[Bug]: Lexing indented annotation

[ShortDevelopment]

Issue

   @annotation

is parsed as Token.Text.Whitespace( ) + Token.Operator(@) + Token.Text(annotation)

Expected behavior

Token.Text.Whitespace( ) + Token.Name.Decorator(@annotation)

Potential solution

Remove the check for newline / whitespace (^[ \t]*)

openqasm-pygments/src/openqasm_pygments/qasm3.py

Lines 77 to 78 in d6d0e07

	(r"^[ \t]*#?pragma", token.Comment.Preproc, "pragma"),
	(r"^[ \t]*@\w+(\.\w+)*", token.Name.Decorator, "annotation"),

Note: Lexing of indented pragmas works but includes the whitespace in the Token.Comment.Preproc token

Pros

• Match official grammar
• Don't include whitespace in parsed token

Cons

• Annotations might be highlighted within a line (Breaking change)

[jakelishman]

Thanks - I agree that the whitespace shouldn't be included in the token.

The trick here is that those two objects must be the first non-whitespace entries on the newline, and they're invalid if not. Do you know off-hand if Pygments accepts a "non-matching group" syntax or a match-start/match-end notation?

If not: there's a trade-off to be made between false positives in the lexer and complexity, but I hope it doesn't come to that.

[ShortDevelopment]

The main problem is that indented annotations are recognized as whitespace + @ operator