Looks like this produces sensible results for interpolation with 0, though I'm
not really convinced about the results from, let's say, 1px to 2000px in the
attached test-case, I would've expected a linear interpolation from that to go
through normal length interpolation.
css-transforms-1 says:
> Two transform functions with the same name and the same number of arguments
> are interpolated numerically without a former conversion. The calculated
> value will be of the same transform function type with the same number of
> arguments.
>
> Special rules apply to <matrix()>.
Which is what we do... I was going to file a spec issue but turns out that it's
already addressed in css-transforms-2:
https://drafts.csswg.org/css-transforms-2/#interpolation-of-transform-functions
Which says:
> The transform functions <matrix()>, matrix3d() and perspective() get
> converted into 4x4 matrices first and interpolated as defined in section
> Interpolation of Matrices afterwards.
Differential Revision: https://phabricator.services.mozilla.com/D4942
Per bug 1322189 we really should. I've copied the setup we have already for
translate / scale, but we should really clean this up a bit more I'd think.
In any case, probably skew should be matched as well...
Bug: 1464615
Reviewed-by: hiro
MozReview-Commit-ID: Jky5k8HVfuH
It's not sound to insert random matrices in random positions in the transform
operation list.
I cannot make any sense of what the old code was trying to do.
Bug: 1458715
Reviewed-by: hiro
MozReview-Commit-ID: 5BtCiueEPlR
DOMMatrix needs to convert a specified transform list into a matrix, so
we could rewrite to_transform_3d_matrix by generics for both specified
and computed transform lists.
Besides, we have to update the test case because we use Transform3D<f64> to
compute the matrix, instead of Transform3D<f32>, so the result will be
the same as that in Gecko. Using 0.3 may cause floating point issue
because (0.3f32 as f64) is not equal to 0.3 (i.e. floating point precision
issue), so using 0.25 instead.
Sometimes, we want to use extremely small pixel value in transform, e.g.
translate(calc(0.001px)), which should be rendered correctly together
with scale(100000). This patch only fixes this case for Servo because
Stylo still uses Au on the Gecko side, even if we pass pixel value into
FFI directly in the previous patch.
We replace Au with CSSPixelLength for the length part of
computed::CalcLengthOrPercentage. Therefore, it would be easier to use
CSSPixelLength for all other LengthOrPercentage{*} types.
First, we define computed::CSSPixelLength which contains a CSSFloat, a
pixel value, and then we replace computed::Length with CSSPixelLength.
Therefore, the |ComputedValue| of NoCalcLength, AbsoluteLength,
FontRelativeLength, ViewportPercentageLength, CharacterWidth, and
PhysicalLength is CSSPixelLength.
Besides, we drop NonNegativeAu, and replace computed::NonNegativeLength
with NonNegative<computed::Length>. (i.e. NonNegative<CSSPixelLength>)
We could use this method to convert a TransformList into a Matrix, and
use this matrix for computing distance for Stylo and rendering the transform
for Servo.
This is an equivalent of nsStyleTransformMatrix::ReadTransforms in Gecko.
