Stuart D. Gathman wrote:
It doesn't solve the problem of (broken but common) DNS
servers that don't respond when queried for a type99 -
resulting in timeout.
That's an "implementation detail", something that has to
be hidden by pretending that it's no problem as good as
possible. For the first query it's unavoidable, so what
can be done about further queries ?
You mentioned that almost every SPF implementation will
have its own kind of a cache. Can it make up some "do
not try -q=spf" entries for the relevant domains ? Bye