On 04/24/2015 01:31 PM, Petr Vobornik wrote:
On 04/24/2015 07:19 AM, Jan Cholasta wrote:
Dne 23.4.2015 v 22:18 Nathaniel McCallum napsal(a):
On Thu, 2015-04-23 at 14:12 +0200, Petr Vobornik wrote:
On 04/23/2015 12:24 PM, Petr Vobornik wrote:
If unbind was called when disconnected it raised:
AttributeError: 'NoneType' object has no attribute 'unbind_s'
AttributeError is not a public error and therefore it prevented
ldap2.destroy_connection() to be called multiple times.
Note: this issue also prevented rpcserver.change_password from
Therefore I think that there might have been an error in recent
refactoring and if #4991 was not run on master then there might
been other issue, which probably have been fixed by the
After discussion with Honza, the approach was changed.
Also I've added patch which removes unnecessary incorrect code which
revealed the regression.
Additional testing shows that these patches actually don't fix the
original issue of #4991. See
0823 - ACK
0824 - ACK
I would prefer if the connection was closed manually in patch 824, IMO
it is a good practice to release resources once you are done with them
just in time, and I don't think you can always trust the automatic
disconnect at the end of request.
Changed (also in user-status command).
Manage your subscription for the Freeipa-devel mailing list:
Contribute to FreeIPA: http://www.freeipa.org/page/Contribute/Code