Op vrijdag 03 december 2010 10:21:38 schreef andre999: > Maarten Vanraes a écrit : > > Op donderdag 02 december 2010 23:09:14 schreef Morgan Leijström: > >> ADDING to what i just wrote: > >>> A week or so old news but I saw it now... > >>> > >>> <http://wiki.mandriva.com/en/2011.0_specs_proposal#Swith_to_rpm5> > >>> <http://ostatic.com/blog/whats-coming-in-mandriva-2011> > >>> <http://proyvind.net/wordpress/?p=71> > >>> > >>> Citing proyvind from last link: > >>> "...I think we will end up being able to be the awesome alternative in > >>> the forked world, hoping for interest of other distros, actively > >>> inviting their participation and offering to help merge,..." > >> > >> - i forgot to paste the best part: > >> > >> "I’m curious about what stance Mageia will take on this though, I’m > >> certainly hoping they will adopt our work and allow for us to help them > >> and contribute back and forth, if not I fear they’ll miss out on several > >> changes in just perl-URPM done recently during rpm5 rewrite, and even > >> more so those planned when rpm5 upgrade situtation settles.." > >> > >>> There are also other goodies in the middle link i hope we can coopreate > >>> with them on. > > > > well, i'm not adverse to cooperation, but RPM5 is in my opinion FAR too > > unstable to be used for use on our first release. > > > > I would say, let mandriva find the bugs first and let it give rpm5 a > > testrun with their users; by the time we hit our second release, we will > > have a for sure stable version of it. > > I've been following RPM5 (and the associated urpm4) on cooker for a > while now ... > All 3 stories come from the person integrating rpm5 into Mandriva - he > was recently hired to do just that, with full support of management. > To me it looks like rpm5 itself won't be the big problem - rather > changing the urpmi tools and adapting the build system to work with it. > Not sure which way Mageia should go at first, but the newer versions of > their tools will be based on rpm5. > It was interesting how cooker wasn't warned of the changes - suddenly > many developers were left unable to do any anything, until Per fixed the > bugs he created. And then they learned about the planned switch. > A perfect example of how NOT to do things :)
indeed, noone said anything, but that is how i felt. The thing is, that cooker testers are a bunch of people who do things in more or less the same way; i fear regular users might do stuff another way. imo, i think/fear Mandriva will (after they release) have some RPM5 issues that regular users will be impacted of. I would take the sure road, and keep our working tools at level before the upgrade; and follow up the stuff that happens. imo, if we don't use rpm5 right now, we lose nothing; i'm sure it'll be an improvement; however, is it necessary at this moment? i think not. imo the risk outweighs the benefits at this time. especially since we're short on time. > It is good that Per is open to cooperation - his experiences will be a > big asset when we do follow suit - which I expect we will, sooner or later. > Maarten, have you taken a close look at it ? I'm had discussions with Per around the time of the forking. however, i (and others) feel that we need to structure ourselves first before going down that route. I think after the release (and org structure will be stable) we could very well start the talks, imho. thing to keep in mind is that Mandriva is backed by a company and we by our non-profit organisation, the last thing i would want to do is taint the organisation, so we need to be careful there.
