Convert work to pbranch
Peter Arrenbrecht
peter.arrenbrecht at gmail.com
Fri Jan 2 01:54:06 CST 2009
[cc'ing list]
On Fri, Jan 2, 2009 at 7:08 AM, TomBa (hg list)
<a3nbau902 at sneakemail.com> wrote:
> How can I convert existing mq patch series or perhaps set of (linear)
> mercurial changesets to pbranch?
> I am looking for something similar to hg qimport --rev.
> I could apply patches manually and issue hg pnew one by one, but I am
> looking for something better.
This does not currently exist. But I can see the need. Cannot tackle
it right now, though. If you want to give it a try, I suggest two
forms:
One where a linear MQ patch series is converted to a linear pbranch
DAG, following the series file.
One where the MQ patch series is rearranged into a true DAG, using a
pre-written patchgraph file as input to define the relationships.
>
> How to convert pbranch patches to regular changesets in order to push
> changes to central repo?
> If the best way to hg pexport --queue, and hg qfinish -a?
Yes, if your patches are linearized.
>
> I am thinking of giving pbranch a try (I use mq a lot), for now primarily
> for:
> - the promise of better merges than in mq
Check.
> - perhaps easier way of moving changes between patches or splitting patches
> or reordering patches
Check for moving (in my opinion).
But reordering will be harder. The pbranch way is to keep unrelated
patches unordered from the start (parallel patches in the DAG). This
is why I suggest convertion MQ to pbranch following a prewritten
patchgraph file.
> - perhaps easier sharing work between different locations, when working from
> home (edit/graphical-merge/graphical-diff at home - but can't build/test
> here; push/pull to work).
Check. Collaboration (even between two copies of yourself) is one of
the main drivers behind pbranch.
> However, I want a way to convert (if needed) between untried pbranch and
> known-to-work-for-me mq.
>
I guess the more important part is the option to ditch pbranch again,
and that already exists via `hg pexport --queue`.
-parren
More information about the Mercurial
mailing list