[Imap-protocol] Handle failures during STORE command execution

[Norman Maurer]

Hi Mark,

thank you for your response. You are right it  should not happen
frequently. But as we are implementing different storages its good to
know that we don't Need some kind of transactions.

Bye
Norman

2010/6/23, Mark Crispin <mrc+imap at panda.com>:

> On Wed, 23 Jun 2010, Norman Maurer wrote:

>> What should I return if the flag update fails for for message 3 only?

>> Do I still return the new flags for the rest of the messages (+just

>> return the old flags for message 3)  and response with ok or do I need

>> to rollback all the changes made during processing the command and

>> response with an error ?

>

> The former (return new flags for the changed messages, old flags for

> message 3).

>

> It should be extremely rare that an operation succeeds for some messages

> but not for others.  If you can abolish that possibility, so much the

> better.

>

> Normally, if a STORE fails to do something, it is because the mailbox is

> in a readonly state and thus no messages are changed.  But note the

> possibility of session-only flags with public mailboxes.

>

> With all that said, the entire reason for STORE returning untagged FETCH

> responses with updated flags is to report what actually happened.

>

> -- Mark --

>

> http://panda.com/mrc

> Democracy is two wolves and a sheep deciding what to eat for lunch.

> Liberty is a well-armed sheep contesting the vote.

>