1297986 - 4D matrix multiplication assumes w = 1.

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[Nicolas Silva [:nical]]

but w is not necessarily equal to 1. It looks like a massive foot-gun. If you really want to enforce w=1, you should be using 3D points instead.

Let's fix the multiplication to take w into account.
Any objection?

Comment 1

[Nicolas Silva [:nical]]

Attachment: Don't assume w=1

r=kats since APZ seems to be the principal user of the 4d transformation.

Comment 2

[Kartikaya Gupta (email:[e-mail])]

Comment on attachment 8784795 [details] [diff] [review]
Don't assume w=1

Review of attachment 8784795 [details] [diff] [review]:
-----------------------------------------------------------------

Redirecting to matt, because he did the relevant w-coordinate changes in APZ and my 4D math knowledge is lacking.

Comment 3

[Matt Woodrow (:mattwoodrow)]

Comment on attachment 8784795 [details] [diff] [review]
Don't assume w=1

Review of attachment 8784795 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/2d/Matrix.h
@@ +838,5 @@
>  
>    template<class F>
>    Point3DTyped<TargetUnits, F> operator *(const Point3DTyped<SourceUnits, F>& aPoint) const
>    {
> +    Point3DTyped<TargetUnits, F> result;

I think we can just keep the old code for this, unless we think we really need the performance boost of skipping the w multiply.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/07336fed8904
Don't assume w=1 when transforming 4d points. r=mattwoodrow

Comment 5

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/07336fed8904