unless anyone jumps in soon with objections or comments, i plan to
commit changes that switch to using a session id and events

On Nov 16, 2007 8:01 PM, Robert Burrell Donkin
<[EMAIL PROTECTED]> wrote:
> ATM the current design for MailboxListener (see 1) is method-oriented
> - each event is reported by calling a specific method. extension
> requires a new interface. i think that this API should be replaced by
> one accepting an event. switching to a messaging API loosens coupling
> and allows novel messages to be created without changing the
> MailboxAPI.
>
> opinions?
>
> ATM mailbox events work by UidChangeTracker noticing that some data
> that a mailbox has retrieved differs from it's cached value
>
> this approach causes problems for protocols such as IMAP that require
> events generated by the current session to be handled differently from
> events resulting from the actions of other sessions. IMO passing the
> listener throw is a poor choice of design and a workaround for deeper
> issues.
>
> IMO an explicit notion of MailboxAPI sessions is going to be required
> (setting aside the current interface names). when a mailbox performs
> an operation, it should notify listeners of that operation by posting
> a suitable event containing a session id.
>
> opinions?
>
> - robert
>
> 1. 
> http://svn.apache.org/repos/asf/james/server/trunk/core-library/src/main/java/org/apache/james/mailboxmanager/MailboxListener.java
> 2. 
> http://svn.apache.org/repos/asf/james/server/trunk/core-library/src/main/java/org/apache/james/mailboxmanager/tracking/UidChangeTracker.java
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to