Tortoise SVN having hernias over existing local copies.

Hi,

We have been using Tortoise SVN for about 9 months.

We have a development team in Australia, and one in UK.

Both sites have local copies of the source code (in identically
structured trees) and we use the central repository (on a Linux Box)
as a way to store revisions and a mechanism for exchanging changes.

However, we keep upsetting Tortoise in the following way :

1) Someone emails a new source record direct to the other party,
(before the periodic subversion commit has been performed) and the
receiver then installs it in their local directory.

2) The sending party subsequently does a SVN Commit.

3) Later again, the receiving party attempts to do a SVN Update
only to be told "Cannot create object - file of same name already exists"
which brings the SVN Update to a screaming halt, and leaves things
in a pickle.

Subsequent attempts to re-update the receiving end produce such memorable
messages as :

"File not found: transaction '7y', path '/BP/WS/<some existing file name>'

(where the file name in question is NOT one that was changed locally or
remotely recently BUT is present in the directory that was being updated
at the time the update aborted).

Tortoise Cleanups do NOT solve the problem, and essentially the only thing
we have found to fix it is to delete and re-check-out the affected directory. :-((

I HAVE SOME BASIC QUESTIONS THEREFORE :

1) If Tortoise can happily merge changes into existing files, and identify potential
conflicts etc, why on new items can't it just say "OK, that file is already there
and is identical" or bypass it, or warn and keep going ? WHY does it abort the
download process ? Isn't this a BUG / unnecessary limitation ??

2) Is there anything that can be done to fix the pickle without having to delete
the whole directory and check-out again ?

3) Are we missing the basic boat here ? Is tortoise designed to cope with the
basic situation of BOTH development teams having full local copies ? Are
we procedurally missing something very basic ? It seems that we are
constantly running into problems with the tool in the way that we are trying
to use it.. that suggests to me that we are not using it correctly, or it is not
the correct tool for the situation we face.

Is it as simple as always running updates and commits in a certain order ?

Any simple knowledgeable pointers greatly appreciated !!

Warren Kinny
Managing Director
Exodus Systems, Horsham, England
wjkinny@resmaster.com

--