era eriksson wrote:
Any way of adjusting your
rule to work with ver 3.10? Will dropping the backslash in $\MATCH work?
Most often it will. There is no way you can be sure.
It does. And the rule caught two spams for me in its very first day of
operation, today :)
For instance, the
following well-known address:
*(_at_)qz(_dot_)to
(yes, that is a valid address)
I gasped when I saw it first, some time ago, but now I know :) I only
wonder what _system_ it is that allows an asterisk for a mailbox name (user
name?) Never mind...
doesn't match itself when interpreted
as a regex. $\MATCH would give you \*(_at_)qz\(_dot_)to -- the first backslash
is
really significant here, while the second will only prevent accidental
matches on eg. qzoto or whatever. (You are somewhat unlikely to see
addresses like that on most of the Internet, but one rather common
case which also breaks without $\ is user+mailbox(_at_)host(_dot_)com -- the
plus, when interpreted as a regular expression, means "one or more
r:s" but there is nothing there to match the literal plus.)
I understand that. I think the danger of misfiling mail in this way is
rather small, though, because the particular rule we're discussing will be
used primarily as spam-catcher, and I haven't seen spammers using 'plussed'
addresses (they'd most likely be forged anyway, so why bother). And since
instead of autodeleting spam I have for some time been saving it to parse
and bitc^H^H^H^Hcomplain, I'll be ok.
Thanks, era! Thanks, Jari, too :)
.marek
--
"This is all very interesting, and I daresay you already
see me frothing at the mouth in a fit; but no, I am not;
I am just winking happy thoughts into a little tiddle cup."
(Vladimir Nabokov)