Hello,
I am currently working on mailto notifications (so I am not entirely
off-topic;) and came accross a problem:
Are argument hnames in the query part of mailto URIs caseful?
RFC 2368 does not tell if the argument hnames resulting from splitting
the query part into pieces are caseful or not. Neither does
draft-duerst-mailto-bis-02. I did not get a response from asking the
authors.
In general, RFCs describing specific schemes must state if anything
is to be considered caseless. Nothing is said -> caseful.
Hnames are caseless -> caseless.
Who will get a message resulting from the URI below and how will it
look like?
mailto:?to=user1(_at_)example(_dot_)com&TO=user2(_at_)example(_dot_)com
Using caseful names, the message would go to user1(_at_)example(_dot_)com, but
listing both recipients in the header, in case the client trusts the
message. In case it does not and allows only safe headers, the message
would go to user1(_at_)example(_dot_)com and only listing that address in the
header.
Any hints towards references I overlooked will be appreciated.
Please respond by private mail and I will summarise.
Michael