As far as I remember from hg documentation, it's better to create a clone at the top repository level (ex: kde4-specs-dev -> kde4.2-specs-dev) rather than branch. It's just a matter of team consensus.
If you want to avoid head merging, try using patch queues (http://blogs.sun.com/sunwg11nprg/en_US/entry/using_mercurial_queues_extension) hnhn Adriaan de Groot wrote: > On Monday 30 March 2009 11:27:16 am hajma wrote: >> I'm a bit confused about the branches, could someone please give me the >> missing steps? I don't want to break things - again ;-) > > I'd say "clone (then you're in the default branch) and do stuff, like edit, > hg > add, hg rm, and then hg commit followed by hg push. Do not touch the hg > branch command; do use hg pull -u regularly." That's just basic Mercurial > usage. It's easy to become branchy, that's one of the characteristics of a > DVCS. > > > But just bang away - it's easy to *unbranch* as well. > > That said, it's probably a good idea to weigh in on Luc's thread about SCM > usage in general; it seems that Mercurial does present conceptual problems of > its own. Unlike SVN, where we couldn't have branches like we do now (and > don't want) at all. > -- Jan Hnatek jan.hnatek at sun.com
