Open cdoublev opened 4 weeks ago
Do you think they could escape this segregation and move to the global
values
?
Yes, that would seem a good thing to do.
There should be a way to tell that this is an extension of a base definition. For properties, we have newValues
. Perhaps we could record the syntax in a newValue
property here, although I note that the semantics are slightly different: for newValues
, these are additional values, whereas the extension here seems to always overwrite the base definition.
I presume this would require parsing the link before extracting the text content, and carry it forward to set
rule.href
, which does not seem ideal.
An additional complication is that these <pre>
blocks may contain more than one production rule :( That is what the split on reSplitLines
a couple of lines after the one you referenced attempts to take care of:
https://github.com/w3c/reffy/blob/30ab8e13c8f5b645cb2f1a4a91237a752fb5ae17/src/browserlib/extract-cssdfn.mjs#L568-L574
I also do not know if the extracted definition
href
would need to point to the other spec.
For property extensions, we simply do not put capture any href
. I would suggest to do the same.
When a spec extends a production rule whose LHS links to its definition in another spec,
reffy
reports it as a dangling production rule:<color-mix()>
<cross-fade()>
<keyframe-selector>
<radial-size>
Do you think they could escape this segregation and move to the global
values
?I presume this would require parsing the link before extracting the text content, and carry it forward to set
rule.href
, which does not seem ideal.I also do not know if the extracted definition
href
would need to point to the other spec.