[PATCH 3 of 7 Series-F] performance: speedup computation of unserved revisions

[Dave S]

Sorry, more minor stuff ...

On Tue, Jan 8, 2013 at 10:54 AM,  <pierre-yves.david at logilab.fr> wrote:
> # HG changeset patch
> # User Pierre-Yves David <pierre-yves.david at logilab.fr>
> # Date 1357327145 -3600
> # Node ID 66e2c7a2eeb8e0ce02ec87fc1ca3197efafafe20
> # Parent  fdd02003dac5afd722565ebff18fcb56a3524bd8
> performance: speedup computation of unserved revisions
>
> In they current state revset call can be very costly as we test predicates on
> the will repository.

Same comment as in #1 (and which also applies to #2)

> The "unserved" filter is used in multiple application and

"applications"

> serve in particular in some branch cache loading operation.

I think you're using "serve" to refer to the subsystem under "hg
serve", but I keep trying to read it as the verb "to serve", and I'm
not sure how make it read better.

> We better make it
> fast.
>
> The current changeset drop revset call in favor of direct testing of the phase
> of changeset.

As in #1, "calls"

/dps

-- 
test signature -- please apply at front gate on Tuesdays only.