On 02/03/2021 18.28, John R Levine via mailop wrote: > On Tue, 2 Mar 2021, Otto J. Makela wrote: >> I read this as meaning most implementations will let you only have >> two NOERRORs, and then it's game over. As I said, I doubt SPF was >> intended to cause this side effect. > > Hm, missed that, it does seem wrong. > > On the other hand, if you're going to support IPv6, it seems to me > that it you put host names in your SPF record, those names should > have both A and AAAA records. As other people have pointed out, > using the IP addresses is often a better idea anyway. In this case, we just host a server for this customer domain, and generate emails for them from it. SPF has been set appropriately. Because the customer only has IPv4 email servers, WE cannot send out email through our standard multi-protocol support email servers.
If the other email services in the SPF at some point go IPv6 this might change, but it's a bit strange we have to account for this. This same effect also precludes any other servers in the SPF from ever wanting to go to IPv6. Unintended consequences, as I said. I'm kinda disappointed that the IPv4/IPv6 "never the twain shall meet" divide is still so strong, just treating names as symbols to be resolved (by whatever protocol) would have made so much more sense. -- /* * * Otto J. Makela <o...@iki.fi> * * * * * * * * * */ /* Phone: +358 40 765 5772, ICBM: N 60 10' E 24 55' */ /* Mail: Mechelininkatu 26 B 27, FI-00100 Helsinki */ /* * * Computers Rule 01001111 01001011 * * * * * * */ _______________________________________________ mailop mailing list mailop@mailop.org https://list.mailop.org/listinfo/mailop
