On Mon, 2008-09-08 at 10:04 +0200, [EMAIL PROTECTED] wrote:
> Clearly trapping the special case of only scaling in one dimension is a
> worthwhile optimisation. That would seem to be a separate issue from
> fundementally changing the scaling algo. Could I suggest you break these
> two changes into separte patches.
You misunderstood my patch then. The patch is not an optimization, not
at all. It just fixes a bug in what used to be the current code (used to
be as that patch has been committed to SVN trunk in the meantime). The
code simply treated this case wrongly. It wasn't slow, it created an
obviously wrong result.
We will have time again in the next development cycle to improve this
code. Or we might decide to use GEGL for scaling. For now, I believe
that the code in trunk handles all cases reasonably well and we should
stick to it for the upcoming 2.6 release. Of course, if someone finds a
corner-case that the code handles incorrectly, please let me know about
Gimp-developer mailing list