On 06/08/2012 08:12 PM, Murray S. Kucherawy wrote:
> On Fri, 8 Jun 2012, Ken Murchison wrote:
>> As an alternative to changing the dkim_header() behavior, here is a
>> patch that reverses the order of the headers when using dkim_chunk()
>> so that they get processed last-to-first.
>
> Have you found that this is necessary? If it is, there's probably a bug
> with dkim_chunk(), because it works fine through the more traditional
> interface. But really, dkim_chunk() just breaks the header and body
> apart and then feeds them in-order to the other parts of the API, so
> this should not be a problem in either case.
My patch shouldn't be necessary if the library already does last-to-first.
In testing, I had my verification code using dkim_header() in reverse
order and it was failing to verify stuff created by dkim_chunk().
Obviously this was me misusing the library.
--
Kenneth Murchison
Principal Systems Software Engineer
Carnegie Mellon University
Received on Sat Jun 09 2012 - 15:38:30 PST