Hector Santos writes:
Arnt Gulbrandsen wrote:
Hector Santos writes:
Excuse me if I am not following you, but the compares are
implementation specific. Like Finch pointed out, Sieve doesn't
care.
Sieve suggests treating localparts case insensitively. Any user may
follow the suggestion or explicitly choose something different from
the default.
This is not specific to any implementation: Any conforming sieve
processor behaves this way.
So what you are saying is that if we implement Sieve, to be Sieve
compliant, our implementation must provide user preferences in
regards to case?
Yes. And if the users copy the examples they'll find on the net, their
scripts will be case-insensitive. To quote the last script snippet I
saw:
if address :is :localpart "from" "zabbix" {
fileinto "Zabbix";
}
Users who know how to (ha ha) can write case-sensitive code by adding
:comparator "i;octet", so each script's case treatment is in principle
implementation-specific.
All this doesn't contravene RFC 5321. But it does suggest that the
overall email architecture doesn't treat localparts as completely
opaque.
Arnt