Mismatch bcd key for url() function

[skyclouds2001]

MDN URL

https://developer.mozilla.org/en-US/docs/Web/CSS/url_function

What specific section or headline is this issue about?

Browser compatibility

What information was incorrect, unhelpful, or incomplete?

the bcd key refers to the <url> css type, not the url() function

What did you expect to see?

fix the bcd key

Do you have any supporting links, references, or citations?

No response

Do you have anything more you want to share?

No response

MDN metadata

Page report details

• Folder: en-us/web/css/url_function
• MDN URL: https://developer.mozilla.org/en-US/docs/Web/CSS/url_function
• GitHub URL: https://github.com/mdn/content/blob/main/files/en-us/web/css/url_function/index.md
• Last commit: mdn/content@919d97a
• Document last modified: 2024-11-19T12:34:07.000Z

[hamishwillee]

@skyclouds2001 This is confusing, but I think it is correct. The <url> value (data type) takes a value that is the url() function - they are indivisible. So they have the same data.

@estelle Is the documentation of the <url> type in any way useful - I mean does this actually every appear in "code" or do you only ever see url()? If not, it might make sense to not document it and point to the URL function everywhere.

[skyclouds2001]

The <url> value (data type) takes a value that is the url() function - they are indivisible. So they have the same data.

currently it is true, as src() didn't support by any browsers, so <url> = <url()>; but the spec defines: <url> = <url()> | <src()>, so I think we should make <url> and url() have different BCD items

see also mdn/content#36839

Is the documentation of the <url> type in any way useful - I mean does this actually every appear in "code" or do you only ever see url()? If not, it might make sense to not document it and point to the URL function everywhere.

Personally I think it is better to still keep them as two pages at present

[hamishwillee]

@skyclouds2001 Perhaps! I am not sure because even if <url>=<src()> I still see not much value in <url> except as a syntax glue for the spec. We might not have value for it in the docs.

But anyhow, the question here is really about BCD in the first case. There is no url() BCD entry so we can't do anything with the docs unless the BCD team agree that it makes sense to do so. Transferring the issue.

BCD compat team, do we need a separate url() feature given the argument above. My feeling is perhaps this would be a subfeature of <url>, but I really don't know.

[estelle]

@estelle Is the documentation of the <url> type in any way useful - I mean does this actually every appear in "code" or do you only ever see url()? If not, it might make sense to not document it and point to the URL function everywhere.

What @skyclouds2001 said.
That page as written is useless, but the page is needed. we could do a much better job
For example "The value is either of the following:" and then there is only one value?!?!
We do document (or try to document) all the data types, and <url> is a data type, so must be documented. It is found in multiple property values, so needs its own page (there are some data types that are only on the property page for which they are a value)

[hamishwillee]

Great. So I guess the first thing is (still) to get a feature for the url() function and look at the docs for this as a separate problem.

[skyclouds2001]

For example "The value is either of the following:" and then there is only one value?!?!

I guess this is because previously there is also a src() function, but since it is not supported by any browsers at present, it was removed via mdn/content#36839.

See also the note just below that paragraph.

[skyclouds2001]

Great. So I guess the first thing is (still) to get a feature for the function and look at the docs for this as a separate problem.url()

yes, and I guess the css url() function should have the same bcd info as the css <url> type