DAVE CROCKER: do are you watching this list? this needs to happen on
yours too
So I just took a look around and they still haven't released the new version
of mailman that doesn't strip DKIM signatures. In the mean time you can
add:
try:
del GLOBAL_PIPELINE [GLOBAL_PIPELINE.index ('CleanseDKIM')]
except:
pass
# This is the pipeline which messages sent to the -owner address go through
try:
del GLOBAL_PIPELINE [OWNER_PIPELINE.index ('CleanseDKIM')]
except:
pass
into your /etc/mm_cfg.py mailman init script.
Mike
Michael Thomas wrote:
Greg Hewgill wrote:
On Mon, Feb 18, 2008 at 04:04:04PM -0800, Murray S. Kucherawy wrote:
Something ate the signature, because it arrived here unsigned.
Indeed, that seems unfortunate. I note that I am not currently folding
the DKIM-Signature line to fit within any particular margins, so
perhaps something ate it for that reason.
What kind of facilities are there available for DKIM interop testing?
http://testing.dkim.org seems to contain pre-RFC4871 code and content.
The reflector at testing.dkim.org should be using rfc4871 as of about a
month
ago. Some of the other material is out of date on there.
If this is to do with python and/or mailman, here is what I have observed:
1) A version of mailman a while ago was gratuitously stripping DKIM
signatures. They claimed that they were going to back that fix out...
I haven't verified that they have actually done and released that
2) There's something bizarre in the python 2822 digestion routines
which gratuitously reflows headers, including DKIM headers wrapping
in the middle of words. This causes signatures to break, and is
extremely
annoying. A bug should be filed against that library.
Mike
_______________________________________________
dkim-dev mailing list
dkim-dev(_at_)mipassoc(_dot_)org
http://mipassoc.org/mailman/listinfo/dkim-dev
_______________________________________________
dkim-dev mailing list
dkim-dev(_at_)mipassoc(_dot_)org
http://mipassoc.org/mailman/listinfo/dkim-dev