[Pixman] [PATCH 12/14] pixman-filter: Fix several issues related to normalization

[Bill Spitzak]

The attached patch changes the normalization to the version I was
suggesting. I added some print statements and discovered that the "residual
error" was always zero, after some analysis I figured out that this version
will always produce a set of values that sum to pixman_fixed_1 (unless the
filter is unrealistically wide, far larger than 2^20, to get double
precision errors to sum large enough to change this). So I removed the
new_total and residual error and left a comment in there.

I am having trouble with IMPULSE.BOX producing values that are all equal or
very close together. Do you know of a way to get gnuplot to include y==0 as
otherwise the plots are very misleading unless you read the axis values.
Also this produces "Warning: empty y range [1:1], adjusting to [0.99:1.01]"
messages, possibly these are not an actual error.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/pixman/attachments/20160413/5efac2cb/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: patch
Type: application/octet-stream
Size: 2642 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/pixman/attachments/20160413/5efac2cb/attachment.obj>