On Mon, 2004-02-23 at 18:35 +0100, Daniel Roethlisberger wrote:
> > What little patience I originally had for the fundamentally broken
> > concepts of SPF and SRS has been reduced significantly over the last
> > day or so. TBH I think I'm going to turn off my implementation and
> > simply declare my forwarding machines to be SPF-incompatible, like 98%
> > of the rest of the world.
>
> As broken as they may be -- ultimately, it is *my* problem if messages
> forwarded through my box will not get delivered because of the receiving
> MTA using SPF or equivalent homegrown origin filters, as are currently
> in use by some largish providers (GMX.net, for instance).
That's a decision which is yours to make. It coincides with my thinking
a few days ago.
Now, however, I am inclined to lump those who reject my mail due to
SPF-type checks along with those whose mail _I_ reject due to their
rejection of MAIL FROM:<> or RCPT TO:<postmaster> in callouts -- in the
'don't care' pile. They're evidently not thinking straight; if at all.
I do appreciate that you may not have that luxury.
> It does not
> matter who's fault it technically is; users want their mail delivered
> without having to listen to technical excuses. I know no better solution
> than rewriting the envelope sender, as I described here [1], but if
> somebody has a better approach, I'd be delighted to hear it.
Your solution suffices to work around the brokenness introduced by SPF,
and is basically the same as what I have implemented; thanks for the
excellent documentation. The 'shortcut' scheme advocated elsewhere is
the one to which I refer as fundamentally broken.
> [1] http://www.roe.ch/spam/return-path-rewriting.xml
--
dwmw2
