[dkim-dev] Do you have an interoperable DKIM of the recent spec?

[Michael Thomas]

Mircea Purdea wrote:
> Dave Crocker wrote:
>> I'm trying to find out how many implementations of DKIM there are, 
>> based on any of the recent versions of the spec.  Doesn't have to be 
>> as late as the one issued out of the working group, but something close.
>>
>> I don't care about the status of the code (prototype, alpha, beta, 
>> released, hack, whatever), product plans, release schedules or 
>> anything else.
>>
>> All I want to find out is that you have some code that has 
>> interoperated with someone else's code.  I guess "interoperating" 
>> with a test suite will qualify, for my purposes.
>>
>> The purpose of this exercise is very simple:  For anyone who claims 
>> that the current spec can't be implemented, I want to feed back to 
>> them a number that says how many already have been done.
>
> I have a reflector set up at 'dktest' at 'exhalus.net', which signs 
> with ietf-07 and verifies everything from allman-01.
>
> My implementation is normally bleeding edge, so to speak, i.e. it 
> always follows the latest released draft; this being so, I found that 
> most reflectors usually ignore my signatures. Anyway, you're all free 
> to try out the reflector and let me know if you have any interop issues.
Same with mine at dkim-test at mtcc.com. I'll accept signatures up v=1.0,
as will cisco's. The sendmail milter, I found, is getting tripped up on:

v=
q=dns/txt (just takes q=dns)
a=rsa-sha256 (it looks like it's a compile time option, but the 
reflector doesn't)

It works just fine on my verifying theirs though (modulo the NULL body
issue which everybody seems to have a problem with...)

       Mike