On 31.03.2009, at 13:17, Göran Krampe wrote:
Hi!
Andreas Raab wrote:
Folks - Just as a side note to anyone who is interested in distributed version control systems, I thought it's interesting to see that they get more and more widely adopted. This just popped up on Python-dev: http://mail.python.org/pipermail/python-dev/2009-March/087931.html I'm wondering if this has any impact on our thinking with respect to Monticello, change sets etc?
I am interested in these beasts, and Keith also has quite a bit of experience. We use Hg in Gjallar - although very little. I have also used bzr a bit, no git yet - but it is just a matter of time since Github has really brought git "to the masses". Those are the three "top contenders" although Darcs has a special place in my heart, not for it's magical smarts (well, its cherry picking is way cool of course) but mainly for its tremendous easy of use.
As you may know OpenSolaris, OpenJDK and Mozilla all use Hg (IIRC) and Ubuntu pushes Bzr. These three are all very good, battle tested and fast. Bzr and Hg being written in Python, and git in C.
The Quilt/patch queues in Hg etc is an example of inspiration for Deltas, when Deltas are fully operational such queues can easily be made using them.
IMHO there is a lot to be learned from these tools. Git and Darcs for example represent two different "philosophies" - git following the "don't be too smart, better to be slightly dumb but predictable"- idea and Darcs following the "very, very smart doing magical stuff"- idea. :)
With Deltas I was also inspired by Linus/git - the idea to be able to do relatively nice merging without using history like MC does - thus, "not too smart, but predictable".
I've been working with git for a while now, first in Sugar and lately at VPRI too. I like it very much (compared to svn at least, no experience with hg or bzr), and the best thing is that it is really really simple inside. Highly recommended reading:
http://www.newartisans.com/2008/04/git-from-the-bottom-up.html
- Bert -