[RFC] Represent Loom threads using branches

[James Henstridge]

2008/5/7 Robert Collins <robertc at robertcollins.net>:
> On Wed, 2008-05-07 at 11:19 +0800, James Henstridge wrote:
>  > 2008/5/1 Robert Collins <robertc at robertcollins.net>:
>  > > On Thu, 2008-04-24 at 08:25 -0400, Aaron Bentley wrote:
>  > >  > -----BEGIN PGP SIGNED MESSAGE-----
>  > >  > Hash: SHA1
>  > >  >
>  > >  > I've been using the loom plugin pretty heavily, and I still feel like
>  > >  > loom threads should be branches.
>  > >
>  > >  My initial feelings are this is going to be hugely more complex for
>  > >  managing the loom history, and have significant performance implications
>  > >  too. (I expect large loom scenarios to number in the hundreds of
>  > >  patches). It will also remove the deliberate simplification of threads
>  > >  (which is not an implementation detail, its a deliberate design
>  > >  decision).
>  > >
>  > >  I will read your mail in more detail over the next couple of days.
>  >
>  > In one of my uses of looms so far, I have been pushing individual
>  > threads to separate locations.
>
>  Could you enlarge on why you need to do this ?

1. Some of the branches I've got in the loom started life elsewhere
and I want to keep those locations up to date.

2. The code review process being used for the project (Storm) is to
register the branch with the associated bug report.  I suppose I could
do that with a loom branch, but then the reviewers would need to have
the loom plugin and I'd need to tell them which thread(s) I want them
to review.

James.