wayne wrote:
I just realized that I *do* remember one SPF implementation
that doesn't (didn't?) support the the format of the sc=
modifier that Roger suggest
It's not a question of what you or James or Roger _could_ do.
Of course you _can_ implement v=spf2 whenever it pleases you.
But you can't modify implementations in the wild. So if an
idea like _positional_ modifiers (sc= or protocol-03 style) is
not part of all running v=spf1 code, you cannot add it now to
v=spf1.
For exactly the same reasons why applying PRA on v=spf1 doesn't
work: It BREAKS future v=spf1 sender policies trying to use
positional modifiers. New implementations would evaluate it
as expected, old implementations ignore the unknown modifier,
and then results are incompatible for the very same policy.
Sometimes your mail arrives, sometimes it's deleted. Or other
effects. So please tell me, do all users of libspf2 interpret
sc=pra correctly ? If the answer is NO, because you have not
yet implemented sc=pra, then the effects are unforeseeable for
publishers of v=spf1 sender policies trying to use sc=pra.
if you are so knowledgeable about SPF implementations
I'm absolutely sure that libspf2 does not support sc=pra, this
idea is less than one day old.
I *really* don't think we should be trying to invent a new
SPF-classic.
v
So why are you claiming that sc=pra in v=spf1 policies is no
problem with existing implementations ? ^
And why do you think that I should read your or anybody else's
code before I can know this ? It's obvious that it can't work
in v=spf1, just like new mechanisms won't work in v=spf1. It
would break v=spf1 sender policies trying to use this. Or do
you have a magic wand, can you recall all running v=spf1 code
worldwide for an emergency fix ?
All you said was completely beside the point. What would your
code installed yesterday do with a new sender policy published
after adding sc= to v=spf1 like "v=spf1 sc=pra -all" ? Bye.