On Mon, 19 Sep 2005, Peter N. Lundblad wrote:
> On Mon, 19 Sep 2005, Daniel Rall wrote:
...
> > Sure, this wouldn't be a deep copy (with history), but a convenience for
> > shallow copies which avoids a 'svn add' after shell-level 'cp'.
> >
> And if we someday want to implement true copies between repositories,
> don't we have "painted ourselves into a corner" then? We would have to
> preserve this semantic, just to avoid the user typing a command now.
I don't think the semantic would be worth preserving. When someone runs
'svn cp wcA/foo wcB/bar', they want the full history. The fact that the
convenience implementation didn't copy the entire history would be more
of an annoyance than a desired behavior, but possibly less annoying than
the error message they're going get with 1.3.0. I look at it more as a
step in the right direction. That said, I'd rather have the full "copy
with history" than the incremental step, but it's significantly more effort,
unless implemented in some cheesy fashion like a series of commits of every
revision of a resource (which would not preserve as much meta data as might
otherwise be possible).
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Sep 19 22:53:23 2005