Speaking as a user of darcs, I would say that I'm pretty happy with it, and the only thing you could do to screw it up would be to put out a new release which produces bad patches or corrupts repositories.
Speaking as an organizer of the Tahoe-LAFS project, I have to say that the Tahoe-LAFS developers are unsatisfied with darcs and are rapidly moving to git (although some of them are also experimenting with bzr). Why are they doing that? 1. performance (probably just due to that Mac OS X performance problem. Can't find a ticket for it.) 2. http://bugs.darcs.net/issue992# short, secure identifiers 3. github 4. wanting to appeal to developers who are motivated by familiarity or fanboyism (i.e. new developers show up and, even though they don't know anything about darcs other than that it is not git, they ask if we can please switch to git) Speaking as a developer and release manager on open source projects, I would advise you that Reinier's original post seemed wise and that it works better for the whole community to focus on one thing at a time (inasmuch as possible), so make stable releases from trunk and encourage everyone to focus their attention on the stable release during that process. Also, I personally think that the overall pace of development on darcs will go faster if you just stop accepting patches into trunk unless they come with full unit tests. Test-driven development is a way to speed up development (over the medium term, say, over one release cycle of a stable release), not an attempt to improve quality at a cost in development speed. Regards, Zooko _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
