where to base patch series off of (was: [PATCH 0/6] Lex, Yacc and explicit declarations of dependencies.)

[Ralf Wildenhues]

Hi Stefano,

* Stefano Lattarini wrote on Tue, Jul 06, 2010 at 11:36:26AM CEST:
> At Tuesday 06 July 2010, Ralf Wildenhues wrote:
> > Please give me another week to review the series;

> OK; but note that this message wasn't meant to hurry you: if this
> change goes in automake, it's not gonna need a fast review, but rather
> a careful one, even if it's slow (as is usually the case with careful
> reviews).

OK.

> I just thought that the feature could be a reasonable candidate for a
> new branch, which BTW would have given me a first opportunity to mess
> around with remote branching and pushing without touching/endangering
> the master/maint branches.

Indeed.

> > for that, it
> > shouldn't be necessary to post updated patches nor push a branch;
> > I have your patches in an old branch sitting here.

> Agreed.  But I think it would be a good idea anyway to edit the
> ChangeLog of the [PATCH 2/6] to cite also Solaris 10 /usr/xpg4/bin/make
> rather than only Heirloom make.

Sure.  I didn't mean that I would push the old branch; just that for
reviewing I won't necessarily need an update.

> > If you haven't
> > heard back from me by next Monday, then feel free to push a
> > branch, but please rebase it against maint not master, if that is
> > at all possible for the changes in this series.

> Yes, it's definitely possibile.  But why maint?  I'm not so sure the
> change should to be considered as maintainance/bugfixing only...

Oh, I didn't mean that the patch series should be committed to the maint
branch.  Rather, that it should be committed to a new branch which
itself is based off of the maint branch, rather than based off of
master.  This doesn't mean that the patch series should be merged into
maint at some point into the future, but that it *could* be done; IOW, I
would like to be able to postpone the decision about where to merge the
branch to.  A more likely scenario, namely that it is merged into master
only but not maint nor branch-1.11, is still possible, and rarely more
work.

Does that make sense?  If not, I can try to draw a picture to clarify
the situation.

Cheers,
Ralf