[PATCH 15 of 21 RFC] revlog: allow importing revision deltas based on a censored file tombstone

[Michael Edgar]

On Thu, Sep 11, 2014 at 1:58 PM, Matt Mackall <mpm at selenic.com> wrote:
>
> On Wed, 2014-09-10 at 20:26 -0400, michaeljedgar at gmail.com wrote:
> > # HG changeset patch
> > # User Mike Edgar <adgar at google.com>
> > # Date 1409776499 14400
> > #      Wed Sep 03 16:34:59 2014 -0400
> > # Node ID 1e85b2cd488fb1b13517345b3ff425e0da7b48b9
> > # Parent  8adefa28d6fb5816ff9623408ef53c1d2d70f884
> > revlog: allow importing revision deltas based on a censored file tombstone
>
> It's great that you're thinking ahead this far, but we don't need to see
> all of this in your initial patchbombs. Shorter batches are better as
> reviewer attention is scarce.

That's good to know, I wasn't sure how best to share all this work.

One question: what should I do when later patches justify decisions
made in earlier patches? For example, this patch and its 3 successors
require the "metadata" attribute on CensoredNodeError which you noted
as questionable in patch 3 of 21.

Should I first send out a shorter branch without that attribute, then
add the attribute when I send a branch which uses it? That certainly
makes sense for this example, given that there were 12 patches between
introduction and use.