Re: [whatwg/dom] Allow more characters in element/attribute names and prefixes (PR #1079)

@mathiasbynens commented on this pull request.



> +
+ <li><p>Return true.
+</ol>
+
+<p class=note>This concept is used to validate [=/element=] [=Element/local names=], when
+constructed by DOM APIs. The intention is to allow any name that is possible to construct using the
+HTML parser (the branch where the first [=code point=] is an [=ASCII alpha=]), plus some additional
+possibilities. For those additional possibilities, the ASCII range is restricted for historical
+reasons, but beyond ASCII anything is allowed.
+
+<div class=note>
+ <p>The following JavaScript-compatible regular expression is an implementation of the above
+ definition:
+
+ <pre class=lang-javascript>
+  /^(?:[A-Za-z][^\0\t\n\f\r\u0020/>]*)|(?:[:_\u0080-][A-Za-z-.:_\u0080-]*)$/u

I think JavaScript doesn’t support character class ranges without the end of the range being explicit, so I’d change both occurrences of `[\u0080-]` into `[\x80-\u{10FFFF}]`.

Is `[A-Za-z-.:_\u0080-]` meant to refer to step 4? In that case, it’s missing “ASCII digits”. Suggestion: `[A-Za-z0-9-.:_\x80-\u{10FFFF}]`.

- Valid namespace prefix: `[\0\t\n\f\r />]`
- Valid attribute local name: `[\0\t\n\f\r /=>]`
- Valid element local name, step 2.1: `[^\0\t\n\f\r />]`
- Valid element local name, step 3: `[:_\x80-\u{10FFFF}]`
- Valid element local name, step 4: `[A-Za-z0-9-.:_\x80-\u{10FFFF}]`

Piecing things together:

```js
/^(?:[A-Za-z][^\0\t\n\f\r\x20/>]*)|(?:[:_\x80-\u{10FFFF}][A-Za-z0-9-.:_\x80-\u{10FFFF}]*)$/u
```

-- 
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/dom/pull/1079#discussion_r892159848

You are receiving this because you are subscribed to this thread.

Message ID: <whatwg/dom/pull/1079/review/999449047@github.com>