[PATCH 0 of 3] hbisect: tolerate unusual but consistent bisect situations

[Matt Mackall]

On Sun, 2011-07-17 at 00:40 +0200, Alexander Krauss wrote:
> In certain situations (when the earliest good revision is not an
> ancestor of the earliest bad one), bisect incorrectly claimed
> "inconsistent state". Apparently, it was assumed that such situations
> cannot occur, which is only true if the user strictly follows the
> bisect sequence suggested by Mercurial. Adding other information to
> the bisect state via "hg bisect -g REV" easily produces such a
> situation.
> 
> Patch 3 in this series adresses the issue and adds a test. The rest
> are minor clarifications.

Looks great, queued.

-- 
Mathematics is the supreme nostalgia of our time.