Re: Problem with upgrade from 0.2.0.0 to 0.3.1

[Eli Barzilay]

On Jul  3, Dan Nelson wrote:
> In the last episode (Jul 03), Eli Barzilay said:
> > [...]
> 
> That block of code would only fire if flag_bucket is set, though,

Right.

> which means you have a -b or -B option on your commandline, which
> implies that you actually wanted the described behaviour :) Note
> that the inside block sends the message to the argument of the -b/-B
> option, not the original recipient, so I don't think it could be
> causing what you're seeing.

OK, there's definitely some mis-communication, and possibly some
documentation problem...  (I did read through the man page, through
other docs that came with the FC5 rpm, and through some of the web
page.)

The description of `-b' (and therefore `-B' too) is:

       -b spamaddress
              Redirects tagged spam to the specified email address.  [...]

So every email that is spam goes to the specified email.  That's
something I want: I want to hold spam in a local file to review from
time to time.  However, this file can get filled pretty quickly during
Spam storms, and I never bother with anything above 8 or so.  Then I
see this:

       -r nn  Reject  scanned  email  if it greater than or equal to nn.
              [..]

Sounds perfect.  *My* understanding was that with a standard 5.0 SA
threshold and `-r 10', anything below 5.0 gets delivered, things
between 5 and 10 are caught by the above -b spamaddress, and anything
beyond that is rejected with *no trace*.  Now, that can be defective
understanding on my part, but the following paragraph:

              For example, if  you  usually  use  procmail  to  redirect
              tagged  email into a separate folder just in case of false
              positives, you can use -r 15 and reject flagrant spam out-
              right while still receiving low-scoring messages.

seems to say just that.  What I obviously missed is the "use procmail"
part.  And finally I see that the original message in this thread was
about using `-u', not `-b'.

Obviously, I understood wrongly...

If I *now* understand correctly, the when `-r' is used, then there is
no difference between `-b' and `-B' (flag_bucket_only is not checked).
Also, `-b'/`-B' have nearly no effect when `-r' is used -- it looks
like the only difference is in how the message gets delivered to the
spambucket, which seems pointless.

So I suggest one of the following:

(1) Assuming that my last conclusion is correct, then there is no need
    to have the code that I removed, so
    1a. remove that code,
    1b. explain it in the `-r' description.

or

(2) If there is some difference (between `-b xxx -r nn' and `-b xxx')
    that I'm missing, then
    2a. clarify the `-r' description to say what the difference is
        explicitly,
    2b. add another flag for not passing messages to the spambucket --
        an obvious choice would be `-R' which will be "same as `-R'
        except ...blah-blah-blah...".

If this makes sense I will be happy to write and send patches (except
for #2a which is a difference I don't see now....).

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                  http://www.barzilay.org/                 Maze is Life!