[EAI] UTF=x IMAP capabilities

[Barry Leiba <[e-mail]>]

At today's EAI meeting, I took an action item to start a short
discussion about the various UTF8=x capabilities in the IMAP doc (
http://tools.ietf.org/html/draft-ietf-eai-5738bis ).  We've agreed to
remove UTF8=USER and UTF8=APPEND, so I wanted to analyze the remaining
ones, and look at the implications.

UTF8=ACCEPT says that the server accepts UTF8 strings and allows
selection of mailboxes in UTF8 mode.  It's the basic "I support the
EAI UTF8 stuff" capability.  Good.

I presume the intent is that when a client that supports EAI talks to
a server that supports EAI, it will just use the UTF8 commands
throughout, yes?

Given that, I see little value in UTF8=ALL.  This says "all mailboxes
can be UTF8 SELECTed." What value is there to a client in knowing
that?  I suppose it can eliminate the need to use the UTF8 options in
the extended LIST command...  and there's no benefit at all if
extended LIST isn't supported.  I suggest that we also eliminate
UTF8=ALL.

That leaves UTF8=ONLY, which is reasonable to leave.  And so we have
three cases:

1.  No EAI support (no UTF8=x capabilities listed).

2.  EAI support, but also support of non-UTF8 versions (UTF8=ACCEPT).

3.  EAI support, and it's all there is (UTF8=ONLY).

So I suggest that we just spell it out that way, and make UTF8=ACCEPT
and UTF8=ONLY be mutually exclusive: there can be one, the other, or
neither, but never both.

Barry