ned+dkim(_at_)mauve(_dot_)mrochek(_dot_)com wrote:
Mark Delany wrote:
I agree that 'simple' is the easiest - in the original DomainKeys
drafts there was a sample perl implementation that took 4 or 5 lines
of code. Unfortunately because sendmail milter has a bug (that SMI
have promised to fix) some of those who are bound by milter
implementations want to bias the specification to match up with the
limitation of milter.
I expect that most of us have moved well beyond that sort of
mono-culture mentality, and I know the SMI folk are *not* the ones
pushing for this constraint, but nonetheless, it seems to live on in
pockets.
I personally am not in favor of using that as show stopper even though I'm
impacted, but since we want to get things deployed as quickly as possible
a proposal that breaks milter's assumptions should certainly viewed in the
light that it will make adoption that much harder.
Sendmail milter is not the only interface with this problem. There is at
least one other that has the same problems as milter; I think it was
qmail but would have to look it up for certain.
And let's please not forget that even if this got fixed tomorrow the amount of
time it takes to significantly deploy new MTA versions is very long - far
longer than we can afford to wait.
On this front, sendmail 8.12 doesn't give milters a good way to control
where a
header is added either, so making any sort of position-dependent
assumptions of
DKIM-signatures is on pretty shaky grounds, which I think this proposal
assumes.
Mike
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html