UnicodeIDStart::
any Unicode code point with the Unicode property “ID_Start”
UnicodeIDContinue::
any Unicode code point with the Unicode property “ID_Continue”

https://tc39.github.io/ecma262/#prod-UnicodeIDStart
@mathiasbynens ah, thanks! 鉂わ笍
I'll try to get to this today, I have discussed a possible fix with @rbuckton that should be fairly easy to implement.
Involving yet another @mathiasbynens's library, of course (in this case, "he"). Idea is in cases where we slice the source text to pass to grammarkdown we first decode HTML entities (other code paths where we don't slice the source text use innerHTML and so get decoded for us)
Fixed bterlson/ecmarkup#117.
Most helpful comment
Fixed bterlson/ecmarkup#117.