Vlad Georgescu wrote:
> Karl Fogel wrote:
>> Vlad Georgescu <vgeorgescu_at_gmail.com> writes:
>>> I was thinking of introducing a new subcommand ("svn relocate") that does the
>>> same thing as switch --relocate does now.
>>>
>>> A switch that only rewrites URL's without updating wouldn't be correct
>>> if there are changes between the source and target paths, would it?
>>> You'd end up with an out of sync working copy.
Only if the new URL and the old URL point to different repository locations.
But that's true of 'svn switch --relocate' today.
> I guess my point was, svn switch --relocate is wrong because --relocate is not
> (and cannot be) a flag for svn_client_switch, instead it does a completely
> different thing, takes different arguments, etc. So, it seems sane to move it
> to a different subcommand. svn sw --relocate would still work, for backward
> compatibility reasons, but then we can pretend it never existed. This change is
> orthogonal to whether my patch is applied or not, although, like you said,
> applying the patch while doing nothing about --relocate would only add to the
> confusion.
Completely agree. The decision to put --relocate onto the existing switch
was fueled by folks with an inherent fear of subcommand proliferation. The
result was a really, really confusing UI.
I agree that making a new subcommand would be better than once again trying
to overload 'switch'.
--
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Received on 2008-08-28 16:07:49 CEST