On Thursday 15 February 2007 09:49, Michael Schwendt wrote: > So, the "upstream" rule here is violated. Instead, you want to change > thousands of spec files because of a theoretical problem that can only > affect users who don't set up a per user build environment on a multi-user > system? Loyalty finds an end here. The mktemp buildroot in spec files > would be really disappointing decision. > > $ rpm -qf /usr/lib/rpm/redhat/macros > redhat-rpm-config-8.0.45-6 > $ rpm --eval %_tmppath > /var/tmp > > could be modified more quickly than applying a bigger buildroot patch > to the code. Sure, that's an option. I think the mktemp based approach found some problems. We haven't ratified anything, we're open for discussion. I was perfectly happy with a rule of "Is there a BuildRoot defined, and it isn't /? Yes? PASS" however others within the Packaging Committee wanted to protect users against the scenarios you described. -- Jesse Keating Release Engineer: Fedora
Attachment:
pgp0d3ET4VetQ.pgp
Description: PGP signature
-- fedora-extras-list mailing list fedora-extras-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-extras-list
