Re: [Aperture-devel] UIDNOTSTICKY and incremental IMAP crawling

[Christiaan F. <chr...@ad...>]

Sounds good. I believe Message-ID's are "reliable enough": 
astronomically small chance of a collision.


Antoni Mylka wrote:
> I'd like to resurrect this thread.
> 
> Christian Reuschling - the original reporter of this issue suggested
> that the IMAP crawler could check if the UIDs are persistent and if
> they aren't it could fall back to constructing URIS from the values of
> the Message-ID header present in each message. This is against the
> http://www.rfc-editor.org/rfc/rfc5092.txt but without UIDVALIDITY the
> UIDs are useless anyway because uris generated from them do not
> "identify" the mail.
> 
> I say we can do it.
> 
> any objections?
> 
> 2008/9/19 Antoni Mylka <ant...@gm...>:
>> To all ye imap gurus out there.
>>
>> I've been investigating the issue number: 1989505 with broken
>> incremental crawling. I've been unable to reproduce this bug for more
>> than 3 months. Only now did the DFKI folks create a test IMAP account
>> backed by the mh mail store - a source of the problems. I've found
>> some interesting things.
> 
> ...
>