In <5AFF355A-1E02-11D9-A608-000393A56BB6(_at_)glyphic(_dot_)com> Mark Lentczner
<markl(_at_)glyphic(_dot_)com> writes:
Now, it has been proposed that it would better return PermError at the
start of processing if any mechanism is unrecognized. I actually
agree with this logic. However, the draft documents the SPF v1 common
understanding, and the above logic has been in SPF drafts for ages.
The %{h} macro variable has also been in the SPFv1 drafts for ages,
and yet it was removed.
In *theory*, changing the logic for unknown mechanisms is a much
larger change than the %{h} macro variable. However, having searched
through actually deployed SPF records, I found no one that correctly
used an unknown mechanism (I don't consider "ipv4" to be correctly
used), but I have found people who use %{h}.
So, in *practice*, cleaning up the unknown mechanism is a much less of
an incompatible change than removing the %{h} macro variable.
This is all stuff that I've told Mark/Meng before, but on IRC. I am
repeating it here not because I think Mark/Meng will pay any more
attention this time, but because not every is on the #spf IRC channel.
-wayne