Open trusktr opened 3 years ago
getClientRects()
has no problem returning the actual sizes of a non-atomic inline, so I think that ResizeObserver should just stop pretending that the size is zero and just populate the arrays with the fragment sizes.
It may be fine (or even required by compat) to keep contentRect
being an empty rect or something special, but populating borderBoxSize
, contentBoxSize
and devicePixelContentBoxSize
as normal seems useful.
In a true W3C Fashion, a confusing API that only work 30% of the time IMO they should just scrap HTML & CSS altogether and start over with a proper rendering API and Markup language I've spent days trying to make something working in a WYSIWYG component and make it react to zoom & resize changes with no success a very stupid thing that takes me 5 minutes to accomplish in Flutter or Kotlin Jetpack or with any other Sane UI Rendering APIs
The resize-observer spec says
I am guessing this is because such elements are split into pieces (rows).
I see that the latest
ResizeObserver
spec allows it to return an array of sizes when content is split across columns. Based on this, it seems feasible to also have it give us an array of the sizes of the pieces of a non-replaced inline element's rows.Or it could just give us the size of the bounding box, but this would make the use cases more limited, though the limited use cases could still be useful.
Is it simply that nothing has been decided on how to handle multiple pieces of a non-replaced inline element?