Ralf Corsepius wrote :
> Makes your example work
> fine. FYI, I never used /var/tmp for _tmppath when I used to build
> on real systems... now with mach and mock, I don't care anymore.
If _tmppath would default to ~/<somedir>, or if redhat was shipping an
rpm with a safe %buildroot, none of these issues wouldn't have popped
up.
Current real world is different: Users set %_topdir in .rpmmacros but
nobody sets %_tmppath.
Educate users! And in this case it's even easier since it's not the
average Joe, but packagers or more advanced users. We've been educating
them to use mock now anyway, right? :-)
> I'd really prefer we try to move forward here, i.e. get some
comments
> on my point n°2 about trying to set a sane (non empty like currently)
> default for BuildRoot, so that we can simply forget about it in a few
> releases.
IMO the answer is clear: We are playing with symptoms. It should be
solved inside of RPM. Unfortunately the responsible parties seem to
prefer to push this issue to users, instead of solving it.
Then let's try and solve it! See my last email, and let us know what
you think (default BuildRoot, %clean and %install first step). I should
have also added a %defattr default, at least for the user/group of
root/root... but let's try one step at a time ;-)
Matthias
--
Clean custom Red Hat Linux rpm packages :
http://freshrpms.net/
Fedora Core release 5.90 (Test) - Linux kernel 2.6.17-1.2431.fc6
Load : 0.39 0.38 0.42