on my x86_64 machine(s) - I'm hoping Paul can help shed some light on whether this needs fixing now or later. Paul?
Oh wait a second, it seems that the bug was about the ia32e-linux platform not looking in x86_64-linux/macros; but that is apparently fixed (?). Will, this does indeed seem to be fixed to the contrary of my earlier comment #2; though I can't pin down exactly which E:V-R it was fixed in (if indeed it was). Nothing in the %changelog seems to show a patch or fix applied. :|
AFAIK this is not going to horribly break anyone's system, so we'll fix it with an update soon after F7 is released.
The problem is fixed in the rpm-4.4.2-ia32e-install.patch which (appears) to be first included in (these are cvs tags) rpm-4_4_2-27 (revision: 1.1)
It does break the occasional spec, making it difficult for people to build packages on x86 hosts. For instance: http://marc.info/?l=courier-users&m=117682266604494&w=2
User pnasrat's account has been closed
Reassigning to owner after bugzilla made a mess, sorry about the noise...
Fixed in rpm-4.4.2.2(-rc1) in next rawhide push.