Re: [exim] completely disabling checks for specific options?

On 9/17/05 1:18 PM, "OpenMacNews" <OpenMacNews@???> wrote:

> hi fred,
>
>> I must be missing something. If the option is unset, there is no
>> host list to check. It sounds like you're suggesting a new set of
>> boolean options of the form "ignore_option_X", but why?
>
> hmmm .... when you put it that way ...
>
>> Are you thinking checking the state of such an option would be faster
>> than directly checking whether the corresponding host list option "X"
>> is unset?
>
> well stated question. and, i suppose for this one particular option, a valid
> point.
>
> hence the 'in general' qualification ...
>
> as i mull your point over a bit, i'd have to say the "real" reason this came
> up
> is that i'm actually trying to pay detailed attention to debug output while
> setting up a new server.
>
> got the book, got the spec, got the filter spec, got the wiki, got google ...
> got a margirita. i'm prepared!
>
> i was in the midst of setting up my acl_smtp_XXXs. i'd read that
> 'host_reject_connection' had been obsoleted, and was thrown for a minor loop
> while trying to ferret out why the check was still running, and whether it was
> causing me any sort of grief w.r.t. my acl conditions.
>
> the end-result is, of course, no, it wasn't.
>
> as i ramble on, i guess it's just that -- given what i'd read/understood -- i
> did not expect the option to appear in the debug output, and was simply
> confused by its presence. honestly, not that it was 'faster' ...
>
> that said, i still think there's some value in knowing that an option that you
> have NOT enabled and/or is "obsoleted/deprecated", can arguably cause some
> conflict with other things you're doing is, in fact, NOT running its checks --
> and potentially causing problems.
>
> either this makes some bit of sense, or its just that i'm suffering vestiges
> of
> "noob disease".

I think you asked a reasonable question.

You will find that Philip works quite hard (except across major divides like
Exim 3 to Exim 4 (and to a much lesser extent Exim 2 to Exim 3 <I don't
recall that we *had* to change anything for 2 to 3>, and perhaps 1 to 2) to
avoid breaking existing configurations with new versions. One still *could*
use host_reject_connection in Exim 4, so simply removing it at Exim 4 time
would at least have risked breaking workable configurations.

Philip has occasionally let "us" (here on the list) talk him into
breakage...usually to the regret of Philip and at least some of us here on
the list.

--John