[Top] [All Lists]

Re: Review comments on draft-martin-managesieve-04.txt

2003-10-21 13:44:12

On Fri, Oct 17, 2003 at 09:08:41AM -0700, 
ned(_dot_)freed(_at_)mrochek(_dot_)com wrote:

It was recently mentioned on the imapext list that it is time to get
the managesieve protocol spec moving forward. So I decided to take a look
at the specification as it stands. My comments are as follows:


A few comments:

As I mentioned in another note, I think it would be worth making the
interface a little more generic: rather than being sieve-specific,
allow it to be used for whatever sort of scripting language is
officially available on the mail server.  There could be more than one
such language available.

I would also like to see support for address-space for the
authenticated user.  A user might have different controls for various
email addresses that are mapped to that user's account.  My own
account here receives mail for a number of different addresses and
domains, each with its own control area and associated filter(s).  I
doubt that I'm the only one that would like to be able to individually
manage filters for different address spaces under my control.  Thus
I'd like to see this interface take multiple address/filter spaces
into account.  Each address space might also (or might not, depending
on the server configuration) have its own namespace for filters, so
that "myscript" for user(_at_)example(_dot_)com could be a different file than
"myscript" for user(_at_)example(_dot_)net .

Perhaps an optional argument to PUTSCRIPT/GETSCRIPT/SETACTIVE/etc that
indicates the specific address space:

    PUTSCRIPT :address-space "user(_at_)example(_dot_)com" "myscript"
    PUTSCRIPT :address-space "user(_at_)another(_dot_)domain "myscript"
    SETACTIVE :address-space "user-extended(_at_)example(_dot_)com" "myscript"

HAVESPACE:  Is there really a point to this?  PUTSCRIPT will already
tell you if the operation fails due to quota.  There's no guarantee
that resource availability at HAVESPACE-time will be the same as the
availability at PUTSCRIPT-time.  Not only that, but I could imagine
that a for given implementation you would not be able to predict the
required space for a script (say, if PUTSCRIPT caused a compiled
version to be stored along with the source).  For a variety of
reasons, HAVESPACE doesn't necessarily tell you what would happen if
you try to upload a script.

  putscript "myscript" {nnn+}
    nnn bytes

I'm not really fond of that byte-counted argument imapism in this
context.  Am I the only one?  I'd much prefer to see some sort of
terminated content a la SMTP.