On Thu, May 28, 2015 at 11:30 AM, Daniel P. Berrange <berrange@xxxxxxxxxx> wrote: > On Thu, May 28, 2015 at 12:16:19PM +0200, Christophe Fergeau wrote: >> On Wed, May 27, 2015 at 04:15:57PM +0100, Zeeshan Ali (Khattak) wrote: >> > Having said that, I don't think this is so much better than the the >> > solution we already merged that it'd be worth delaying releases for >> > (which were due day before yesterday). >> >> For what it's worth, libosinfo is not following GNOME release cycle. >> The fact that GNOME Boxes maintainers decided to depend on libosinfo.git >> is in a way Boxes problem. Making a libosinfo release to accommodate >> Boxes makes sense, though it I'd prefer if this was anticipated and we >> started preparing for a libosinfo release the week before for example. >> >> Rushing some last minute patches in, and insisting that the patches have >> to go in before a release is pushing things a bit too far in my opinion. > > I think we're probably getting to the point where we should consider > splitting the metadata off from the code. There's rarely a compelling > reason to rush out new releases of the code, but the metadata we could > practically be releasing on a weekly basis given the number of distros > we track. Since new distros are generally additions only, there is > less risk of regresssions to existing distros, when pushing out fast > releases of the metadata whenever needed. Good idea and I've been meaning to making that happen for a while but in this case, we couldn't have just done that as making Fedora 21-22 install scripts working, involved adding new API and getting apps to port to that API for it to work. -- Regards, Zeeshan Ali (Khattak) ________________________________________ Befriend GNOME: http://www.gnome.org/friends/ _______________________________________________ Libosinfo mailing list Libosinfo@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libosinfo
