Thorsten Leemhuis wrote:
Michael Schwendt schrieb:
> fedora-usermgmt is not about fixing something, but about adding a
> feature. Well, that's my point of view. I'm not a hardcore advocate of
> using it everywhere. But I don't understand why a simple EPEL steering
> decision is wrapped into a crusade against an optional tool.
Mainly for two reasons afaics:
- Because the rules in EPEL imho should be as identical to the rules
from Fedora as much as possible, as everything that differs between the
two will make life harder for users and packagers (for example a package
that currently uses fedora-usermgmt in Fedora could not simply be build
for EPEL without adjustments)
- Because having a tool like fedora-usermgmt that solves a particular
problem is IMHO not worth much, if half of the Fedora packages use it,
while the other half doesn't
So having a solution for Fedora (use it everywhere, don't use it at all,
use another solution that fixes the problems fedora-usermgmt tries to
solve) would be the best for everyone.
I vaguely recall seeing a spec someone wrote that had some conditionals
added to determine if it should use fedora-usermgmt or simply run
useradd. Extra overhead, but it'd let the same spec be viable for both
fedora and epel w/o having to either bring fedora-usermgmt to epel or
kill fedora-usermgmt altogether (though I'm in favor of the latter. ;)
--
Jarod Wilson
jwilson(a)redhat.com