1734479 - [css-transforms] WPT test transform-interpolation-005.html fails in Firefox, due to issue with 2d vs 3d identity matrix

Status: RESOLVED DUPLICATE of bug 1850841

(Core :: CSS Parsing and Computation, defect)

Description

[Daniel Holbert [:dholbert] (PTO until 2/24)]

transform-interpolation-005.html has 4 subtest-failures in Firefox:
https://wpt.fyi/results/css/css-transforms/animation/transform-interpolation-005.html?label=master&label=experimental&aligned&q=firefox%3Afail

All of them are of the form:

assert_equals: expected "matrix ( 1 , 0 , 0 , 1 , 0 , 0 ) "
but got "matrix3d ( 1 , 0 , 0 , 0 , 0 , 1 , 0 , 0 , 0 , 0 , 1 , 0 , 0 , 0 , 0 , 1 ) "

...for animations to [none]. So it looks like we're producing an identity matrix of a different dimensionality (2d vs 3d) vs. what the test expects.

This is part of the compat2021 test set, for what it's worth:
https://github.com/Ecosystem-Infra/wpt-results-analysis/blob/master/compat-2021/css-transforms-tests.txt

Comment 1

[Daniel Holbert [:dholbert] (PTO until 2/24)]

Side note: I think this test's expected $FOO but got $BAR wording happens to be backwards here.

The error is phrased as "expected [2d transform syntax] but got [3d transform syntax]", but in fact it looks like the test is expecting the 3d transform syntax, and it must be mixing up the args for assert_equals at some level of abstraction.

So I suspect our interpolation is producing a 2d identity matrix, but the test is expecting a 3d identity matrix (which is a reasonable expectation, given that the test is animating from a 3d transform to none).

Comment 2

[Daniel Holbert [:dholbert] (PTO until 2/24)]

transform-interpolation-verify-reftests.html (in the same directory) looks like it might be failing for the same underlying reason -- its failures are all mismatches between a matrix(...) and matrix3d(...) expression.

https://wpt.fyi/results/css/css-transforms/animation/transform-interpolation-verify-reftests.html?label=master&label=experimental&aligned&q=firefox%3Afail

Comment 3

[Boris Chiou [:boris]]

I missed this bug. I'm pretty sure Bug 1850841 should fix this. :)

Comment 4

[Thomas Wisniewski [:twisniewski]]

@dholbert, is this fixed now? The two WPTs you link to above seem to be passing, at least.