[svk-devel] Re: 2.0-pre "svk pull" tries to merge back to mirror source?

[Derek Atkins]

"David Glasser" <glasser at MIT.EDU> writes:

> On 10/11/06, Pazu <pazu at pazu.com.br> wrote:
>> Still, svk shouldn't smerge *to* the remote repo during a pull. AFAIK, a
>> pull should merge to the local repo, while a push would merge to the
>> remote repo.
>
> Well, here's the question:
>
> svk cp //mirror/project/trunk //local/project/branch
> svk co //local/project/branch
> svk pull
>
> svk cp //mirror/project/trunk //mirror/project/branch
> svk co //mirror/project/branch
> svk pull
>
> The fact that one branch happens to be stored locally and the other
> happens to be stored on a server doesn't say to me that "pull to the
> branch" should be different in the two cases.
>
> On the other hand, people certainly do get bitten by "svk pull makes a
> merge happen from a totally unexpected copy, on the server".

This is where I see the difference..  I don't think of 'push/pull'
as handy wrappers around smerge. I don't consider it syntactic
sugar.  I consider them specific operations around mirrors.

Maybe I'm alone in that view, or maybe I'm in the minority.  Perhaps
it's because I've used Teamware and Bitkeeper and that's what
push/pull means in those contexts.

But to me, I believe that push/pull is specific to mirroring and
therefore in your examples above /I/ believe that "svk pull" should do
different things.  In the first case is should be sync; smerge; update
and in the second case it should be sync; update.

> --dave

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available