Julian Mehnle wrote:
> It's obviously unspecified (as unfortunate as it may be)
The current state of the art wrt 2821bis is that it will
adopt the 2822 concept, one or more dot separated labels.
Which is BTW what I originally asked in November 2004:
Is the "1" in subdomain 1*( "." subdomain ) intentional
or a typo. It clearly was intentional, but the author
now changed his mind after checking out the same ideas
as we over the years: optional trailing dot everywhere
but mandatory for TLDs, no trailing dot anywhere unless
it's a TLD, did I miss something ?
So now it will be no trailing nowhere (as is in 2821 and
2822), but allowing all FQDNs. Figuring out if a single
label is garbage, something local, or a FQDN is left as
an exercise for implementors.
We've to wait for the next draft what the prose exactly
will be.
> the test suite should not require any specific behavior.
Well, so far nobody proposed that "crash" or "TempError"
is acceptable. I've tracked this issue directly below
your similar problem with invalid domains:
http://www.openspf.org/RFC_4408/Errata#permerror-invalid-domains
http://www.openspf.org/RFC_4408/Errata#TLD
There's yet no problem statement for your case, is this
about crap like target-name foo..bar (adjacent dots) ?
> the spec doesn't say "MUST be an FQDN" or anything to
> that effect.
<target-name> is clearly underspecified, and as that has
confused several developers it's also a clear erratum.
But just stating "underspecified" is lame, let's talk
about it again in some weeks when we know what 2821bis
will (or rather would) say.
Frank
-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?list_id=1007
> It's obviously unspecified (as unfortunate as it may be)
The current state of the art wrt 2821bis is that it will
adopt the 2822 concept, one or more dot separated labels.
Which is BTW what I originally asked in November 2004:
Is the "1" in subdomain 1*( "." subdomain ) intentional
or a typo. It clearly was intentional, but the author
now changed his mind after checking out the same ideas
as we over the years: optional trailing dot everywhere
but mandatory for TLDs, no trailing dot anywhere unless
it's a TLD, did I miss something ?
So now it will be no trailing nowhere (as is in 2821 and
2822), but allowing all FQDNs. Figuring out if a single
label is garbage, something local, or a FQDN is left as
an exercise for implementors.
We've to wait for the next draft what the prose exactly
will be.
> the test suite should not require any specific behavior.
Well, so far nobody proposed that "crash" or "TempError"
is acceptable. I've tracked this issue directly below
your similar problem with invalid domains:
http://www.openspf.org/RFC_4408/Errata#permerror-invalid-domains
http://www.openspf.org/RFC_4408/Errata#TLD
There's yet no problem statement for your case, is this
about crap like target-name foo..bar (adjacent dots) ?
> the spec doesn't say "MUST be an FQDN" or anything to
> that effect.
<target-name> is clearly underspecified, and as that has
confused several developers it's also a clear erratum.
But just stating "underspecified" is lame, let's talk
about it again in some weeks when we know what 2821bis
will (or rather would) say.
Frank
-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?list_id=1007