Use of named branches

Sebastien Lucas sebastien.lucas at gmail.com
Sat May 5 17:47:21 UTC 2007


On 5/5/07, Guido Ostkamp <hg at ostkamp.fastmail.fm> wrote:
> Hello,
>
> > Here is your second example:
> >
> > hg glog
> >
> > o  changeset:   1:9a534b2e2d3b
> > |  branch:      seb-dev
> > |  tag:         tip
> > |  summary:     commit to seb-dev
> > |
> > @  changeset:   0:f057905ed14e
> >   branch:      trunk
> >   summary:     commit to trunk
> >
> > In the second example, it is impossible to merge with trunk, because
> > seb-dev is directly descended from trunk and there is no diverging
> > change.
>
> sorry, but that doesn't sound logical. A file has been added so that the
> working copy is different, and you say, that's no change?
>
> It you like to develop in branches, I think the aforementioned scenario is
> quite a common one.
>
> How do you then get the change back into your mainline (=trunk) branch, if
> not by 'hg merge'?
>
> Regards
>
> Guido

I have to agree with Guido. Coming from CVS or SVN it really seems
illogical. I somehow understand Daniel's explanation, but I'm really
not convinced of the usefullness (I hope it's an english word) of
named branches in that case.

I would consider more logical that any unmerged branch count as a head
and you have the possibility to open/close (by merging or dropping) a
branch.

I also understand that branching with hg clone solves everything but
it really seems overkill (to handle, to backup) for me alone.

Anyway thanks you both for your answers.

Sebastien



More information about the Mercurial mailing list