Le 10/05/2011 01:34, Matthew Brush a écrit :
On 05/09/11 11:12, Colomban Wendling wrote:
Le 09/05/2011 19:35, Jiří Techet a écrit :
I'd say that VCS migration and bug tracking system migration should be done separately and independently. Migration of the bug tracker is a lot of work while migration to git is quite easy. I'd also be rather cautious before moving the bug tracker to GitHub. At the moment they are offering hosting of open source projects for free but there's no guarantee it will be like that in the future as well. This is no problem with the git repository if they get evil - you can always upload the repository somewhere else and update a few links on the geany's home page. However with the bug tracker it would be a much more painful process.
Well... this makes sense, but having the but tracker on SF and the code on GitHub seems a bit like a suboptimal option -- though since SF don't really link bug tracker and VCS maybe it'd not really change anything.
From what I can tell, the majority of the bugs in the SF tracker are either closed, open but will never get resolved or no longer apply to current versions, so I don't know how much of a big deal it would be to start moving away from it, of course always leaving it (possibly read-only?) for reference.
Maybe, need to check but might not be that painful (BTW, don't GitHub offers a SF BT import feature? :D)
But the point on the possible future of GitHub is important IMO. if we have no guarantee for the long-term viability -- and when I read you I read "I'd not be really surprised if it happened" --, do we really want to use this? I mean, if we need to switch to another official repo next year because GitHub decided not to continue to provide (free) hosting for us, it'd not be really good.
Speculating on the future of any of the project hosting sites is just that, speculation. They have different business models, like SF with ad revenue, GitHub with private paid accounts, Gitorious with extra services (and probably $ from Nokia), and Google Projects with Google's plan for total world domination.
If I had to make a guess, I'd say it would be more likely for SF to go belly up due to lousy services, mass exodus to better project sites and it not being financially worthwhile for GeekNet.
Put simply, AFAIK, none of these projects sites offer a guarantee that they will not shutdown, go paid only, or otherwise change their services, so I don't think speculation should be a primary factor in deciding on a project site.
Agreed as said in another mail, apart that I doubt SF will really die, just maybe become even more crappy by the years.
I haven't either checked the other sites (Gitorious, ?) deeper, maybe they are good candidates if we don't want the BT functionality? don't know -- apart that I already have and account on Gitorious and wasn't scared by their policy.
I can't say I'm personally opposed to Gitorious, but to me it just seems like a stripped-down version of GitHub, missing lots of the cool features. Of all the project hosting sites I've used though, the only two I really dislike are SourceForge and Launchpad followed farther by Google Projects.
Well, again, I have no real opinion on this, apart that yeah, GitHub *seems* (haven't tested it) to have more cool features. I was suggesting something else only because of the speculations about GitHub's future ;)
Anyway, I think we should wait for Nick's opinion, and probably again Enrico and Frank ones about the BT stuff.
Cheers, Colomban