sid-milter-discuss

now "DNS timeout (opcode 99)"

[<fre...@va...>]

>
>It means DNS times out out while trying to do a query for record of
>type SPF (opcode 99).  You must be using my 0.2.8 patch with NH_SPF99
>defined.

Yes, now i remember! Accually i'm using all the extensions in your =
patch.
Have you withdrawn some other things from that patch aswell?
Thanks for the reminder!

>However, my extension to use opcode 99 records was withdrawn becausse
>it's broken.=20

What was the problem that made it broken?

/P

Re: now "DNS timeout (opcode 99)"

[Dick St.P. <stpeters@NetHeaven.com>]

> >It means DNS times out out while trying to do a query for record of
> >type SPF (opcode 99).  You must be using my 0.2.8 patch with NH_SPF99
> >defined.
> 
> Yes, now i remember! Accually i'm using all the extensions in your patch.
> Have you withdrawn some other things from that patch aswell?

NH_TYPOFIX and NH_INCLUDEFIX are incorporated into the 0.2.9 release.
It turned out that NH_INCLUDEFIX made NH_REDIRECTFIX unnecessary, so
it's unnecessary for 0.2.9.

NH_SPLITRECORDFIX was a workaround, not a real fix.  Eventually I
tracked the actual problem to a single missing line in the 0.2.8/0.2.9
code.  I'm expecting that to be fixed in the next release, and I'm
also hoping the NH_BACKFIX will be in that too.

That would leave only one 'fix', NH_IPFIX, which may never have been
necessary anyway.  I only saw a single instance of the problem it
addressed, and that might have been a side effect of one of the other
problems (or possibly even a side effect of one of my fixes!).  Of
course, it could also be that the reason I never saw another instance
is that I installed the fix.  Either way, the fix is benign even if it
may be unnecessary.

The situation for my extensions is more complicated, for reasons that
are not entirely technical and are still muddled.  All I'm prepared to
say is that they need revisiting.

> >However, my extension to use opcode 99 records was withdrawn becausse
> >it's broken. 
> 
> What was the problem that made it broken?

I don't remember exactly, but it involved breakage in cases where
there was no opcode 99 record.  Given the extreme rarity of opcode 99
records, I didn't feel any more-common-case breakage was tolerable.

--
Dick St.Peters, stpeters@NetHeaven.com