<html>
<head>
<base href="https://bugs.freedesktop.org/" />
</head>
<body>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW --- - cairo-tor-scan-converter is off by half of a subrow."
href="https://bugs.freedesktop.org/show_bug.cgi?id=84396#c14">Comment # 14</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW --- - cairo-tor-scan-converter is off by half of a subrow."
href="https://bugs.freedesktop.org/show_bug.cgi?id=84396">bug 84396</a>
from <span class="vcard"><a class="email" href="mailto:sixtysix@inwind.it" title="Massimo <sixtysix@inwind.it>"> <span class="fn">Massimo</span></a>
</span></b>
<pre>With the last patch, comparing the output of the test a8-clear
with its 180°-rotated version, shows that all pixels computed
via sub_row are correct, while those with full_row are slighlty
wrong.
I've obtained the opposite result not altering the way e->x.quo .rem
is computed, but adjusting the rest.
The problem is that to share the same walker between full_row and
sub_row, it would be necessary to use a grid with twice as many y
samples as the number of subrows. Cell corners for full_row and
cell centers for sub_row.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are the QA Contact for the bug.</li>
</ul>
</body>
</html>