You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Should elements only report paints once to Container Timing? #72
In Chromium the behavior is inherited from ElementTiming (and sharing some logic with ICP/LCP). An element will only report one time. For images, it is when the load is complete and we get the paint for that. For text it only reports once for the text aggregator.
ICP has some open questions about what to do here.
As Container Timing is inspired on the idea of LCP, we should likely explicitely state in the specification that the elements will only report once, in the same way LCP does (or refer somehow to that part of LCP specification).
I don't think this is correct. If an element paints again in a different part of the container, that paint should likely be reported, otherwise the portion of the container that is painted could grow without knowing what element contributed that growth.
3. I don't think this is correct. If an element paints again in a different part of the container, that paint should likely be reported, otherwise the portion of the container that is painted could grow without knowing what element contributed that growth.
Current Chromium behavior:
If an element paints again in a different part of the container, without being reattached, then it will not report to container timing.
If it is reattached, then it will count again though.
We will eventually have to discuss how to cover repaints, to make the API cover more cases.
Right, I think my understanding of the spec is basically that it already does this correctly, an element moves down (for example a new element is inserted above it, it's top value is changed, etc.) without reattachment, the paint is reported again, for that same element and that portion of the container is now considered painted.
Activity
Report each element's first paint, and again after a reinsertion
Bas-moz commentedon Nov 5, 2026
I don't think this is correct. If an element paints again in a different part of the container, that paint should likely be reported, otherwise the portion of the container that is painted could grow without knowing what element contributed that growth.
jdapena commentedon Nov 5, 2026
Current Chromium behavior:
We will eventually have to discuss how to cover repaints, to make the API cover more cases.
Bas-moz commentedon Nov 5, 2026
Right, I think my understanding of the spec is basically that it already does this correctly, an element moves down (for example a new element is inserted above it, it's top value is changed, etc.) without reattachment, the paint is reported again, for that same element and that portion of the container is now considered painted.