[css-contain-3] Rename 'none' to 'normal'

[fantasai]

Since we decided elements with container-type: none still have style containment, should we rename none to normal?

[SebastianZ]

I guess that's actually a duplicate of #7202.

I'm for that, since we now actually have some kind containment as initial value.
I want to point out again, though, that we still have the issue with the other keywords not reflecting that they do not overwrite style containment. But that's probably worth it's own issue.

Sebastian

[SebastianZ]

I want to point out again, though, that we still have the issue with the other keywords not reflecting that they do not overwrite style containment. But that's probably worth it's own issue.

I've now created #7403 to discuss that.

Sebastian

[lilles]

If so, we also need to consider:

• Should 'none' still not be a valid container-name and reserved for later?
• Should 'none' be kept as a way to turn off all container-types including 'style'?
• Make 'normal' an invalid as a custom-ident

[anttijk]

I'm opposed to this change. If every element is is a style container (not "have style containment" which would be a web breaking change) then the concept of "style container" has no meaning and value 'none' remains the correct one.

[css-meeting-bot]

The CSS Working Group just discussed rename container-type: none to normal, and agreed to the following:

• RESOLVED: Change initial value of 'container-type' to "normal"

The full IRC log of that discussion

<TabAtkins> Topic: rename container-type: none to normal
<TabAtkins> fantasai: Since we made style containment the default on every element, the initial value of container-type is "none".
<TabAtkins> fantasai: WAs wondering if we wanted to name that as "normal", which would possibly let us shut off style containment later if needed
<TabAtkins> fantasai: No strong opinion.
<fantasai> TabAtkins: I agree that renaming to normal works
<TabAtkins> miriam: Agree and also don't ahve strong opinion
<TabAtkins> astearns: In the issue antti said they weren't opposed to the change
<TabAtkins> astearns: But I think they weren't considering the case where we add names to style containers
<fantasai> s/weren't/were/
<TabAtkins> astearns: So we could put into the issue that we're okay with considering it?
<TabAtkins> smfr: If there's more context, I can ping antti and get his response
<TabAtkins> astearns: My take is that it's not that container-type has nothing to do with style containers, it's just that everything is a style container, and the only thing container-type *is* doing for style-only containers is letting you set a named style container.
<TabAtkins> astearns: I think antti is thinking we've broken the connection to the property entirely
<TabAtkins> astearns: would be good to find out
<TabAtkins> astearns: Simon, do you want to ping antti and we resolve next week?
<TabAtkins> smfr: Believe he's on vacation so might not get a response by next week
<TabAtkins> astearns: Alternately we could resolve now and let them bring up an objection later if it's still there?
<TabAtkins> smfr: I think I'd prefer that this issue have enough context that antti can read it and be sure he udnerstands the issue
<TabAtkins> smfr: But with that, I think I'm okay with resolving and letting him bring up an issue
<TabAtkins> fantasai: The two reasons are: 1) because style is a type of container, saying "none" is disingenuous
<TabAtkins> fantasai: 2) we might want a way to turn off style container on an element, and using "none no-style" is somewhat awkward. "normal no-style" is more neutral
<TabAtkins> fantasai: We're also not allowed to use container-names of "none" and we should keep that restriction, and also exclude "normal".
<TabAtkins> astearns: So proposed resolution is to change container-type's initial value to "normal" ,but still resolve "none" as a future value
<TabAtkins> ??: How about 'auto' instead of 'normal'?
<TabAtkins> fantasai: Could work, but 'auto' usually implies more magic
<astearns> s/??/ntim/
<TabAtkins> astearns: I slightly prefer normal over auto, but no great argument either way
<TabAtkins> I'm lightly for "normal" for the reason fantasai gave
<TabAtkins> ntim: No strong preference, just a thought
<TabAtkins> astearns: So any objections to normal?
<TabAtkins> RESOLVED: Change initial value of 'container-type' to "normal"
<TabAtkins> fantasai: TAb suggested in IRC that we just blanket restrict none/normal/auto in custom idents and I propose we do that.
<TabAtkins> astearns: Makes sense, but do we have those allowed in the past that would cause compat issues?
<TabAtkins> fantasai: Possible but seems unlikely.
<TabAtkins> chrishtr: What does that mean?
<fantasai> TabAtkins: A custom-ident can never be keyword 'initial'; all global keywords are restricted.
<fantasai> TabAtkins: we'd just add this to the list of keywords
<fantasai> chrishtr: Could it break content?
<lea> +1
<fantasai> TabAtkins: if ppl are currently using those names, yes
<fantasai> TabAtkins: but they are very generic, so it seems unlikely
<fantasai> TabAtkins: but we can check
<fantasai> astearns: We should open a new issue

[dbaron]

The new issue mentioned above is #7431.