Bug in RecolorImage() when calculating alpha channel

[rockcreekdan]

I believe there is a bug in RecolorImage() in fx.c when calculating opacity.

For an order 5 matrix, the code calculates opacity as:

recolor_pixel.opacity=(MagickRealType) QuantumRange-k[15]*pixel.red+
k[16]*pixel.green+k[17]*pixel.blue+k[18]*(QuantumRange-
pixel.opacity)+k[23]*QuantumRange;

So for a matrix like this:

00 01 02 03 04
05 06 07 08 09
10 11 12 13 14
15 16 17 18 19
20 21 22 23 24

opacity = QuantumRange - 15R + 16G + 17B + 18A + 23

But I think it's either missing a set of parentheses or needs minuses instead
of all the pluses:

opacity = QuantumRange - ( 15R + 16G + 17B + 18A + 23 )
opacity = QuantumRange - 15R - 16G - 17B - 18A - 23

If I adjust my matrix so that the G, B, and A channel weights and the constant
are negative from what I want, then it works as expected. (That is, it works
as I would expect it to work without my changing the signs on my matrix.)

The same bug appears to be in the order 6 matrix for CMYKA.

I'm using the following version, but the above code is from ImageMagick 6.4.1

Version: ImageMagick 6.3.7 02/19/08 Q16 http://www.imagemagick.org
Copyright: Copyright (C) 1999-2008 ImageMagick Studio LLC

[magick]

We have a patch for the problem you reported in ImageMagick 6.4.1-4 Beta available sometime tomorrow.