Murray S. Kucherawy wrote:
> On Wed, 6 Jun 2012, Ken Murchison wrote:
>>>> 1. Is there a way to keep the library from requiring a From:
>>>> header, other that modifying the source? This obviously makes it
>>>> email-centric.
>>>
>>> The From: header is required as mentioned in Section 5.4 of RFC
>>> 6376. The only way around that is to modify the source.
>>
>> Right. I wasn't sure if there was some function call or back-door
>> that I wasn't seeing in the source.
>
> SM's correct, we don't have a way to do that since we're pure to DKIM
> specs and focused on email, so we force "from" into the list. It sounds
> like your patch might be useful for future extensions though, so I hope
> you'll contribute it back to us. :-)
I haven't fleshed anything out yet because I'm not sure if this would go
into something layered on top of libdoseta, or should be a patch to
libopendkim which switched between "email" and "ischedule" service modes.
> If you always want to add an "h=" instance of a name whether it's
> present or not, you want DKIM_OPTS_OVERSIGNHDRS. That's an experimental
> feature of the library you have to request at build time in v2.5.x by
> saying "--enable-oversign" to ./configure.
I will look into this. Is is still experimental in 2.6?
> DOSETA is idle because of lack of uptake. However, since you have a use
> case, I'll pass that on and maybe we can get some movement on it.
Has an API for libdoseta been fleshed out? Where do you see the break
between libdoseta and libopendkim? I assume that this is just a matter
of splitting the existing code and adding a little "glue".
--
Kenneth Murchison
Principal Systems Software Engineer
Carnegie Mellon University
Received on Thu Jun 07 2012 - 12:06:12 PST