Clarify: are gradients considered contentful?

[noamr]

Currently according to the spec, background images are considered contentful in terms of measuring FCP. The definition of images in CSS also includes gradients. I'm not sure they should be - they're more similar to background colors than they are to images in terms of what "content" is - but in any case the spec should maybe clarify this.

[npm1]

I tend to agree that gradients are not very contentful (given that colors aren't). I did a quick test, Chrome currently does consider them as contentful. But I'm open to changing it, and we can mention this in the spec either way.

[rniwa]

Related, we really need to more rigorously define what it means for an image to be rendered, or even what an image means in this context.

For example, [https://drafts.csswg.org/css-images-4/#image-values]( type) defines the CSS syntax for specifying images in places like background-image, and one of the concepts defined there is of an invalid image where url is pointing to a non-existent image. We need to define under which scenarios, anything specified by type is considered as an image and is rendered.

And then for each one of those properties, we need to define under what condition a specified image is considered as rendered. For example, background-image has an associated background painting area. What happens when this background painting area is 0px by 0px?

A border image with border-image-width set to 0px presumably doesn't count?

All these things need to precisely defined.

[noamr]

I tried to address this here: #66

[noamr]

I believe that #66 addresses this, closing.

[npm1]

@noamr so they will not be considered contentful, correct? Once you confirm I can add a test (Chrome will fail it at the moment).

[noamr]

@noamr so they will not be considered contentful, correct? Once you confirm I can add a test (Chrome will fail it at the moment).

Correct. Only loaded images will be considered contentful.

[rniwa]

Please don't close an issue until the relevant PR is landed. It's very confusing.

[noamr]

Please don't close an issue until the relevant PR is landed. It's very confusing.

OK, sorry :)

[noamr]

Closing now that the PR is landed. https://w3c.github.io/paint-timing/