As one of the downstream distro packagers, I follow both the tarball and rpm
revisions carefully. Please be aware that changeslike the one proposed impact
us and ought not be made without some announcement so we can know what is going
on and adjustour packaging code accordingly. Right now I'm using the code
below for our rpm's which is quite similar to the code in slurm.spec.
Occasionally I see a -2 revision, but only rarely. Since most users are
probably not using rpm build tools at all, it might not makesense to expose
that number always.# Upstream tarballs use an additional release number
%global ups_rel 1
%if "%{ups_rel}" == "1"
%global name_version %{name}-%{version}
%else
%global name_version %{name}-%{version}-%{ups_rel}
%endif
Source0: http://www.schedmd.com/download/latest/%{name_version}.tar.bz2
On Monday, June 10, 2019, 12:42:45 AM EDT, Kevin Buckley
<kevin.buck...@pawsey.org.au> wrote:
This actually just tripped me up on a Cray, but I belive the observation
is still worthy of discussion.
If I take the
slurm-19.05.0.tar.bz2
tarball from the SchedMD download site, and then do a direct RPM
build on it, so
rpmbuild -ta slurm-19.05.0.tar.bz2
what I end up generating are the following RPMs
slurm-19.05.0-1.x86_64.rpm
slurm-contribs-19.05.0-1.x86_64.rpm
slurm-devel-19.05.0-1.x86_64.rpm
slurm-example-configs-19.05.0-1.x86_64.rpm
slurm-libpmi-19.05.0-1.x86_64.rpm
slurm-openlava-19.05.0-1.x86_64.rpm
slurm-pam_slurm-19.05.0-1.x86_64.rpm
slurm-perlapi-19.05.0-1.x86_64.rpm
slurm-slurmctld-19.05.0-1.x86_64.rpm
slurm-slurmd-19.05.0-1.x86_64.rpm
slurm-slurmdbd-19.05.0-1.x86_64.rpm
slurm-torque-19.05.0-1.x86_64.rpm
where I note the extra "-1" that's crept in to the RPM names.
Going back through the tarball history (yes, I've kept a lot
of then) suggests that, were there to be a "-2" release of
slurm 19.05.0, then the tarball name would be
slurm-19.05.0-2.tar.bz2
and that it would then match the name of the generated RPMs.
Is there anything, other than a tacit admission that there may
sometimes need to be a "-2" or greater revision of any given
release, preventing SchedMD from distributng the first tarball
in a release with the "-1" revision in the tarball name ?
Kevin
--
Supercomputing Systems Administrator
Pawsey Supercomputing Centre
Tel: +61 8 6436 8902
SMS: +61 4 9970 3915
Eml: kevin.buck...@pawsey.org.au
