Re: evaluated result is wrong

[Stephan Kirsten]

Alan DeKok schrieb:

Stephan Kirsten wrote:

I have a question regarding this debug log output. I'm a bit confused
about this logic operations and the boolean result:

  "signed 32-bit integer"

okay, thanx for pointing me on this.


where can i read more about this limitation?

i've read in the dictionary.freeradius.intern file that these attribute are integer, but i wasnt sure about the size and obviously it can handle 64bit values. the only limitiation i've found was in rlm_expr, where it says, it can handle unsigned 32-bit integers

Sun Dec 20 15:44:46 2009 : Info:        expand: %{control:Tmp-Integer-5}
-> 2785017856

  That's larger than 2^31. :(

obviously :/

So how does this come together? The debug log is attached to this mail.

I hope someone ca clarify this problem

  I'll take a look at making these checks signed 64-bit, or unsigned
32-bit for 2.1.8.

that would be great, also enhancing expr for signed 64-bit integer would be appreciated :)

And it's of course good to know, that freeradius is working as designed :)

-Stephan

smime.p7s
Description: S/MIME Cryptographic Signature

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html