[css-contain] contain:size needs to mention its effect on aspect-ratio

[tabatkins]

Contain 1, section 3.1 currently says:

Replaced elements must be treated as having an intrinsic width and height of 0.

However, it makes no mention of the element's aspect ratio. I believe the ideal behavior is to treat the element as not having an intrinsic aspect ratio.

(This was brought up in #5550)

Agenda+ because this is a Rec-level spec now. ^_^

[fantasai]

Fixing this will close #5550

[frivoal]

I agree on the substance of of the issue, and think we should fix this in L2.

However, I'm not sure we need to patch it up in the Level 1 spec as well, rather than just in the Level 2. As I see it, there's nothing wrong to be fixed in L1, it just doesn't define a particular thing, which will now be defined in the next level. We should absolutely go back and fix mistakes / contradictions in earlier levels, but having earlier specs be a tad vague, and later ones be more specific seems perfectly normal to me.

It's not particularly hard to fix l1 as well, so I won't object if y'all insist, but I just don't see the need.

[frivoal]

@fantasai said on IRC:

I think it's a mistake in L1 to have not mentioned it
That means it's not affected
It's a concept that's existed in much earlier specs
not something new
so by leaving it out, that implies it's not affected

Good point, I now agree. :)

[css-meeting-bot]

The CSS Working Group just discussed contain: size and aspect-ratio, and agreed to the following:

• RESOLVED: contain:size suppresses intrinsic aspect ratio

The full IRC log of that discussion

<fantasai> Topic: contain: size and aspect-ratio
<TabAtkins> github: https://github.com//issues/5585
<fantasai> florian: I think this was oversight in the original specification
<fantasai> florian: contain:size suppresses intrinsic size, mentions width/height, but forgot to state that it also suppresses intrinsic aspect ratio
<fantasai> florian: So should say so
<fantasai> florian: Note this is not about the explicit 'aspect-ratio' property
<fantasai> florian: but about the intrinsic one
<fremy> lgtm
<fantasai> lgtm
<fantasai> TabAtkins: seems obvious, but this is a REC so need WG approval
<dlibby_> q+
<fantasai> astearns: proposed that contain:size suppresses intrinsic aspect ratio
<astearns> ack dlibby_
<fantasai> dlibby_: Would it be possible that 0/0 gives us the right behavior for this aspect ratio?
<fantasai> TabAtkins: what do you mean by both zero?
<fantasai> fantasai: having an aspect ratio vs having an infinite aspect ratio is different
<fantasai> florian: We're not doing 0/0, we're doing "no aspect ratio"
<fantasai> cbiesinger: what if we have 'auto' in the aspect ratio in the property?
<fantasai> TabAtkins: you wouldn't ignore auto, but you would look up intrinsic aspect ratio and see that you have none
<florian> q?
<fantasai> RESOLVED: contain:size suppresses intrinsic aspect ratio
<TabAtkins> fantasai: We get to the be the guinia pigs for modifying a Rec
<TabAtkins> fantasai: Do we want to reoslve publishing an updated Rec that contains a candidate change?
<TabAtkins> chris: Doesn't it have to be published under a particular license?
<TabAtkins> fantasai: Only if you're adding new features, not fixing errors
<TabAtkins> florian: there is another change we're likely to do to the same level of this spec
<TabAtkins> florian: *after that*, sure, but let's resolve just once
<fantasai> TabAtkins: so no publication yet, but soon. happy to guinea pig
<fantasai> florian: another change in terminology is proposed
<fantasai> florian: and another one about the definition of contain:size being phrased sufficiently vaguely that mats didn't disagree with what we were trying to do , but wasn't sure what we were trying to do
<fantasai> florian: and we found some potential things we might want to change about how it affects grid tracks
<fantasai> florian: not on agenda today, but can discuss later