[css-pseudo] Should ::before / ::after really be part-like?

[emilio]

When @tabatkins introduced part-like pseudos in the spec (0f41712), it was added that:

Both ::before and ::after are part-like pseudo-elements; there is no restriction on what properties apply to them.

While it is true that there are no property restrictions on these pseudos, I'm not sure they should be part-like pseudo-elements.

part-like pseudo-elements should have other properties (like supporting further pseudos, or a bunch of pseudo-classes that ::before and ::after do not support).

So I'm not convinced ::before and ::after should be marked as such?

[emilio]

cc @dbaron

[dbaron]

I'm inclined to agree although I'm not sure yet. I was expecting to look into this as part of the "audit" I suggested in #10794.

[dbaron]

I think making ::before and :after part-like means that ::before::before should work. Supporting potentially infinite levels of ::before/::after seems like it might be problematic. (Though maybe not?)

[Loirooriol]

Note that ::before::marker works (well, browsers don't support the selector yet, but the ::before can originate a marker box).

[tabatkins]

Hm, I'm fine either way.

[josepharhar]

I thought that part-like means that the main/parent element of the pseudo-element has a UA shadowroot, and there is an element in that shadowroot which the pseudo-element points to, but based on this discussion I guess that's not true, thats just how its implemented in blink.

[kbabbitt]

I think making ::before and :after part-like means that ::before::before should work. Supporting potentially infinite levels of ::before/::after seems like it might be problematic. (Though maybe not?)

FWIW, the spec currently states this is invalid. https://drafts.csswg.org/selectors-4/#sub-pseudo-elements:

Unless the corresponding sub-pseudo-element is explicitly defined to exist in another specification, pseudo-element selectors are not valid when compounded to another pseudo-element selector. So, for example, ::before::before is an invalid selector, but ::before::marker is valid (in implementations that support the ::before::marker sub-pseudo-element).

[css-meeting-bot]

The CSS Working Group just discussed [css-pseudo] Should ::before / ::after really be part-like?, and agreed to the following:

• RESOLVED: ::before and ::after are not ::part-like but are tree-abiding

The full IRC log of that discussion

<dholbert> dbaron: this is a subtopic of previous one. emilio filed a specific issue about ::Before and ::after
<dholbert> dbaron: I agree with him, that ::before and after are not part-like
<kbabbitt> +1
<dholbert> fantasai: right, it surprised me to find that in the spec
<dholbert> emilio: sounds good
<andreubotella> +1
<dholbert> PROPOSED RESOLUTION: ::before and ::after are not ::part-like
<kbabbitt> q+
<flackr> +1
<Rossen6> ack flackr
<Rossen6> ack fantasai
<Rossen6> ack kbabbitt
<dholbert> kbabbitt: but they are tree-abiding?
<dholbert> emilio: right
<dholbert> PROPOSED RESOLUTION: ::before and ::after are not ::part-like but are tree-abiding
<dholbert> RESOLVED: ::before and ::after are not ::part-like but are tree-abiding