Fernando Nasser wrote:
> Not all under our control, not all authors still around etc. And these
> fails are subtle and sometimes only detected too late. The truth is that
> the regexp syntax is almost impossible (actually, I believe it is
> impossible) if the prefix is not there.
> There isn't enough information to know if it is the same package or not.
> People are forced to add explicit had-coded code for each exception.
They shouldn't hardcode it, they should query the actual package or
metadata.
> It is always confusing for humans to track where things are coming from if
> subpackages have different names. If one is not in a situation to run a
> rpm -qi on the subpackage .rpm one may never know what produced it.
It's also in the XML/SQLite repository metadata (repomd). Try e.g.:
repoquery -s oxygen-cursor-themes
> The only reason I don't propose %{name} should prefix all subpackages as a
> guideline is that we'd have to fix our geronimo-specs huge spec file ;-) I
> actually think I should regardless...
I think there are many valid reasons to use %package -n. RPM supports it for
a reason.
Kevin Kofler
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
