Committer workflow

[Stephen J. Turnbull]

Dominique Quatravaux writes:
 > On Thu, Mar 6, 2008 at 12:00 PM, Stephen J. Turnbull <stephen at xemacs.org> wrote:
 > >  Yes.  That's the right thing to do for the reviewers, especially given
 > >  how strongly array-oriented Mercurial's UI is.  (Hint: show how to
 > >  retrieve the hash ID of the parent of a non-merge commit.)
 > 
 > Note to reviewers: if the linearly growing revision counter in
 > Mercurial bugs you, don't use it.

It's not that *presence* of array indicies bugs me; it's that
*absence* of ancestor operators is a serious bug in a DAG processor.

 > To get Git-style revision modifiers (eg tip^^), put this in your .hgrc:
 > 
 >  [extensions]
 >  parentrevspec =

Ie, this was added as an afterthought.

It seems to be implemented fully[2], though, so I feel a lot better.
Apparently with a big enough stick you *can* beat Mercurial into
behavior sensible enough to be usable.


Footnotes: 
[1]  By which I do *not* mean "conformant to git usage", although I
appreciate the fact that it is.  What I mean is that in light testing
all paths tried through a couple of merges were traceable.