On Tue, 03 May 2011 14:47:21 +0200, Axel Dörfler <axeld@xxxxxxxxxxxxxxxx> wrote:
Brecht Machiels<brecht@xxxxxxxxxxx> wrote:I've seen several independent benchmarks (they should be easy to find), and git is indeed the fastest and it's repositories are the smallest.It doesn't really matter which tool is the fastest at a specific action, but if the tool's performance is acceptable. While I guess that git will be faster overall compared to hg, I don't expect it would be a performance issue to use either tool for our repository. And while I would usually prefer hg over git, I think the missing hard link feature of BFS simply disqualifies hg for now. And why not use git then, if it comes down to preference, anyway?
It may seem that I'm pushing git, even though I'm not exactly a very active contributor. I don't really care whether Git or Hg ends up being chosen. But I don't see the reason for stretching this decision so long. Performance-wise, there's no reason for choosing one over the other. But does a vote make sense as long as hard links are not supported? It's obvious git is the only option, really.
As soon as the choice is made between Git or Hg, a decision can be made on using the GitHub service or similar and get this over with.
Cheers -- Brecht