Greg Troxel wrote:
> Your proposed change makes me nervous, because it seems to break the
> property that all merges are recorded. But, I think the rationalization
> is that it can't ever matter if changes that aren't actually changes are
> recorded. I'm trying to articulate the rules to see if I can convince
> myself this is ok, and having trouble. (I know I'm being redundant in
> my comments below, but I'm just barely following myself.)
>
I think we're basically talking about two different, rather orthogonal
use cases for recording merge info.
In the first case, you want to record the formal, logical history of
merges in the repository; that implies recording mergeinfo for every
merge, even if it did not result in changes in the target. That would
imply a "formally complete CM system".
In the second case, you only record operations that actually had an
effect on the repository. Subversion has always taken that approach for
commits, where unchanged files do not get a new version in the repo.
(We're even more "leaky" with revprop changes.) So from the point of
view of consistency, it makes sense not to explicitly record merges
where they had no effect.
-- Brane
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2380747
Received on 2009-08-06 11:10:42 CEST