Re: [SVNMERGE][PATCH] handle -r X-Y, where X > Y

On Tue, 16 May 2006 04:19:08 +0530, Giovanni Bajo <rasky@develer.com>
wrote:

> Madan U Sreenivasan <madan@collab.net> wrote:
>>> [[[
>>> Accept -rX-Y, for all svnmerge commands, where X > Y.
>>> Before this patch, empty RevisionSet()s were created when X > Y.
>>>
>>> * contrib/client-side/svnmerge_test.py
>>> (TestCase_RevisionSet.test_reverse_order_input_string): New
>>> test. Try creating a RevisionSet() with a string range, in
>>> descending order.
>>>
>>> * contrib/client-side/svnmerge.py
>>> (RevisionSet.__init__): Sort the input revision range before
>>> filling in RevisionSet._revs.
>>> ]]]
>
>
> Rationale?

When we have the 'rollback' feature in place, it would make more sense for
the user to say

svnmerge rollback -r45-30

than say.

svnmerge rollback -r30-45

     Though, the rollback functionality should be written to handle both
appropriately.

> "svn merge" accepts the reversed order with a different semantic
> (reversed merge).

I only know that 'svn merge' understands a reverse merge if you simply say
-r 45-30. Am sorry, could you explain this different semantic?

> I'm worried about the confusion that can issue.

I encountered this when coding for rollback (as above). There are no tests
specifying the range in reverse order. So, I was thinking, "Maybe nobody
else is using this, but NOT that nobody else needs it, so I put it in
place".

> I am +1 on a
> patch that errors out when X > Y until there is agreement on how to best
> handle
> this.

IMHO, Lets discuss this, but lets not error out unless we decide to do so.

Regards,
Madan.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue May 16 07:28:17 2006