reftests: Allow large totalPixels for fuzzy matching

[kainino0x]

For these tests the fuzzy matching is only to allow for slight differences like rounding errors. We don't care about how many pixels are different.
https://web-platform-tests.org/writing-tests/reftests.html#fuzzy-matching

Also adds fuzzy matching to two more canvas_colorspace tests because it's not important to have extremely precise results for any test that does color conversions or blending (at least right now).

I haven't tested this exact change but in the linked Chromium CL below we saw the following:

01:01:43.047 7604 Found pixels_different: 500, max_channel_diff: 1
Allowed pixels_different; 0-400, max_channel_diff: 0-2

01:01:44.062 7604 Found pixels_different: 700, max_channel_diff: 2
Allowed pixels_different; 0-400, max_channel_diff: 0-2

So this should definitely work.

Some other tests that we're not currently passing for other reasons render quite large areas, so we set totalPixels very large. For example the canvas_colorspace tests render twelve (128 * devicePixelRatio)-squared canvases.

Issue: #1045, https://chromium-review.googlesource.com/c/chromium/src/+/5134470

---

Requirements for PR author:

• All missing test coverage is tracked with "TODO" or .unimplemented().
• New helpers are /** documented */ and new helper files are found in helper_index.txt.
• Test behaves as expected in a WebGPU implementation. (If not passing, explain above.)

Requirements for reviewer sign-off:

• Tests are properly located in the test tree.
• Test descriptions allow a reader to "read only the test plans and evaluate coverage completeness", and accurately reflect the test code.
• Tests provide complete coverage (including validation control cases). Missing coverage MUST be covered by TODOs.
• Helpers and types promote readability and maintainability.

When landing this PR, be sure to make any necessary issue status updates.

[greggman]

lgtm

I find it strange that totalPixels is any difference even if they are under maxDifference but 🤷‍♂️

[kainino0x]

Yeah, it's unintuitive, but that's how it's defined. I think they wanted to define it such that you could use "fuzzy" matching to almost-precisely match a particular expected error. But I don't know if it actually is ever used like that.
https://web-platform-tests.org/writing-tests/reftests.html#fuzzy-matching