[cairo-bugs] [Bug 15349] bad clipping with EXTEND_NONE

[bugzilla-daemon at freedesktop.org]

http://bugs.freedesktop.org/show_bug.cgi?id=15349





--- Comment #7 from Owen Taylor <otaylor at redhat.com>  2008-04-07 06:28:29 PST ---
Created an attachment (id=15740)
 --> (http://bugs.freedesktop.org/attachment.cgi?id=15740)
My suggested patch

This patch moves the rectangle intersection outside of the identity-matrix
optimization, and also replaces the open-coded intersection (which would
have been buggy if the rectangles didn't intersect at all say) with
cairo_rectange_interact().

I get a failure with the xlib version of filter-bilinear-extents with
this patch, but looking at the output, it's a small variation evenly
distributed over the image and I think attributable to getting hardware
acceleration of the scaling.

For a full test run, I'm not really sure what failures are "expected"
on my hardware/software and what not, so testing yourself would be
suggested. (I saw an extend reflect failure that bothered me a bit,
but I can't see how it could have anything to do with this patch 
since extend reflect doesn't even get down to the section of code
that I changed. But if that's a regression, then the setup code for
the test must be breaking somehow.)


-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.