Allow colons in nmtokens #365
Conversation
The current definition of the `nmtoken` production doesn't match the XML's one. The same is true of `name`, but names are less of an issue. `nmtokens` are frequently used in LDML to define attribute values which are likely to be used a variant keys and option values.
|
It's worth noting that even without this change, it's still possible to use literals for values which don't match our |
| name = name-start *name-char ; based on https://www.w3.org/TR/xml/#NT-Name, but cannot start with U+003A COLON ":" | ||
| nmtoken = 1*name-char ; equal to https://www.w3.org/TR/xml/#NT-Nmtoken |
There was a problem hiding this comment.
Is the colon in {:functionName} syntax sufficiently valuable to justify deviation from exact match of the XML productions? If so, I question whether alignment with XML is actually well-motivated, and if not then perhaps it would make more sense to use characters excluded from Nmtoken:
!"#$%&'()*+,/;<=>?@[\]^`{|}~
[later commentary removed because it related to Nmtoken rather than Name]
The current definition of the
nmtokenproduction doesn't match the XML's one. The same is true ofname, but names are less of an issue.nmtokensare frequently used in LDML to define attribute values which are likely to be used a variant keys and option values.This was originally spotted by @mihnita in #364 (comment). I suggested fixing it in one of the two following manners:
This PR implements the first solution.