Noel J. Bergman wrote:
> Alternatively, we get the respository that represents that
> folder, and we
> store the mail without needing to make any other changes to
> the API.  Folder
> hierarchy is a naming issue.

Is the MailetContext intended to represent
1) A facade collecting together a few operations whose actual
implementations are spread across various components of the underlying MTA
OR
2) A simple bridge to the underlying MTA components???

Other than the experimental changes, it seems to be uniquely (1). Adding
methods that answer components is mixing in (2). Personally, I don't see
that we need to do this to resolve the sub-folder issue.

Supplementing the current...

void storeMail(MailAddress sender, MailAddress recipient, MimeMessage msg)

with...

void storeMail(MailAddress sender, javax.naming.Name destination,
MimeMessage msg)

...enables any storage mechanism whose locations can be specified using a
JNDI Name to be used.
During initialisation, a Mailet would discover what JNDI destinations are
supported by the current environment.

If we are heading towards JNDI anyway, why not?

-- Steve


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

Reply via email to