- Previous message: Mattia: "Re: Problem with postfix and mysql"
- Maybe in reply to: mouss: "Re: [OT] Quarantine notices"
- Next in thread: mouss: "Re: [OT] Quarantine notices"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
|
|
|
From: Matt Fretwell (no email)
Date: Sun May 15 2005 - 18:31:28 EDT
mouss wrote:
> > running for this single specific purpose? Obviously it would
> > be http *or* mail, (http taking precedence), not both methods in
> > parallel.
> the thing is: If it is not simple, it is not email.
True. Thats why I have acquiesced and taken the suggestions/corrections
given :) I always take good, pertinent advice and use it well,
(hopefully).
> so, don't even try to think of a model that is not simple for users.
> also, do not try to implement 70 schemes. find out one that is usable
> and implement it. all the rest is of only of interest to theoriticians.
> in the present case, it is no surprise that only imap and http are
> candidates. smtp is usable,
I already had the smtp type option working, (minus a few points that
Viktor pointed out, that are general, method indifferent, problems).
Thats why I wanted to know if it is worthwhile leaving in, even though I'm
now going down the mail notification with single URI pointing to,
(minimised), html index route.
Would leaving the mail only index option be useful for the paranoid type?
( The people who refuse to run anything unless it is chrooted and has been
torture tested for two years :) Or, do I just rip the smtp release mech
out, and pretend that bit never existed?
> but it requires a lot of work (one thing is how to make sure only the
> right user deletes/releases his message. this isn't that simple).
Even with http, I would assume that requires passwording the
directories/indexes, to be effectively secure?
Cheers Mouss,
Matt
|
|
|