Josip Rodin wrote:
>> To avoid a memory leak, the pointers should be initialized before they
>> are used. I've committed a change to the v2.1.x branch which fixes this.
>
> This didn't seem to end up in 2.1.12, why is that?
2.1.12 came out before the patch was done. The patch went in on
Nov
Josip Rodin wrote:
> On Wed, May 09, 2012 at 01:55:10PM +1200, Jasen Betts wrote:
>> the file /usr/share/freeradius/dictionary.mikrotik
>> is not included from /usr/share/freeradius/dictionary
>
> This problem is fixed in 2.1.12.
>
>> also the file /usr/share/freeradius/dictionary.mikrotik
>> is
Josip Rodin wrote:
> That just doesn't seem safe enough when an analogous bit of parsing code in
> cf_section_parse() does:
..
> and cf_item_parse() can fail, so it stands to reason that *data can
> remain unwritten, and should not be free()'d.
The solution is to write to data. The reason is th
The configuration file for the module, raddb/modules/passwd says:
# hashsize - hashtable size. If 0 or not specified records are not
#stored in memory and file is read on every request.
#This configuration is *not* recommended, as it can be
#very slow. The
Steve Kemp wrote:
> I'd keep it unless there is a significant reason not to - grave
> bugs that can't be fixed, etc - I don't regard the gcc test as
> a good enough reason to remove it alone.
I've put an updated version on my new web site:
http://deployingradius.com/pscan/
The only thing
Josip Rodin wrote:
> OK, so I tried it, and it actually doesn't work for me. I HUP'ed my FR
> server process running since Oct 16, that logrotate had previously rotated
> out and cleared out. After the HUP, the /var/log/freeradius/radius.log file
> is still zero-sized, while the files used by linel
Josip Rodin wrote:
> Ah, no, here we go. I had stopped and started it again, and then it was able
> to tell me:
>
> Thu Nov 11 15:10:24 2010 : Info: Received HUP signal.
> Thu Nov 11 15:10:24 2010 : Info: HUP - Re-reading configuration files
> Thu Nov 11 15:10:24 2010 : Error: Unable to open file
Configuring freeradius to also listen on a IPv6 interface (in this case
::) does not work.
Here is the error message I get when starting freeradius in debug mode:
---
listen {
type = "auth"
ipv6addr = :: IPv6 address [::]
Josip Rodin wrote:
> On Mon, Nov 08, 2010 at 08:16:50AM -0500, Gedalya wrote:
>> Package: freeradius
>> Version: 2.1.9+dfsg-1+b1
>>
>> After logrotate rotates /var/log/freeradius/radius.log, radius.log stays
>> at 0 size, further logging does not work. This seems to be due to a
>> change in fre
Josip Rodin wrote:
>> The server needs to be HUP'd for the log rotation to work. It
>> previously did *not* re-open the log file on HUP. Now it does.
>
> No, but that's not what I'm saying. Previously it just kept writing the same
> file *without* a HUP.
No... it wrote to the same *filename
Josip Rodin wrote:
> Hi,
>
> Stephen Gran mentioned to me yesterday that the OpenSSL license exemption
> was in the works, that none of the contributors had complained about its
> addition.
>
> When can it be made official? :)
I think we should be able to do it for 2.1.8.
Alan DeKok.
--
Josip Rodin wrote:
> % sudo strace freeradius -X
> [...]
> socket(PF_INET6, SOCK_DGRAM, IPPROTO_IP) = 5
> getsockname(5, {sa_family=AF_INET6, sin6_port=htons(0), inet_pton(AF_INET6,
> "::", &sin6_addr), sin6_flowinfo=0, sin6_scope_id=0}, [28]) = 0
> setsockopt(5, SOL_IPV6, IPV6_PKTINFO, [1], 4) =
Josip Rodin wrote:
> After a bit more digging, it seems you need a simple:
...
> This seems to be so per RFC 3542 from 2003, which obsoletes RFC 2292
> from 1998.
Ah... you're using 2.1.10. The git "v2.1.x" branch already has a
better fix which is cross-platform.
--
To UNSUBSCRIBE, email to
Josip Rodin wrote:
> The first test works, but the second doesn't. The second one should actually
> work, but the problem is in the definition:
C compilers now have built-in "offsetof" to fix precisely this problem.
See the attached patch. You'll have to re-run "autoconf" to
re-generate the
14 matches
Mail list logo
