Corruption issue from filesystem exception.

[Doug Philips]

On or about Monday, February 09, 2009, at 03:28PM, Matt Mackall indited:
>On Mon, 2009-02-09 at 21:16 +0100, Sune Foldager wrote:
>> On 09/02/2009, at 18.02, Matt Mackall wrote:
>> > There's no such thing as a Mercurial-style rollback for strip because
>> > it's deleting data. Rollback simply truncates back to a known-good
>> > point. Any real undo-like functionality for strip would be  
>> > vulnerable to
>> > the very same unexpected things that would cause strip to break in the
>> > first place.

Wait a minute here, something doesn't seem right...
Anything that is added to Mercurial can be un-added (rollback).
Mercurial is append only, so the rollback "checkpoint data locations" will -never- be invalidated by any future additions to the repo.
So, how "bad" would be it, philosophically, for each commit to store its own rollback data?
No, that isn't strip (considering alternating commits on independent branches), but it sure seems a lot less risky and much more core-Mercurial-y than strip...
(If this has been shot down before, pointers to, or search text for, the email archives appreciated).

--Doug