Meng Weng Wong wrote:
[about %{t} timestamps]
it's not meant to be human readable; it's meant to give a
hint to the explanation CGI.
Okay, it's for explanations by URL, that makes sense. Maybe
mention it in an example. I'd still prefer the format...
| The syntax is: YYYYMMDDhhmmssZ
| Example: 19990609001326.34352Z
...as in RfC 3161, but if you're sure that the CGI knows all
about 20380119031407Z (RfC 2626), the timestamps should work.
I'll look into the new ABNF of protocol-01 later. In the new
SUBMITTER-03 bounces can still go to an unchecked MAIL FROM,
so that's not yet in a state where I could consider it to be
"compatible" with SPF classic.
PRA-00 is almost the same algorithm as before, so it does not
yet handle the common "missing Sender:" case. The "MAY" in
2476 8.1 still exists in draft-gellens-submit-bis-00.txt
The PRA algorithm without a solution for the "missing Sender:"
problem on the side of the PRA-aware recipient won't work.
Mail providers can't publish a SPF2.0/PRA sender policy, if
this doesn't work with legacy MUAs + 3rd party "MAYbe not" MSAs
of their users.
SYMPA lists: They have an Errors-To: matching the Return-Path:
But a special Errors-To: PRA-step for SYMPA makes no sense, a
general "missing Sender: derived from Return-Path:" PRA-step
would work for everybody.
Bye, Frank