Re: [Nmh-workers] Weird behavior with non-ascii code in headers

[Ken Hornstein]

>> Sure, that would have been bad.  But if instead the invalid character
>> had been replaced by a ?, everything would have worked fine.
>
>That doesn't remove the possibility, it just makes it very,
>very unlikely.  If we're going to change something, we might
>as well do it right.  And, we wouldn't have to advise users
>to not use "?" in aliases.

Um, wait ... I'm confused.  How would it have _not_ completely removed
the problem?  In Valdis's case, the full header would be (after replacing
the "bad" character with a ?):

To: Mr Foo ? Bar <address@hidden>

The email address would still be fully qualified, and aliases would never
be checked.  Am I missing something?

--Ken