Add overflow shorthand feature (with clip values)

[ddbeck]

This corresponds to https://caniuse.com/css-overflow

This is a new feature. Here are some ideas for reviewing it:

• Is this a reasonable identifier for the feature?
• Does this have a reasonable spec link?
• Does this have a reasonable caniuse reference?
• Are the mdn/browser-compat-data features plausible pieces of the feature as a whole?
• Are any pieces missing?
• Are any of the listed features later additions, part of a distinct sub feature, or otherwise excludable?

[tidoust]

CSS Overflow Module Level 4 extends overflow to also apply to "replaced elements". That's still being worked upon, I think, and seems at risk.

Assuming this goes through, the definitions of the CSS properties would remain unchanged, the scope of their application would change. I'm wondering how we could represent that evolution once css-overflow-4 becomes the base level. Would we just update the URL to https://drafts.csswg.org/css-overflow-4/#propdef-overflow?

[ddbeck]

@tidoust I think the evolution would look something like this:

1. When the first implementer ships overflow for replaced elements, then I'd expect a new feature to appear in BCD for the relevant properties (e.g., css.properties.overflow.applies_to_replaced_elements).
2. At that point, we'd create a feature here (something like overflow-for-replaced-elements). We'd indicate it has some relationship to the overflow-shorthand (see Feature parent relationships #217 for more on that question).
3. The march of time happens and browsers eventually implement the feature.
4. A day comes when the overflow-for-replaced-elements feature is mature enough (and developers have ceased to distinguish it) that we merge the feature into the overflow-shorthand feature. (I imagine we'll have a guideline to cover this, something like a year after a feature has achieved some very high level of support.) See also: Consider handling feature identifier changes or aliases #91.

Does that feel like a practical approach? Or do you think we need to do some future-proofing today?

[tidoust]

Does that feel like a practical approach? Or do you think we need to do some future-proofing today?

That works! I wasn't clear that there could be a more "abstract" (in the sense of not being a 1-to-1 mapping to a property value) entry in BCD such as css.properties.overflow.applies_to_replaced_elements. That seems a good approach to me.