Re: [Fwd: Re: [Monotone-devel] Lack of conflicts checking]

[Nathaniel Smith]

On Wed, Aug 08, 2007 at 09:40:27PM -0600, Derek Scherger wrote:
> Sorry, I don't entirely follow how the mark on the live side will tell
> us if it has seen a content change that the dead side has not.

That's the definition of a mark, it's the revision where the change
occurred, and so if it's an ancestor of the other side, the other side
has seen the change, if it isn't an ancestor of the other side, then
the other side hasn't seen the change.

> Do we just look to see if the common ancestor has a different content
> mark than the live side does? I think this would say that the live side
> has seen a change that the dead side has not.

No, no, this is *-merge, there's no common ancestor idea there at all.
We just compare look to see if marks are in a given revision's
ancestry.  (And as an efficiency hack, the way we actually implement
the "is this mark an ancestor of this parent revision?" check is via
uncommon ancestor sets, maybe that's confusing you?)

-- Nathaniel

-- 
The Universe may  /  Be as large as they say
But it wouldn't be missed  /  If it didn't exist.
  -- Piet Hein