Html: Lowercasing of element type selectors

[gsnedders]

The spec currently says:

When comparing a CSS element type selector to the names of HTML elements in HTML documents, the CSS element type selector must first be converted to ASCII lowercase. The same selector when compared to other elements must be compared according to its original case. In both cases, the comparison is case-sensitive.

Selectors says:

The attribute name in an attribute selector is given as a CSS qualified name

As written, I believe this means that:

@namespace xhtml "http://www.w3.org/1999/xhtml";

XHTML|span {
  color: orange;
}

should be matched as xhtml|span.

See also: https://github.com/whatwg/html/issues/3372

[gsnedders]

https://github.com/web-platform-tests/wpt/pull/13738 has tests for this, as well as plenty of other tests for the HTML ns specificity of this.

[NOVALISTIC]

In your example, the type selector is not XHTML|span, but span. The former is a CSS qualified name; the spec says only to lowercase the type selector portion, not the entire qualified name. So the problem isn't in the type selector but in the namespace prefix XHTML.

Section 3.3 of css-namespaces-3 states that namespace prefixes are case-sensitive, so XHTML not matching the xhtml as declared in the @namespace rule seems correct to me.

[annevk]

Per https://drafts.csswg.org/selectors/#type-selector @gsnedders is correct. It seems like we should write something like "ASCII lowercase type selector's identifier".

Not sure what to write for attribute selectors though.

cc @tabatkins @fantasai

[gsnedders]

So https://drafts.csswg.org/css-namespaces-3/#css-qualified-name makes me think we want "the CSS element type selector's local name must first be converted to ASCII lowercase", though the definition of local name (and namespace prefix) isn't defined as a term in Namespaces.

[tabatkins]

Yeah, the namespace component is part of the type selector, per https://drafts.csswg.org/selectors/#type-nmsp.

This subsection should probably be slightly rewritten to use terms from DOM, yeah?

[NOVALISTIC]

I stand corrected, @annevk @tabatkins, so is the DOM text just confusing in that case?

[annevk]

@tabatkins the subsection in HTML or Selectors? I think Selectors could be clearer that a type selector is a struct that always has a name and sometimes a namespace (though perhaps it should be namespace prefix, it's a little unclear when resolution happens and what ends up in the data model). And then HTML could be updated to lowercase the type selector's name only.

[gsnedders]

@annevk Is that any less clear than cases where DOM uses qualified names? I don't think it really is?

[annevk]

@gsnedders yes, DOM has dfns for all these bits and is specific about using them.

[fantasai]

I think this should be fixed by using the wording @gsnedders proposes in https://github.com/whatwg/html/issues/4155#issuecomment-435347987 . CSS Namespaces is fairly old,, so it doesn't have DFNs the way newer specs do (and it wasn't anticipated that a host language should be able to manipulate part of a qualified name e.g. by lowercasing it prior to matching, it's a pretty weird layering violation). But @gsnedders' wording seems pretty clear to me.

[domenic]

Thanks! @gsnedders any chance we can convince you to turn that into a pull request? :)

[gsnedders]

@domenic I meant to do that several weeks ago. I, uh, probably should. :)