On 21 Mar 2000, at 14:41, Philip Hazel wrote:
> On Tue, 21 Mar 2000, Kai Henningsen wrote:
>
> > Well, the director in question will have some test of validity for
> > the truncated localpart. When there are several possible ways to
> > truncate, and only one gives a hit, "correct" is pretty clear, if not
> > always practical.
>
> No, that is precisely what does NOT happen. The director is not run if
> the suffix does not match.
That's not the interesting case. The only interesting case is if more
than one possible suffix matches.
> That's what I said before. It is purely a
> textual matching process. The validity of what is left as the local part
> is not checkable at this time.
I did say it's not always practical.
> > And, having done "take the longest match from a list" code myself,
> > it's also pretty easy to do. Hmm. Would you accept a patch that made
> > exim do so?
>
> Is it worth it? How often does this problem arise in practice? [As it
I have no idea, but when it does happen, it can be hard to figure out
because, IMNSHO, the behaviour is quite counter-intuitive.
> happens, I have re-organized that code recently, for other reasons, so a
> straight patch won't work in the next release.]
There would be several possible solutions to this.
> (I hope you aren't now also going to suggest a change to regular
> expression matching in PCRE to go for the longest possible match... :-)
PCRE mirrors Perl re behaviour, no? So it also has * vs. *?
behaviour. And anyway, that doesn't seem to have any relevance
whatsoever that I can detect.
Regards - Kai Henningsen
--
http://www.cats.ms
Spuentrup CTI Fon: +49 700 CALL CATS (=22 55 22 87)
Windbreede 12 Fax: +49 251 322312 99
D-48157 Muenster Mob: +49 161 322312 1
Germany GSM: +49 171 7755060
