Linux workflow during the « bitkeeper debacle »

At the beginning of this month, Larry McVoy of Bitmover fame stated that his company was going to phase out the free (as in beer) Bitkeeper client that a number of Linux developers use. As a result of this, Linus Torvalds took a week to figure out alternatives to Bitkeeper, and finally, not finding anything that fit his needs, started writing git. Git development has rapidly intensified, with other developers joining. As this is not a standard SCM and doesn’t compete with existing solutions like Arch, Darcs, or Bitkeeper, it works for them.

However, even if the influence of this change on kernel development has been quite well contained, it looks like it « fuzzied » the process a bit. I currently more patches than ever (still not much, but this is significantly more than usual) in my patchset. Some of these are from BenH or David Brownell, the cpufreq ones are required to get 2.6.12-rc3 to even compile on PPC. The ones that are from me have been pushed to various maintainers but until now, they only landed in -mm. There’s not even a 2.6.12-rc3-mm1 on kernel.org yet…

Looks strange. In my experience, patches found their way to mainline really fast. Let’s hope they’ll get up to speed after 2.6.12!

Update: 2.6.12-rc4 just went out, and it contains all of the patches I needed to get a rock stable sleep support on the iBook. Nice :)