[chore] Port missing Chromatic adaptation tests

[lloydk]

There are 13 tests in the old Chromatic adaptation test suite and 2 in the new test suite

[LeaVerou]

Wow, no idea how that happened! Thanks!

Can you edit your comment with links to both files? (in general if you can add links to these issues, it would help). Thanks!

[lloydk]

Actually, the tests for this suite have been ported but most of them are commented out. Most of the commented out tests fail if they're run.

Old tests:
https://github.com/color-js/color.js/blob/main/tests/adapt.html

New Tests
https://github.com/color-js/color.js/blob/main/test/adapt.js

[svgeesus]

This is because, as the commented-out tests say:

description: These test the linear Bradford adaptations from CATs.js against the matrices <a href="http://www.brucelindbloom.com/index.html?Eqn_ChromAdapt.html">published by Lindbloom</a>

The trouble is that

• Lindbloom uses low precision rounded-off matrices
• Lindbloom uses a slightly different definition of D65 and D50, thus gets different results
• Lindbloom is AFAICT the only source for the "expected" values for adaptation to and from other whitepoints

There would be value in recalculating these to a higher precision, and using those values as "expected" to test for later regressions.