rlm_sql: zero session length

[Josip Almasi]

Hi all,

I see rlm_sql complaining about stop packets with zero session length, and then just ignoring them.

This misleading message actually tells that Acct-Session-Time = 0.
And the code check for <= 0:)
The RFC tells nothing about valid values for it, it simply states it's number of seconds.

Well, these days we can do a lot in a second.
I.e. we can perform a single HTTP GET, in order to download a MMS.
And in that case, we get invalid radacct table, due to intentionally dropped records.

Or, in VoIP, Acct-Session-Time of zero means a call was attempted, but was not accepted.
(details are in another attribute)

Or... anyway, it's not a big deal, though I'd rather have consistent radacct table.

Why is this if statement there anyway, some buggy NASes?
If so, why not just let user handle it in unlang?

IMHO that needs to be removed, or if written for a good reason, made configurable. With rfc-compliant default:)

Regards...