Note why the lexers match digits with [0-9]

Records why \d must not come back, and why the negated classes keep it.

Author: FelippeRoza

ly/lex/html.py

@@ -95,6 +95,7 @@ class StringSQEnd(String, _token.StringEnd, _token.Leaver):
     rx = r"'"
 
 
+# [0-9], not \d: a numeric character reference is ASCII digits only
 class EntityRef(_token.Character):
     rx = r"\&(#[0-9]+|#[xX][0-9A-Fa-f]+|[A-Za-z_:][\w.:_-]*);"
 

ly/lex/lilypond.py

@@ -28,6 +28,9 @@
 from . import _token
 from . import Parser, FallthroughParser
 
+# digit patterns use [0-9]: \d also matches Unicode digits that are not legal
+# LilyPond. \d in a negated class excludes those too, so it stays
+
 # an identifier allowing letters and single hyphens in between
 re_identifier = r"[^\W\d_]+([_-][^\W\d_]+)*"
 

ly/lex/scheme.py

@@ -138,6 +138,7 @@ def test_match(cls, match):
         return match.group() in data.scheme_constants()
 
 
+# [0-9], not \d, which also matches Unicode digits that are not legal LilyPond
 class Number(_token.Item, _token.Numeric):
     rx = (r"("
           r"-?[0-9]+|"