BIND 10 #2410: race condition(?) in dhcp[46]_test.py

[BIND 10 Development]

#2410: race condition(?) in dhcp[46]_test.py
-------------------------------------+-------------------------------------
                   Reporter:  jelte  |                 Owner:  marcin
                       Type:         |                Status:  reviewing
  defect                             |             Milestone:  Sprint-
                   Priority:         |  DHCP-20121129
  medium                             |            Resolution:
                  Component:  dhcp   |             Sensitive:  0
                   Keywords:         |           Sub-Project:  DHCP
            Defect Severity:  N/A    |  Estimated Difficulty:  4
Feature Depending on Ticket:         |           Total Hours:  0
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------
Changes (by stephen):

 * owner:  stephen => marcin


Comment:

 > Doesn't this object die when the function ends anyway?
 It should do, which is why I think the issue is more complicated than it
 seems.  There are several Python unit tests within each file.  If the "pi
 = None" line is omitted, the all  the unit tests report the resource leak.
 With it in, only the first one reports the leak.

 From checks, the unclosed file is not one opened directly by the unit test
 code, it appears to be one opened by the subprocess.Popen object (part of
 the Python library) within !ProcessInfo.  The same leak happens with the
 tests for src/bin/bind10.  Ticket #2403 has been opened for fix
 miscellaneous resource leaks.

 As the current code fixes the problem (and also #2496), I suggest we close
 this ticket and leave the resource leak to #2403.

-- 
Ticket URL: <http://bind10.isc.org/ticket/2410#comment:12>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development