In <00d601c4653e$e855c790$c803a8c0(_at_)tag> "Andrew G. Tereschenko"
<spf-discuss(_at_)spam(_dot_)24(_dot_)odessa(_dot_)ua> writes:
I agree with this.
Current RFC proposal has nothing about maximum SPF record length.
I'm unaware if it 65000 (TXT DNS limit), 1400 or 512.
I would like writer clarify length for a records.
As well there is nothing about DNS string compression.
I have sent Meng patches to the SPF spec to add some stuff about how
publishers who go beyond 512 byte UDP DSN limit will likely have their
SPF record ignored in many cases. So far, the patches haven't stuck,
but there is hope.
In
<6(_dot_)1(_dot_)1(_dot_)1(_dot_)2(_dot_)20040708180912(_dot_)1288b2f8(_at_)81(_dot_)255(_dot_)84(_dot_)73>
Len Conrad <LConrad(_at_)Go2France(_dot_)com> writes:
SPF will look pretty stupid and self-aborting if it can't come up with
a way to stay within 512 bytes.
I have done surveys of SPF records and posted the stats to both this
list and to the IETF MARID list. In practice, there isn't any problem
with SPF records fitting into a 512 UDP DNS packet. If I recall
correctly, 50% of all SPF records are under 35 bytes, and something
like 99% are under 200 bytes. (The maximum SPF record is somewhere
around 350-400 bytes, before UDP headers, DNS headers and other stuff
push the total size over 512 bytes.)
-wayne