I haven't quite figured out what's happening here, so please excuse the vague title, but the output of KotlinLexer.get_tokendefs() certainly seems to be broken.

Reproducable script.

Run the following code (for example with uv run main.py or how ever you manage your dependencies):

# /// script
# dependencies = [
#   "pygments==2.19.2",
# ]
# ///

from pygments.lexers.jvm import KotlinLexer

print(KotlinLexer.get_tokendefs())

And it will print something that looks like something in binary format (screenshot here because GitHub won't accept invalid utf-8 bytes here):

As you can see up until the highlighted Token.Literal.Number everything looks fine but below are random looking bytes.

The corresponding line in the KotlinLexer would be the parsing of identifiers:

# Identifiers
(r'' + kt_id + r'((\?[^.])?)', Name) # additionally handle nullable types