Python: CodeTokenizer splits quoted named-arg values on spaces (NamedArgBlockSyntaxError)

Description

CodeTokenizer splits named-argument quoted values at spaces, so a template like:

{{func key='hello world'}}

fails with NamedArgBlockSyntaxError: A NamedArgBlock starts with a name... — the token for the argument is cut to key='hello and parsing breaks. Any quoted named-arg value containing a space breaks the whole template.

Reproduction

from semantic_kernel.template_engine.code_tokenizer import CodeTokenizer
CodeTokenizer().tokenize("{{func key='hello world'}}")
# NamedArgBlockSyntaxError

Same with double quotes.

Expected behavior

quoted values are single tokens, spaces included, matching how the value grammar is documented (ValBlock is text surrounded by quotes).

Environment

semantic-kernel python main, Python 3.12

Start with CodeTokenizer in semantic_kernel/template_engine/code_tokenizer and run the provided tokenize reproduction for both single- and double-quoted values. Done means each quoted named-argument value containing spaces remains one token and the example templates tokenize without NamedArgBlockSyntaxError.