Nick Leverton <[EMAIL PROTECTED]> wrote: > That's talking about local deliveries though (I've looked at the > code, and the only lowercase calls I can see are in reference to > hostnames and to local users). check_badrcptto is also invoked > when forwarding mail from relayclients to other hosts, and those > hosts may be case sensitive (rare though that is nowadays). > >From interest, I've just done an experiment with two qmail > hosts, and the first one preserves the case of the address > completely in RCPT TO when forwarding as a remote delivery to > the second host.
Yes, I would expect it to preserve the case. I'm not suggesting that the address should be modified, only that the comparison should be case-insensitive. Consider a case where the admin wants to block delivery to [EMAIL PROTECTED] How likely is it that blocking delivery to [EMAIL PROTECTED] as well is going to be a problem? I'd say the chance of that is so vanishingly small that it's not worth making the admin go through the trouble of entering JSmith and Jsmith and JSMITH in the 99.9999% of cases where case-sensitivity is not desired. And I don't see how that violates any RFCs. -- Keith C. Ivey <[EMAIL PROTECTED]> Washington, DC
