In <714EB841B8E8024D91E725F6E3F5F739513DC4(_at_)citadelex(_dot_)corporate>
"Dustin D. Trammell" <dtrammell(_at_)citadel(_dot_)com> writes:
wayne wrote:
I think that checking things token by token is just plain confusing.
Again, I think implementations of SPF MUST check for all syntax errors
including unknown mechanisms. It MAY check for things like missing
records in include:, recursion depth limits, etc.
[Justification for left-to-right *EVALUATION* snipped]
It appears that both you and Meng somehow are under the impression
that I want to get rid of the left-to-right *evaluation* order.
Nothing could be further from the truth.
What I am talking about is only checking for valid syntax on a
token-by-token basis. What I don't like is an SPF record like
"v=spf1 mx <random garbage>" returning "pass" some of the time and
"unknown" other times. This includes things like typos, invalid
syntax, unknown mechanisms, etc.
My reading of the SPF spec is that whether an implementation decides
to check the syntax of the complete record or not is unspecified.
Some implementations MAY return "unknown" anytime there is a
syntax error anywhere in the record, while others MAY return "unknown"
only some of the time.
The SPF spec as it stands now has vague terms like "encountered" and
"while processing". It doesn't say whether, when searching for an
exp= string, finding an unknown modifier means that it has
"encountered", or if an "encounter" only happens during a the
evaluations.
I do not think that this uncertainty is a good thing. When we are
talking about authorization issues, fuzziness and undefined behavior
just gives me the creeps.
Again, I think implementations of SPF MUST check for all syntax
errors.
I think that the SPF spec must be tightened up in this regard. If
unknown mechanisms only trigger an "unknown" status while the
evaluation of the status has not been determined, then it should say
so.
Personally, I also think that allowing some implementations to accept
mechanisms that are not in the spec and do different things is a bad
idea. If unknown mechanisms are allowed at all, I think something
like Mark Shewmaker's suggestion in the "Extensibility" thread is a
good idea.
-wayne
-------
Sender Permitted From: http://spf.pobox.com/
Archives at http://archives.listbox.com/spf-discuss/current/
Latest draft at http://spf.pobox.com/draft-mengwong-spf-02.9.5.txt
Wiki: http://spfwiki.infinitepenguins.net/pmwiki.php/SenderPermittedFrom/
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to
http://v2.listbox.com/member/?listname(_at_)©#«Mo\¯HÝÜîU;±¤Ö¤Íµø?¡