Le 16/09/2020 à 14:30, Derek Wilson a écrit : > https://github.com/systemd/systemd/issues/8967#issuecomment-391459667 > > Apparently the trailing dot "thing" never hits the wire? > > At some point if all DNS clients start doing ridiculous things, do we > worry that it will break server side operations? At what point do > clients abusing protocols start becoming a problem for systems (like > DNS) they misuse/abuse? > > I probably yelled too much in that thread for it to be effective > (sorry), but maybe someone here has a back channel to systemd-resolved > folks and can advocate for proper handling of trailing dots? > > Or maybe I'm bothered over nothing - in which case I'd love to be > convinced. > Oh my ..... Some parts of systemd are really brilliant. But the reasoning of L.Poettering in this case is completely brain-damaged and really stupid. All days I'm fighting with people because of using ambiguous non fqdn denomination to talk about an fqdn entry. All DNS entity are fqdn entity. Non fqdn is only a notation that should have never exist/created. His ping example is exactly what should had not existed. The guru vs non guru part is really pathetic. I'm not sufficiently fluent in English to help you to advocate in the ticket for proper trailing dot handling, I'm fully aligned with your point of view and argument. It is critical for conformant behavior of the resolver for application usage and proper on wire/internet behavior. I hope that some of dns-operation voice will help him to revise his position on the subject.
Emmanuel. _______________________________________________ dns-operations mailing list dns-operations@lists.dns-oarc.net https://lists.dns-oarc.net/mailman/listinfo/dns-operations
