[Imap-protocol] Cyrus and RFC5255

Bron Gondwana brong at fastmail.fm
Tue Nov 1 04:50:00 PDT 2011




On Monday, October 31, 2011 4:37 PM, "Mark Crispin" <mrc+imap at panda.com> wrote:

> On Tue, 1 Nov 2011, Bron Gondwana wrote:

> >> FWIW, I always considered the implementation of string searching to be

> >> implementation dependent. The only requirement was that simple

> >> case-independent substring is recognized as compliant. I never intended

> >> that a server be forbidden from implementing a smarter search (e.g.,

> >> fuzzy) that returns more matching messages.

> > Interesting. That makes me feel a bit more comfortable about ignoring the

> > full RFC 5051 i;unicode-casemap on searches, and only using it for sort.

>

> You can not do that and claim compliance with RFC 5255.


I'll probably make it a skanky toggle then.


> RFC 5255 explicitly requires that you apply i;unicode-casemap in searches

> as part of level 1 compliance.


The response when I mentioned it to our project manager was "it's often nice
not to worry about a vs å when searching - and have it find both".

But that's fuzzy-matching for you. Next thing it will be soundex searches,
and then we wind up with google style "magic".


> If you do not claim RFC 5255 compliance then there is no particular reason

> to implement i;unicode-casemap at all.

>

> However, as far as I know, however, Cyrus always implemented a variant of

> it from its inception; so you would actually be removing an aspect of it

> that was in there from the onset. That is probably not a good idea.


I'd rather not break our users' expectations too fast. People don't like
change much. Hence the toggle. Probably with the bogus default too. Maybe
"rfc5255_strict_search: no" or something. If you don't turn it on, then an
extra dicritical stripping pass and lowercasing pass gets run on the final
rfc5255 compatible data before searching it.

Bron.
--
Bron Gondwana
brong at fastmail.fm




More information about the Imap-protocol mailing list