Error messages in debug on 3.0
I've just ported our config to 3.0 and I'm seeing a few error messages;
they don't seem to be critical but are concerning me.
Specifically I'm seeing:
ERROR: Conditional evaluation failed due to internal sanity check.
...whenever I try to compare against absent attributes. What's the
correct syntax for this now - do I need:
if ((Attr) (Attr op RHS)) {
...or can I ignore the message?
I'm also seeing this with:
if (%{outer.request:Blah})
...if I'm not in a tunnel (in some generic logging policy); what's the
right syntax for that?
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: Error messages in debug on 3.0
On 10/10/13 18:32, Phil Mayers wrote: I've just ported our config to 3.0 and I'm seeing a few error messages; they don't seem to be critical but are concerning me. Specifically I'm seeing: We're also getting: Info: Invalid operator for item Sql-Group: reverting to '==' ...which is logged to radiusd.log. This seems to be spurious - all our comparisons to SQL-Group are, of course using ==. Other than that, it all seems OK! - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: Error messages in debug on 3.0
On 10 Oct 2013, at 18:32, Phil Mayers p.may...@imperial.ac.uk wrote:
I've just ported our config to 3.0 and I'm seeing a few error messages; they
don't seem to be critical but are concerning me.
Specifically I'm seeing:
ERROR: Conditional evaluation failed due to internal sanity check.
...whenever I try to compare against absent attributes. What's the correct
syntax for this now - do I need:
if ((Attr) (Attr op RHS)) {
Yes.
...or can I ignore the message?
I'm also seeing this with:
if (%{outer.request:Blah})
...if I'm not in a tunnel (in some generic logging policy); what's the right
syntax for that?
possibly if (outer.request
Sorry about the errors, I think they should be more descriptive in master
unless Alan changed something when he did the second pass fix up for conditions.
But yes, in general a presence check is now required. It makes it easier to
spot attributes which should be there but aren't.
-Arran
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: Error messages in debug on 3.0
On 10/10/13 18:51, Arran Cudbard-Bell wrote: possibly if (outer.request Hmm, no same thing, and worse it's squashing Module-Failure-Message :o( - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: Error messages in debug on 3.0
Phil Mayers wrote:
I've just ported our config to 3.0 and I'm seeing a few error messages;
they don't seem to be critical but are concerning me.
Specifically I'm seeing:
ERROR: Conditional evaluation failed due to internal sanity check.
That should be fixed. Either it can be deleted, or reworded.
...whenever I try to compare against absent attributes. What's the
correct syntax for this now - do I need:
if ((Attr) (Attr op RHS)) {
...or can I ignore the message?
Yes.
Alan DeKok.
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
Re: Error messages in debug on 3.0
On 10 Oct 2013, at 22:23, Alan DeKok al...@deployingradius.com wrote:
Phil Mayers wrote:
I've just ported our config to 3.0 and I'm seeing a few error messages;
they don't seem to be critical but are concerning me.
Specifically I'm seeing:
ERROR: Conditional evaluation failed due to internal sanity check.
That should be fixed. Either it can be deleted, or reworded.
I've reworded it. The code in master actually tells you what was wrong, but the
changes
were quite extensive so it didn't get ported to v3.0.x.
I've also removed the error on accessing a request which doesn't exist, so
presence checks work as expected.
The actual presence check itself is ok.
In radius_evaluate_tmpl:
case VPT_TYPE_ATTR:
case VPT_TYPE_LIST:
if (radius_vpt_get_vp(request, vpt) != NULL) {
rcode = true;
} else {
rcode = false;
}
break;
Which works for
if (outer.request:Attribute)
and
if (outer.request)
It's just one of the functions called to convert the enumerated request type to
an actual REQUEST threw an error when the request didn't exist.
Arran Cudbard-Bell a.cudba...@freeradius.org
FreeRADIUS Development Team
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
