Remove some commented-out filters that shouldn't have worked in the first place, and improve comments. Not part of the build.

This commit is contained in:
dbaron@dbaron.org 2008-02-17 17:56:42 -08:00
parent 67c4f161d4
commit a4d0989379

View File

@ -51,8 +51,6 @@ Features to add:
** move log loading into popup from viewer UI
Bugs I've hit:
* Problems with k4 on feComposite type="arithmetic"; likely to be fixed by the nsSVGFECompositeElement::Filter changes in https://bugzilla.mozilla.org/show_bug.cgi?id=414784
** when fixed, can go back to the simpler way of computing filter results c1 and c2 (currently commented out)
* dynamic changes to xlink:href of feImage require removal/reinsertion from document: https://bugzilla.mozilla.org/show_bug.cgi?id=417339
** when fixed, can remove the removal and reinsertion near the end of showImages
@ -317,13 +315,16 @@ function show_differences(cb) {
<feFuncG type="linear" slope="-1" intercept="1" />
<feFuncB type="linear" slope="-1" intercept="1" />
</feComponentTransfer>
<!-- c1 will have non-black pixels anywhere that img2 is brighter than img1, and c2 for the reverse -->
<!-- Why doesn't this work?
<feComposite result="c1" in="img1" in2="inv2" operator="arithmetic" k2="-1" k3="-1" k4="1" />
<feComposite result="c2" in="img2" in2="inv1" operator="arithmetic" k2="-1" k3="-1" k4="1" />
-->
<!-- w1 will have non-white pixels anywhere that img2
is brighter than img1, and w2 for the reverse.
It would be nice not to have to go through these
intermediate states, but feComposite
type="arithmetic" can't transform the RGB channels
and leave the alpha channel untouched. -->
<feComposite result="w1" in="img1" in2="inv2" operator="arithmetic" k2="1" k3="1" />
<feComposite result="w2" in="img2" in2="inv1" operator="arithmetic" k2="1" k3="1" />
<!-- c1 will have non-black pixels anywhere that img2
is brighter than img1, and c2 for the reverse -->
<feComponentTransfer result="c1" in="w1">
<feFuncR type="linear" slope="-1" intercept="1" />
<feFuncG type="linear" slope="-1" intercept="1" />