W B Hacker wrote:
>>> Inserting one line of code:
>>>
>>> sender_host_name = sender_host_address;
>>>
>>> ... to utilize existing variables might be the mark of a Hacker (hey it
>>> IS my 'real name')
>> It wouldn't be compatible with peoples configs. You'd need to have a new
>> variable. Something like $sender_host_ptr. However, why bother writing
>> new code and adding further config options when the capability already
>> exists in dnsdb. Feels like bloat to me.
>
> One line with "<variable> = <value>" isn't a whole lot of 'new code', actually.
It's more code than removing a '#' from a makefile. It's also more
documentation and more stuff for people to learn. It's feature bloat.
>>> ..But is 'lighter' than having to build dnsdb just to see if a PTR RR
>>> 'exists'.
>
> I'm still waiting for someone to weigh-in with something dnsdb does that has a
> bit more utility - or at least is harder to substitute for.
1.) A plain PTR lookup without the equivalent forward lookup.
2.) Checking whether or not an SPF record exists for a domain
3.) Setting $smtp_active_hostname to the PTR record for $interface_address
4.) Querying the TOR exit node list
5.) Querying an "emailbl" such as the one described here:
http://mail-archives.apache.org/mod_mbox/spamassassin-users/200904.mbox/%3C49F7A21B.6030203@khopis.com%3E
6.) Checking how many MX records a domain has
7.) Checking if a domain name has NS servers or not
8.) Checking if the NS servers for a domain are listed on some DNS based BL
9.) Automatically checking if the connecting host has connected to the
primary MX or secondary MX for the domain it is attempting to deliver
mail to.
There are literally hundreds of use cases, but I can't be bothered to
spend more than a minute thinking of examples because I know you'll just
go through them all pretending that there are alternative ways of doing
it whilst ignoring the real issue here which is, dnsdb is flexible to do
almost anything you can imagine with DNS, without having to add hundreds
of different configuration options to Exim to do the same thing.
>> By your definition of, "lighter," yes.
>>
>
> Well.. the Exim 4.69-3 binary here is about 740KBytes -DWITH_PGSQL.
*Nobody* cares about the 4KB increase in file size. Not even you. You're
just clinging on to this argument because it's one of the few that you have.
>> It seems that everybody so far has disagreed with you.
>
> ACK. The *ratio* looks overwhelming.
>
> But does that mean we have now heard from all five who want dnsdb in the
> default? Or was it four?
>
> Is the tail wagging the dog?
>
> Or do the majority simply have simpler ways of getting the job done and not
> really care?
I'd suggest that if anybody agreed with you, at least one person would
have said by now. On that note, I'm satisfied that the archives and the
wishlist item show there is support for this idea, and I'm not going to
discuss this any further with you as it's quite clearly a pointless
exercise.
>> I'd suggest you
>> add your negative vote to the whishlist item, but that doesn't look
>> possible unfortunately.
>
> Done. Some time ago.
>
> BTW - personally it affects me not.
>
> I've shifted to use of a customized build anyway.
So have I. That's irrelevant to the discussion. I'm arguing for the
change for the benefit of the Exim community, you're just arguing about
your own personal systems and personal preferences.
--
Mike Cardwell
(
https://secure.grepular.com/) (
http://perlcv.com/)
