Disttag for Fedora 7 and beyond
Michael Schwendt
bugs.michael at gmx.net
Fri Jan 5 19:05:24 UTC 2007
On Fri, 5 Jan 2007 19:09:00 +0100, Axel Thimm wrote:
> > > How many packagers have been bitten by Requires: foo >= 2.0.1 while
> > > they really needed foo >= 17:2.0.1
> >
> > Users, not packagers.
>
> That's the same, or if one get's bitten he bites the next in the food
> chain ;)
The packagers made the packages, becauses the strict dependency works
for them. But it's not transparent for the users.
> > For everyone else including packagers, in the distribution there is
> > only one package "foo" [...]
>
> Nah, that's an arghument against having versioned dependencies at all,
> which is not what we want.
It's an argument against superfluous and desolate versioned deps, yes.
> > > and the next update needed at least 23:3.0. E.g. the epoch
> > > inflation everywhere make it mandatory to start checking all your
> > > versioned BRs and
> >
> > Versioned BRs are not affected, since the RPM Epoch never specifies an
> > API version.
>
> What makes you say this? How about epoch of the perl package itself?
It is specific to the RPM package, not defined by Perl at all.
> They very much define the ABI/API in this case by themselves.
There is no Epoch in Perl's versioning scheme. Not in the old one,
and not in the new one either.
> Furthermore the automated perl dependencies that by definition lack an
> epoch, can become an indirect BR quite easily ...
>
> Ask the people that had to cope with the epochs in perl itself, that
> required special handling in all perl dependency detection and that
> creates redundant parts of redhat-rpm-config that we cannot send back
> upstream, because any epochs are non-portable. So we get to sit on all
> the special handling stuff and this just for automatic depedencies for
> perl the package.
Why do you want to add Epochs to versioned Perl dependencies?
> > Epochs are not evil. Even the old explicit "Epoch: 0" policy solved
> > more problems [of an epoch promotion bug in rpm] than it created,
>
> Ouch, no, let me respectfully disagree 100%. That broke apt, the only
> depsolver back in these days, and until we knew what it was it created
> an epoch inflation never seen again since. And it did solve nothing,
> because the problem was in comparing epoch "none" vs epoch "0",
c.f. http://www.wideopen.com/archives/rpm-list/2003-June/msg00137.html
There are more gems to find.
> and guess what: There were no packages with explict epoch "0" back then.
It broke comparing "no Epoch" with "Epoch!=0". Sub-packages that were
moved into a new main package with an own versioning scheme around that
time were affected, for instance.
More information about the Fedora-maintainers
mailing list