How we deal with Canvas Support

[jasonwilliams]

We have 2 implementations of Container Timing, in both Firefox and Chromium.
Chromium doesn't support canvas right now, so the canvas example yields no PerformanceContainerTiming entries, however the Firefox example does.

The way the Firefox one works is it will yield a single Container Timing entry on the first paint of a canvas element and the size of that entry is the entire canvas, this means we'll only ever get one entry per canvas element as the paint checking isn't more granular than that. I suppose this is working as expected, similar to an image which will take up the full area regardless of what the image is made up off.

Are there any thoughts on this approach?

[jasonwilliams]

What we need to do in our spec is refer to this list of eligible element https://w3c.github.io/paint-timing/#timing-eligible and mention that this list needs to grow, because right now it doesn't cover canvas and video which Firefox supports.

This will need some discussion

[jdapena]

What we need to do in our spec is refer to this list of eligible element https://w3c.github.io/paint-timing/#timing-eligible and mention that this list needs to grow, because right now it doesn't cover canvas and video which Firefox supports.

This will need some discussion

LCP and Element Timing refer to that already, so we can have a similar section.