BIND 10 #2449: Make needed things for bindctl readline support more clear

[BIND 10 Development]

#2449: Make needed things for bindctl readline support more clear
-------------------------------------+-------------------------------------
                   Reporter:  shane  |                 Owner:  jinmei
                       Type:         |                Status:  reviewing
  enhancement                        |             Milestone:
                   Priority:         |  Sprint-20121120
  medium                             |            Resolution:
                  Component:  bind-  |             Sensitive:  0
  ctl                                |           Sub-Project:  Core
                   Keywords:         |  Estimated Difficulty:  2
            Defect Severity:  High   |           Total Hours:  0
Feature Depending on Ticket:         |
  alpha2                             |
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------
Changes (by muks):

 * owner:  UnAssigned => jinmei


Comment:

 Replying to [comment:7 jinmei]:
 > I've committed my suggested solution to trac2449.  Please review.

 It looks fine to me. Please go ahead and merge.

 > I couldn't come up with an effective way of testing this
 > automatically, so I didn't add an explicit unit test.  (Although
 > indirectly, it would cause regression in some lettuce tests if we
 > show the warning even if stdin isn't a terminal, in which sense we
 > can at least know the code was used in the test).  If there's any
 > effective and meaningful way of testing it explicitly I'll add it.

 In this case, we can probably skip it. It doesn't seem to something that a
 test can help. If you must test it, you can move the stuff from the top of
 `run()` into a different method, and make a mock class deriving from
 `BindCmdInterpreter` which just sets a flag if that method is called.
 Then, set `bindcmd.my_readline` both ways, call `run()` (making
 `self.login_to_cmdctl()` fail and hence returning early) and see whether
 that flag is set or not. But I really think this much of testing is
 overkill for just printing a message.

 > I think the documentation update should go to #2305.  I'll make a
 > comment about it in that ticket.

 Verified that this has been added in `trac2305`.

 > I don't think we need a changelog for this.

 Nod.

-- 
Ticket URL: <http://bind10.isc.org/ticket/2449#comment:9>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development