-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Stuart D. Gathman wrote:
Empty domain-spec on exp modifier
[...]
Thus, domain-spec can never be empty. Syntax errors require a permerror
result, which conflicts with the instruct to "proceed as if no exp
modifier was given."
I don't think there is actually a conflict. The "If domain-spec is empty,
[...] then proceed as if no exp modifier was given" just never applies
(since a domain-spec can never be empty). This isn't a conflict but a
null statement, comparable to:
"n is a natural number. If n is larger than 2 and lower than 3, then
<foo>."
AKA
void blah(int n) {
if (n>2 && n<3)
foo();
}
An intelligent compiler would just optimize away the call to foo().
It's not a conflict. However, an erratum removing the redundant "If
domain-spec is empty," part is indeed in order.
Therefore, the grammar for exp should be:
explanation = "exp" "=" [ domain-spec ]
Rationale:
Unknown-modifier is defined as:
unknown-modifier = name "=" macro-string
This means that implementations not supporting the exp modifier would
not give a permerror. Unlike redirect, the exp modifier is not critical,
so giving permerror for an otherwise valid record is counter productive.
No, the grammar makes absolute sense as it is now. The argument to exp= is
a domain where the explanation TXT record should be looked up.
Saying "exp=" without a domain doesn't make any sense whatsoever and
should result in a PermError.
Following Stuart's rationale would mean that we also should not throw a
PermError on syntax errors if some mechanism matches before the faulty
syntax is reached -- after all, if we can get a result, any bad syntax in
later parts isn't critical, is it?
We deliberately decided against this notion quite some time ago.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFFrA7IwL7PKlBZWjsRAuhWAKDD5IZdql/9H0owxiSUukGolilx9gCfdpQ4
3lL+mvdKqikiLhcH1EJe1WQ=
=x0Hj
-----END PGP SIGNATURE-----
-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to http://v2.listbox.com/member/?list_id=735