Fix matrix concat order in polyfill

[gn-dave-a]

The matrix concat operation was doing B . A for A.concatenating(B) which was causing issues when combining transformations.

[khoi]

AffineConcatComparison.zip

Can u take a look at this test project and the unit tests in there.

• concatenating_org vs concatenating_new (the logic in this PR)

It compares the 2 implementations with the original Apple's CoreGraphics one. and it seems like the new logic is failing?

[gn-dave-a]

Had removed that comment from the wrong file - sorted now.

[gn-dave-a]

AffineConcatComparison.zip

Can u take a look at this test project and the unit tests in there.

* concatenating_org vs concatenating_new (the logic in this PR)

It compares the 2 implementations with the original Apple's CoreGraphics one. and it seems like the new logic is failing?

@khoi are you sure the tests are right? Why do any of them check for a delta bigger than the tolerance given the use of abs? If you swap all the assertions to check for an abs delta < the tolerance, the new implemenation passes and the original one fails.

[khoi]

oops, you're right. Verified the changes and it looks correct now. Thanks for fixing