Freeradius-Users Digest, Vol 167, Issue 48

[Daniel Finger]

Am 18.03.2019 um 21:40 schrieb freeradius-users-request at lists.freeradius.org:
> Date: Mon, 18 Mar 2019 13:06:13 -0400
> From: Alan DeKok<aland at deployingradius.com>
> Subject: Re: Please help me understand the changes in
> 	dictionary.rfc4679
>
> On Mar 18, 2019, at 1:01 PM, Daniel Finger<daniel.finger at ewetel.de>  wrote:
>> I am a bit confused about the dictionary.rfc4679
>> There are 2 entries (ADSL-Agent-Remote-Id and ADSL-Agent-Circuit-Id), that were changed from string to octets:
>> https://github.com/FreeRADIUS/freeradius-server/commit/0838f3d58a56f63f715156cea110426e6e514f45#diff-d50b31ca7578df0305cc37e831799353
>>
>> The RFC states:
>> The format of the value field is one of 2 data types, string or integer [RFC2865] (from:https://tools.ietf.org/html/rfc4679#section-3.2)
>>
>> Can you help me understand why this is octet and not string.
>    Because RFC 2038 used "string" for both printable and non-printable attributes.  That was wrong, so I made FreeRADIUS use "string" for printable ones, and "octets" for binary ones.
>    The IETF didn't listen to me (at the time), and chose to use "text" for printable attributes, and "string" for binary ones.  So they're wrong.
>    We don't use the IETF names, and won't use them.  The IETF uses "octets" when referring to binary data everywhere else.  And using "string" is wrong, IMHO.  But that can't be changed now.
>    Alan DeKok.

Thanks for the explanation. I expected that there was a good reason behind
that decision :-)

Then I'll change my configuration to use %{string:ADSL-Agent-Remote-Id} as
in my case there are only printable strings in those fields.

Greetings,
Daniel Finger