On Fri, Sep 9, 2016 at 1:49 PM, Michael Schwendt <mschwendt(a)gmail.com> wrote:
> 16:31:18 <ignatenkobrain>
hyperv-daemons-license-0-0.15.20160728git.fc26.noarch
Only looking at that, what's wrong with it? It seems to adhere to Fedora's
pre-release snapshot scheme with no known %version.
> 16:31:18 <ignatenkobrain> kernel-modules-4.8.0-0.rc4.git3.1.fc26.x86_64
> 16:31:18 <ignatenkobrain> arpack-3.3.0-2.b0f7a60git.fc24.x86_64
> 16:31:18 <ignatenkobrain>
python3-iscsi-initiator-utils-6.2.0.873-34.git4c1f2d9.fc25.x86_64
> 16:31:46 <ignatenkobrain> it's jsut 4 packages out of thousands which are
packaged wrongly because packagers do not understand what they should do
More often, it is ignorance. Packagers, who "do not understand what they
should do" as you put it, ought to ask about it. Preferably, in advance.
However, I doubt the packagers are aware of the RPM Version Comparison
pitfalls.
As part of allowing tilde I described in same page how RPM Version
Comparison works.
Even the cases above don't result in update problems, if the next update
bumps %version or %release in a good way. At most there would be
inaccurate versioned dependencies, which either are superfluous or not
happy about the changes introduced by the snapshots.
A few times per year you can experience upgrade problems where packagers
didn't even notice that their latest builds don't replace the previously
released builds. And once the issue is discovered, don't be surprised if
Epoch gets bumped without any concerns.
It would be a better idea to run a bot and check koji builds for problematic
EVRs (and versioned package inter-dependencies) early.
I do it from time to time
semi-manually ;)
--
-Igor Gnatenko