[css-typed-om] Should StylePropertyMap.set be exposed to worklets?

[darrnshn]

From the original issue, it sounds like only .parse and .parseAll are not exposed to worklets. However, StylePropertyMap.set can take strings as arguments and will perform parsing. Is this problematic if it's called from a worklet?

[nainar]

No, currently StylePropertyMap.set isn't useful to be exposed to worklet. So let's restrict it to Window only for now.

[tabatkins]

We also need to restrict append() and update(), as they have the same problems. At that point the only function left for worklets is delete(), which makes me think we just shouldn't expose the interface at all, and only give them CSSStyleMapReadOnly. Thoughts?

[tabatkins]

(Also, hm, I think that if we don't expose the method, it'll get the default implementation from being maplike? If that's the case, then we do need to expose it, but throw immediately.)

[nainar]

To make sure I understand what you are saying correctly:

PaintWorklet exposed:

• StylePropertyMapReadOnly

Window expose:

• StylePropertyMap

If we do this we don't have the problem of the default implementation being available in PaintWorklet no?

[tabatkins]

Right, then it's just plain not editable at all, so no parsing required. ^_^

(And StylePropertyMapReadOnly is exposed to both Window and PaintWorklet, obvs.)

[nainar]

Fixed PR - #597 TAL?