On Wed, Mar 09, 2011 at 02:27:32PM -0800, Junio C Hamano wrote:
Jeff King [off-list ref] writes:
quoted
Hmm. Is this papering over a bigger problem,...
It is not very obvious to me if redefining the semantics of filtering done
by diff (the current definition is it is purely an output phase thing) is
necessarily a good thing. I agree that the interaction between the output
phase filtering and pruning done by the revision walker machinery is a
fine topic to discuss.
But Matthieu's patch is not papering over anything but is a real fix
within the context of the current architecture.
I consider the current state to be "buggy but we live with it" and not
an architectural decision. But that is simply a matter of perspective. :)
Certainly Matthieu's patch fixes a real problem, and it does not make it
any harder to address the other problems in the future, so I think there
is no reason not to apply it.
-Peff