In <427ADFFE(_dot_)5D94(_at_)xyzzy(_dot_)claranet(_dot_)de> Frank Ellermann
<nobody(_at_)xyzzy(_dot_)claranet(_dot_)de> writes:
Julian Mehnle wrote:
Not that I had a problem with the fact that you changed your
mind and applied the change, but why did you?
Maybe because we read it the same way wanting s/SHOULD/MUST/ ;-)
Yeah, actually I gave up after Frank hammered me also. I confess that
I'm more willing to change the parts of the spec that I wrote than
what others wrote. I'm pretty sure Meng wanted the SHOULD, so I'm
waiting for him to complain that it needs to be changed back. ;-)
People just shouldn't expect SPF to work on invalid input
data.
Invalid input data are characters above 127, or more than 25x
characters. For something used in mail you could restrict it
to ldh-string or similar, but not generally, e.g. the "_" case.
You mean invalid input for DNS queries?
If so, characters above 127 are perfectly valid and can be specified
using the \[...] notation. Each label must be less than 64
characters, and a zero length label must be at the far right (root).
To the best of my knowledge, those are the only two ways you can have
a "malformed domain name".
Even "." can be put inside a label using the \[...] notation, so
"foo.bar.com" is ambiguous and could be anything from two to four
labels (remember to include the zero length label at the far right).
This kind of thing drives the DNS gurus batty because it causes so
much confusion.
-wayne