Sauer, Damon wrote:
and version 2 adds support for something like this:
<my_mail_comes_from_ip>10.0.0.1
<unless_sender_is>bob(_at_)example(_dot_)com</unless_sender_is>
</my_mail_comes_from_ip>
This works unless you have a million names on your whitelist.
I think that is where it should be kept.
Also, you are going to have to add a bunch or <or>'s here
It would seem you missed my point. I wasn't arguing
the example was reasonable or even an extension
that would be considered, just that it providing an
example of where ignoring tags could lead to odd
behavior.
<my_mail_comes_from_ip>10.0.0.1
<or>11.0.0.1</or>
<or>12.0.0.1</or>
</my_mail_comes_from_ip>
Also, lets say that you are a business or ISP that has lots of people that
travel.
<my_mail_comes_from_ip>10.0.0.1
<or>11.0.0.1</or>
<or>12.0.0.1</or>
<visitors>yes</visitors> (this can even be omitted and default to 'yes')
</my_mail_comes_from_ip>
Or, lets say that I am a web-based mail system and you should NEVER expect mail
from my domain to be sent via someone else's network.
<my_mail_comes_from_ip>10.0.0.1
<or>11.0.0.1</or>
<or>12.0.0.1</or>
<visitors>no</visitors>
</my_mail_comes_from_ip>
I guess my point is, the more data that you have to add.. the more sense it
makes to use XML.
Especially considering the diverse OS's out there.
the more sense the SPFv1 exists system makes! You're arguing
directly for needing TCP to transmit the data now since it's
well beyond the 127/255/512 byte limits.