gh-87720 Fix additional header refolding-quoting edge case

[bitdancer]

In this case, the higher level syntactic unit fit on the remainder of the line in un-encoded format, so the encoding check never happened. To fix this, we look inside the current unit to see if has anything that was originally encoded.

I'm not really happy with the method name 'has_token_type', but I haven't thought of anything I like better. Ideas welcome.

• Issue: email: encoded headers lose their quoting when refolded #87720

[bitdancer]

@medmunds Do you feel like reviewing this? :)

[medmunds]

I'm not really happy with the method name 'has_token_type', but I haven't thought of anything I like better. Ideas welcome.

contains_token_type? (Though I'm not convinced that's any better; has_token_type makes sense to me too.)

[medmunds]

This fix LGTM. But I found another corner case while looking at it. Here's the failing test:

--- a/Lib/test/test_email/test__header_value_parser.py	(revision ebde4e897750963b82bcae785949ac11b3877640)
+++ b/Lib/test/test_email/test__header_value_parser.py	(date 1748289921461)
@@ -3101,6 +3101,8 @@
              '=?utf-8?q?a=2C?=\n'
              ' 123456789012345678901234567890123456\n'
              ' <[email protected]>\n'),
+            ('[email protected] (, but definitely not also sent to [email protected])',
+             '[email protected] (, but definitely not\n also sent to [email protected])\n'),
         ]
         for (to, folded) in cases:
             with self.subTest(to=to):

(This case belongs in a new test covering comment folding, but you'll get the point. Guessing I should open a new issue—it's more like the earlier quoted-string problem. I bring it up here in case it might help you spot other xtext variations needing attention.)

(Also, is this on Seth's radar yet?)