On 2021-11-24 05:47 +0000, Claus Assmann wrote: > On Tue, Nov 23, 2021, Aaron Poffenberger wrote: > > > > > I know this is related to Message-ID... > > > > I'll back off this claim because I can't prove it beyond doubt. I've > > Of course you could. I explained how to do it, but (so far) you > haven't done that -- hence all those replies from people being > sceptical -- it would be really nice for those of us who are very > curious about this problem if you would provide real evidence / > data.
I haven't replied to the request until now because Kevin mentioned not being able to look at the issue much this week due to holidays so I took the time to try various options and conduct more research. What I can say now is I tried it multiple times, with plain mutt, no hostname setting, no changes to message_id_format, and regardless of configuration or email domain I've tried, every email has gone through. As noted elsewhere, other email admins have complained of the same reverse DNS problems at ATT, and Network Solutions has previously worked with ATT to resolve bogus reverse-DNS errors. To further bolster the case against my claim, I scanned through the sendmail code (prodigy.net servers reply with "Sendmail Inbound 8.15.2"). AFAICT, at most it validates that Message-ID exists and is not empty. Sure, there could be something else on their side, but I have no knowledge of it. I'm left with correlation. My change in muttrc coincided with some other change. There are better possible explanations, so I'm dropping the claim. Mea culpa. I should've have re-verified the issue and provided concrete evidence to use it as the justification for the patch, or abandoned it in favor of another use case. Cheers, --Aaron
