-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 22 Jun 2011, at 20:04, Jesus Cea wrote:
> On 22/06/11 16:46, Andrzej Szeszo wrote: >> Letting developers to choose their vcs frontend is one nice thing. Git's >> lightweight branches and the way they work are also very appealing. >> Also, more people are familiar with Git than with Mercurial so new >> developers may find it easier to get started with the project if its >> source is managed using Git. > > I think that using several different DVCS would be a source of > "impedance mismatches". > > Do you have hard data to support your comment that GIT is more used that > HG?. Moreover, in our arena? (OS development, distro creation). I don't think Andrzej is suggesting long-term co-existence of git and Mercurial but a transition that starts with using git for new development, with existing consolidations migrated subsequent to that. When he says let developers choose, I think what he's saying is that we should choose for ourselves rather than letting the upstream choose for us: being able to clone from Mercurial to git and vice-versa is something that just works to the point that what the upstream does needn't be decisive (although, as Garrett's pointed out, we have to consider how the SCM is integrated into the build tools). I don't think user populations are a dispositive argument here, as I don't think the difference raises to a level where it implicates viability (i.e. I don't think there's a question of Mercurial dying on the vine for lack of users). If Mercurial really is adequate for our use cases, I see no reason to ditch it because more people are using git, and I think that's not the crux of what's motivating git advocacy. If our use case is clearly understood in its design and well-documented in its use, I really don't give a flip whether someone else using a different product for a different use case has an order of magnitude more users--I wouldn't be working on OI/Illumos if I conceded to such logic, either as in the affirmative or as a simple reversal. That's how I ultimately digest the question of how easy it is to get started with the project: most people are just going to be cloning source quite a bit before they have to worry about merging or committing, so we have plenty of opportunity to let people absorb non-trivial use cases before it resembles a barrier to entry for new developers. This is largely, if not entirely, about whether git support for branching is something that Mercurial isn't delivering something truly essential for OI and can't. Cheers, Bayard -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.12 (Darwin) iQIcBAEBAgAGBQJOAmIoAAoJEHm5cBpJ87dowwQQAL83YfzeuZ6mVToMS+KTIeXr wTZ5fyLZvoAosZO+Zst3kpVW91ufvlmvgFMRcdQeiV/me21u7gzcDTTlZ/Bieucg RTOLRCmcjGepdv+slBxIYe0tUoY51LyaN4GUGaTpodoQilARwVYPebWY/B1RcK7s leqKrWNQpGJkYxgKtKOvaQmIHcBpur6otC25TLwVWcYYWiNUv8WVl5eKW7+maOvi qk931fHy9/5G/y4v6ZJFc0zfrChEOxd3nDW3jY5QimdpoILHd6NH5Ru5ed1mVvih 7i4cFxe7/43/9BgEcyXtp2CXnup3wOboXtNU8SHGGXk9qHXtV0a3pniB2cevU3Ev oBMZCr1XhbdBPmIbVsvK/ILEktGxqpuxJQ5kCuEougBGsFWQiri1m0YmkTW0acUB fsSue0hSb/iFdmcFQ5+ixlH+gcdop4uW2tWW5VWAph5jAoaIby6ESwfpI1KadPpo LUAlovwxQMw6eOmwH7wgLyKLCV3BrAetxn+D9+gbZ/fTztDf7MW2PZYOcpWUb/q1 2SMS7rjYAPm/9AznoYi2qmkQ+rpDI6th/st36xhXDCufRzZ7tIR3zbyaIHgc3znL G240MkTzlmra9Wtk6mG4pjU8OLTma7gXl5FLk5KmRJfaX6ydi4CNhqUiHi+SEjUk u0Tb8kXQLuUz6KqWy5zb =vQGR -----END PGP SIGNATURE----- _______________________________________________ oi-dev mailing list oi-dev@openindiana.org http://openindiana.org/mailman/listinfo/oi-dev
