Gmail to "gmail apps" sync: ValueError: Backend could not find uid for message

chris coleman christocoleman at
Mon Jun 27 10:22:17 BST 2011

>On Sun, Jun 26, 2011 at 11:32 PM, chris coleman wrote:

>> Brandon, Thanks for the clarification.
>> So this issue will be resolved once imaplib2 properly digests the untagged
>> 2nd CAPABILITIES response after login to Gmail.
>> There is still this issue: you seem to be saying that Gmail sometimes (or
>> always) doesn't announce UIDPLUS in the 1st CAPABILITIES response.  This
>> implies SOME users are on Gmail servers that don't support UIDPLUS, so all
>> Gmail servers are obliged to include defer announcing UIDPLUS support until
>> the 2nd CAPABILITIES response.  So the same crash will happen if the user is
>> on one of those servers that do not support UIDPLUS.
>Ah, I see the confusion.
>UIDPLUS was rolled out incrementally, but its now on for all users,
>and has been for at least a year.  We could graduate it to being
>always on and being announced pre-login, we just haven't done it yet.

My opinion: graduating UIDPLUS to be always on and announced pre-login would definitely be appropriate.

@Sebastian: What's your thoughts on fixing imaplib2 or offlineimap, so they respond properly when an untagged CAPABILITIES response arrives at any time.  Just in case, one day a user gets an IMAP server that changes the capabilities. Imagine a cluster scenario, the 1st IMAP server goes down, the load balancer switches over to the 2nd IMAP server, which announces different capabilities.

>And I just logged in and checked our code, my guess at the possible
>failure scenario is wrong, I do believe that we should handle that
>search backup correctly.  I'll have to write a test or two to see if
>there's some other bug there.

Very cool.  It's great to have your input Brandon.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the OfflineIMAP-project mailing list