[css-conditional-3] Are there issues with !important in @supports? (Is the grammar wrong?)

[svgeesus]

@SimonSapin [wrote on www-style in 2013](

Le 30/04/2013 07:44, John Daggett a écrit :

I've updated the CSS3 Fonts draft based on Simon's feedback.

AFAICT, the point of <descriptor_declaration> is to not have the
production for !important. You could, like @supports, refer to
defined in the Core Grammar instead.
Does @supports specifically ban the production? I don't see
the language in there for that. It's a relatively minor point but I
added the production to be clear that '!important' has no function in
@font-face rules (there have been issues brought up in the past in
relation to this).

I just assumed that because @supports defines explicitly
by referring to Chapter 4:

declaration : property S* ':' S* value;

rather than Appendix G:

declaration
: property ':' S* expr prio?
;

But I was wrong. !important is allowed and has no effect. This is
specified in section 6, just before section 6.1:

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

For example, the following rule is valid:

@supports (display: flexbox !important) {
// ...
}

[svgeesus]

@SimonSapin a few questions:

1. Is this just about @supports or is it about !important in descriptors in general
2. Does the point you made by referring to CSS2.1 grammar still hold now that Conditional 3 uses the CSS Syntax module instead?
3. Are you saying that

• !important in descriptors is allowed, valuable, and we should test it
• !important in descriptors is allowed, unfortunately, and we change the grammar to disalow it
• something else?

[tabatkins]

Just to confirm, here's a testcase

Chrome and Firefox both show green (!important is allowed)

[css-meeting-bot]

The CSS Working Group just discussed Are there issues with !important in @supports? (Is the grammar wrong?).

RESOLVED: Closed - no change
RESOLVED: Publish a CRD of CSS Conditional L3

The full IRC log of that discussion

<gregwhitworth> Topic: Are there issues with !important in @supports? (Is the grammar wrong?)
<astearns> github: https://github.com//issues/5559
<TabAtkins> github: https://github.com//issues/5559
<gregwhitworth> TabAtkins: this is going to be similar to the first one
<gregwhitworth> TabAtkins: syntax currently allows a bang at the end, it has no impact on the result
<gregwhitworth> TabAtkins: implementations that I tested ignore it consistently. Should this be confirmed that we're fine with or not?
<gregwhitworth> TabAtkins: Simon found it unfortunate possibly in the past? I feel we're going to lean towards consistency
<gregwhitworth> florian: so you can include it?
<gregwhitworth> TabAtkins: yes Chrome/FF do, just ignore it
<gregwhitworth> florian: that's good, we may have something later that people can check for that
<gregwhitworth> Proposed Resolution: Closed - no change
<gregwhitworth> Resolved: Closed - no change
<bkardell_> WebKit also seems green in the test
<gregwhitworth> chris: we're at 0 open issues, that means we'll be able to publish a CR thing outside of privacy and i18n
<gregwhitworth> astearns: deadline?
<gregwhitworth> chris: no
<gregwhitworth> fantasai: we can publish a crd anytime we want
<gregwhitworth> astearns: are there other changes?
<gregwhitworth> chris: every edit since 2013
<gregwhitworth> astearns: on some that's a short list :P
<gregwhitworth> astearns: should we have a crd to help others
<gregwhitworth> fantasai: I think it's a good idea
<gregwhitworth> Proposed Resolution: Publish a CRD of CSS Conditional L3
<gregwhitworth> Resolved: Publish a CRD of CSS Conditional L3

[Loirooriol]

So is this expected?

CSS.supports("background: green !important"); // true
CSS.supports("background", "green !important"); // false

[fantasai]

@Loirooriol Mind filing that as a separate issue? Afaict the spec isn't particularly clear...

[svgeesus]

@fantasai Closed - no change so which edits?

[fantasai]

@svgeesus Idk, maybe I was looking at the wrong issue or something. :)