[css-contain-3] Should style() queries allow !important flag?

[lilles]

According to the current draft[1], the syntax for style features are the same as for declarations which means !important is allowed as part of the feature. I don't think that would have an effect, but would it still make sense to change the syntax to not allow it?

[1] https://drafts.csswg.org/css-contain-3/#style-container

[tabatkins]

We should disallow, yeah.

[frivoal]

In the case of @supports,

A trailing !important on a declaration being tested is allowed, though it won’t change the validity of the declaration.

I suspect we may want to do the same here: allow the !important syntactically, but ignore it. That way, we're consistent about the allowed syntaxes in the places that let you give a declaration for the sake of checking something rather than for declaring something. That said, if people feel strongly about disallowing it, that seems ok too.

[fantasai]

Disagree in the case of @supports. You're doing a syntax "is it supported" check there, and !important is arguably part of it.

[css-meeting-bot]

The CSS Working Group just discussed [css-contain-3] Should style() queries allow !important flag?.

The full IRC log of that discussion

<astearns> topic: [css-contain-3] Should style() queries allow !important flag?
<astearns> github: https://github.com//issues/7413
<argyle> astearns: who's gonna take this one?
<argyle> florian: the syntax for style container queries, lets you specify a property declaration representing the computed value, and grammar wise, a prop dec can include !important. but that has no effect on th computed value
<argyle> florian: q is, shouuld we ban writing !Important or should be keep it even tho it has no effect?
<astearns> ack fantasai
<emilio> q+
<argyle> fantasai: i think we should disallow this becaues it represents a computed value, it has no concept of where it exists in the cascade
<florian> q+
<argyle> fantasai: i dont feel the same about @supports rules which have a siimilar sytnax, in that case, you're asking is this syntax supported and then it's fine to have !important be part of that. but if add another !something later, it'd return null
<astearns> ack emilio
<argyle> emilio: i think i disagree with fantasai because even tho it does rep the computed value, what you write is a specified value. you can query color: blue, and it's computed then compared, but
<argyle> emilio: it's a specified value right
<astearns> ack florian
<miriam> q+
<argyle> florian: i dont feel strongly, seems to me it's easier if we can copy/paste declarations from where they define and test them, to test. that might drag a !impportant along the way. if we allow it in @supports, could support it here as well. but i dont feel strongly
<astearns> agrees with the copy-paste argument
<astearns> ack miriam
<argyle> miriam: was thinking about resolving variables, custom properties in a query, but they won bring along !important. never mind, i'm fine either way
<argyle> florian: if you're using actual css var, your'e right, but if you're using server side processor that has something similar to variables, it might
<argyle> miriam: that's true
<argyle> astearns: we are out of time, is anyone else convinced byt he copy paste arg to not make a change to the sepc?
<argyle> astearns: or should we continue discussion int he issue?
<argyle> astearns: not going to get reslution to this, we will continue in the issue. bring it back to the agenda onde we have clarity
<masonf> Thank you Adam!

[astearns]

I think it is very likely that authors will copy-paste values (or have them set by build-time variables) that will contain !important and be surprised or not notice that their style() query became invalid. So I think we should continue to allow them.

[lilles]

FWIW, I'm fine with closing this as no change. I have changed our implementation to allow !important and adjusted the WPTs accordingly to match the current spec.