On 13 May 2013 23:42:46 -0600
"Andy Bradford" <amb-fos...@bradfords.org> wrote:

> > That is, it's backwards: you first do some work, then decide to
> > commit and  decide  this commit  should  start  its  own branch
> > rather  than continuing  the current  one,  so  you create  that
> > new branch  while committing.
> 
> Not  exactly backwards,  but  more  of a  convenience.  If you
> suddenly discover that  your changes should  not yet go  into the
> trunk,  you can branch at the moment of the  commit. This makes
> branching extremely easy to accomplish and less of a  hassle to get
> the changes committed without breaking the rest  of the sources.
> With  other VCS you would  have to go through a few  more girations
> to get your newly  changed files committed at the *right* location.

Changing a branch "under a dirty work tree" is achieved with a single
command in Git and two commands in Subversion; not sure about less
mainstream VCSes.  But I digress.  If memory serves me right, initially
Fossil only supported creation of a branch when committing, and
"explicit" branch creation has been added afterwards, per user requests.

I would say this is not about convenience but rather about different
mindsets: some people just don't "feel right" when they can't start a
new line of by *first* creating a branch to work on.  People with
a different mindset are fine with first coding something and then
deciding where this should go.  There definitely was at least one
thread on this list which discussed this matter, with Richard Hipp
chiming in and explaining he has the mindert of the latter kind ;-)
Of course, the convenience argument still holds, just not when it comes
to history of implementation of these features.

And if, again, memory serves me right, explicit creation of a branch
works by creating an empty commit on that newborn branch -- specifically
for the purpose of attaching the necessary tags to it, which are to
designate the branch.  I'm inclined to think this hints at that such a
mode of operation wasn't initially envisioned.
_______________________________________________
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Reply via email to