Robert Burrell Donkin ha scritto:
> On Sun, Dec 14, 2008 at 2:49 PM, Norman Maurer <[email protected]> wrote:
>> Should I just go ahead an commit it ( and open a jira issue) ?
> 
> if we follow stefano's design reasoning then really the VUT code needs
> to be moved from LocalDelivery into a separate mailet

A little history.

1) LocalDelivery was a single mailet including user aliasing code.

2) I splitted LocalDelivery into 2 separate mailets:
   UsersRepositoryAliasingForwarding
   ToMultiRepository
   I did that because the way aliasing worked in past james was a lot
tricky and this was a first step in modularizing it.

3) Norman and I worked on a VUT service because we wanted to let access
to the same "table" from both mailets and smtp fastfail stuff. I still
consider this an *hack*, but there is no easy way to support per
recipient fastfail unless we don't run some of the mailets during the
smtp protocol. I tried proposing changes to the mailet so that they was
more SMTP aware but others never liked the idea.


Also 2+3 happened at the end of 2006 when we (Norman and I) was working
hard to try to release a new version of james with new features.
That features required some care to be completed but we suddenly stopped
because other people didn't want a release and we gave up.

I really sincerely don't remember anything else about the status of what
was committed and what was not there. It's 2 years ago.


"my" design reasoning is simply derived from what was the design when I
approached james and what was written in the stone by people working on
this before me.

Maybe I didn't understand what kind of API you want to change, so maybe
it's better for me to wait for a concrete proposal and then judge/critic
it. IMHO the problem here is not the code or a VUT API issue, this is a
Mailet API issue.

Stefano

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to