[css-houdini-drafts] [css-typed-om] rule on inferring <length-percentage> is too vague

dbaron has just created a new issue for https://github.com/w3c/css-houdini-drafts:

== [css-typed-om] rule on inferring <length-percentage> is too vague ==
The section on [Numeric Value Typing](https://drafts.css-houdini.org/css-typed-om-1/#numeric-typing) says:

> Many specifications use ''[ &lt;length> | &lt;percentage> ]'' instead of ''&lt;length-percentage>'' in their grammar, and specify in prose that the &lt;length> and &lt;percentage> can be combined. For the purposes of matching, these cases should be treated as &lt;length-percentage>. Similarly for &lt;angle-percentage>, etc.

This doesn't seem clear enough to me to lead to interoperable implementations, since it doesn't make it clear what "specify in proces that ... can be combined" really means.  For example, for `line-height` I would think that &lt;length> and &lt;percentage> can be combined, but depending on what your definition of combined is, maybe &lt;number> and &lt;percentage> can't (since they're more semantically different).

I think this needs to be specified more clearly if it's going to lead to interoperability.  It might need to do that by listing the set of "old" grammar definitions to be patched, and requiring that future grammars get it right.

Please view or discuss this issue at https://github.com/w3c/css-houdini-drafts/issues/581 using your GitHub account

Received on Sunday, 28 January 2018 07:16:14 UTC