On Sep 23, 2004, at 12:10 PM, Scott Palmer wrote:
>
>> The only difference between "real tags" and "all that fake copying"
>> is your own concept of what's going on. They both operate in exactly
>> the same way. A copy only takes up "1 unit" of the repository and
>> are exactly what you're asking for, aliases for a revision, it does
>> not duplicate all the files. A "real tag" would also have to be
>> stored in the repository, except in svn it's simply part of the file
>> hierarchy.
>
> There is a difference. The 'copy' style of tag can be changed still,
> it is not 'pinned' to a specific revision.
>
> Maybe the thing to do in Rick's case is to use a property to keep
> track of the last revision that was merged?
>
Or, write a log message that says, "I'm merging revisions X thru Y
now.". That's what the book recommends. You just run 'svn log' the
next time you're getting ready to merge, so you know where to pick up
from.
This is one of the many problems ("the repeated merge problem") taht we
hope to solve when we implement merge tracking. The VC system should
be remembering what was merged where, rather than forcing users to keep
track manually.
By the way: the CVS style of using tags to solve the "repeated merge"
problem still works in SVN. You can create a tag of each tree you're
comparing in the merge, using whatever descriptive names you want. The
next time you come back to repeat the merging process, just compare the
last tag with some new tag.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Sep 23 20:01:42 2004