[Python-Dev] crosses branches?

[skip at pobox.com]

    Sjoerd> Since you have a local change, you cannot use hg pull -u (or
    Sjoerd> rather, hg update).  Hg wouldn't know where to update to since
    Sjoerd> there are 2 heads in the branch you're on: the new head from the
    Sjoerd> server and your own head with your change.

I see lots of heads, but that never prevented an update before.  How do I
tell which two heads are key to this case?  For completeness, here's the
output of hg heads in my 2.5 repository.  I only see one labelled "2.5".
What is the other head to which you referred?

    changeset:   68827:3114f26d5d54
    tag:         tip
    user:        Victor Stinner <victor.stinner at haypocalc.com>
    date:        Tue Mar 22 10:46:35 2011 +0100
    summary:     Issue #11630, issue #3080: Fix refleak introduced by
    ef2b6305d395

    changeset:   68804:f316e6d6271a
    branch:      2.7
    parent:      68801:6b2edc385ffe
    parent:      68803:b99c94261225
    user:        Martin v. L?wis <martin at v.loewis.de>
    date:        Mon Mar 21 10:32:02 2011 +0100
    summary:     null merge

    changeset:   68309:c3caaf979b9e
    branch:      2.5
    parent:      68263:7790ad8332ba
    user:        Skip Montanaro <skip at pobox.com>
    date:        Sun Mar 06 21:31:25 2011 -0600
    summary:     manually expand the defunct HeadURL subversion keyword

    changeset:   68249:4cd9f5e89061
    branch:      3.0
    user:        Georg Brandl <georg at python.org>
    date:        Sat Mar 05 15:09:43 2011 +0100
    summary:     Close 3.0 branch.

    changeset:   68241:b77918288f7d
    branch:      legacy-trunk
    user:        Georg Brandl <georg at python.org>
    date:        Sat Mar 05 14:57:44 2011 +0100
    summary:     Close legacy-trunk branch.

    changeset:   68239:ceec209b26d4
    branch:      2.4
    user:        Georg Brandl <georg at python.org>
    date:        Sat Mar 05 14:56:41 2011 +0100
    summary:     Close 2.4 branch.

    changeset:   68237:364638d6434d
    branch:      2.3
    user:        Georg Brandl <georg at python.org>
    date:        Sat Mar 05 14:55:46 2011 +0100
    summary:     Close 2.3 branch.

    changeset:   68235:61b0263d6881
    branch:      2.2
    user:        Georg Brandl <georg at python.org>
    date:        Sat Mar 05 14:55:23 2011 +0100
    summary:     Close 2.2 branch.

    changeset:   68233:e849d484029f
    branch:      2.1
    user:        Georg Brandl <georg at python.org>
    date:        Sat Mar 05 14:54:46 2011 +0100
    summary:     Close 2.1 branch.

    changeset:   68231:5fd74354d73b
    branch:      2.0
    user:        Georg Brandl <georg at python.org>
    date:        Sat Mar 05 14:54:19 2011 +0100
    summary:     Close 2.0 branch.

    Sjoerd> There are two possibilities: hg merge + hg commit to merge your
    Sjoerd> change with the change from the server, or hg rebase (after you
    Sjoerd> enable the extension and with extra arguments to specify source
    Sjoerd> and destination revisions).  This would take your change and
    Sjoerd> graft if on top of the head from the server.

Looking here at the Common Cases section:

    http://mercurial.selenic.com/wiki/RebaseProject

I can see how I have now apparently got this situation:

     C1 --> C2 --> S1
            \
             \ --> ... Cn

(where S1 is my local change and C? are the central changesets)

Rebase will do this, right?

     C1 --> C2 --> ... --> Cn --> S1

Am I going to have to rebase then every time I re-pull that repository and
incorporate new upstream changes?  Wouldn't I rather want the chain of
revisions to look like this after the rebase step?

     C1 --> C2 --> S1 --> ... --> Cn

Skip