Good news guys. If my testing is correct, this should be fixed by .
I would be glad if you could help me to investigate, how to make DNF
to prefer MRI over JRuby (use Recommends, Suggests or epoch for that,
probably break some bad dependency - there used to be issue with
rubygem-json). That is the right solution for the future. And along
the way, JRuby is doing steady progress to work correctly.
So to say, the fix ATM is easy as BR: ruby (btw this is default
behavior of gem2rpm, so you probably removed this line, because it
worked for YUM by default, but does not work for DNF) and I don't
think that removing ruby(release) is the right step. You would need to
modify all your packages anyway.
Dne 24.6.2015 v 06:37 Dan Allen napsal(a):
> On Tue, Jun 23, 2015 at 9:36 PM, Ken Dreyer <ktdreyer(a)ktdreyer.com
> <mailto:firstname.lastname@example.org>> wrote:
> I think it makes sense to
> just remove the virtual provide from jruby until jruby is more
> I bumped into this same issue. I did find that one solution is to
> first install ruby, then JRuby won't get pulled in as a transitive.
> For some reasons, it's winning over MRI Ruby in the dependency closure.
> To clarify, though, it's not JRuby that's unstable. JRuby is
> incredibly stable. It's the package for Fedora that's unstable (or
> renders it unstable).
> Dan Allen | http://google.com/profiles/dan.j.allen
> ruby-sig mailing list
ruby-sig mailing list