Re: Merge tracking proposal, rev 2

Daniel Berlin wrote:
> They are similar, but
> 1. SVK only stores merge info at the roots

If by "roots" you mean "any directory that was the target of a merge" then yes,
that's right. Can we come up with a different term for this (since your
proposal cascades storage downward from any such "root"), and "root" already has
connotations that conflict with this usage?

> 2. SVK keeps the UUID of the repo in the list
>
> The rest of all the differences can be attributed to the fact that SVK
> has the entire repo around, and we don't, so we need to be more
> efficient on some things.

Thank you. Since your proposal didn't even mention SVK, it seemed almost like
it was developed in a vacuum (which I know was not the case).

> This part of SVK, more than any other, makes attempting to compare
> proposals for storage design somewhat nonsensical, to be honest.

It's there and it already works, so I think it justifies the comparison. As you
say, Subversion will have to do something somewhat different due to the lack of
having the full repository (or a mirror) locally at all times.

John

-- 
John Peacock
Director of Information Research and Technology
Rowman & Littlefield Publishing Group
4720 Boston Way
Lanham, MD 20706
301-459-3366 x.5010
fax 301-429-5747
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org