Re: [tor-bugs] #8865 [Stem]: test_event_handling is unreliable

["Tor Bug Tracker & Wiki" <blackhole@xxxxxxxxxxxxxx>]

#8865: test_event_handling is unreliable
---------------------+------------------------------------------------------
 Reporter:  atagar   |          Owner:  robinson
     Type:  defect   |         Status:  assigned
 Priority:  minor    |      Milestone:          
Component:  Stem     |        Version:          
 Keywords:  testing  |         Parent:          
   Points:           |   Actualpoints:          
---------------------+------------------------------------------------------

Comment(by atagar):

 > Your TODO mentions "sizable sleeps", how large have you set the
 wait()/sleep() in your experiments?

 Oh drats, I confused tests. I didn't bump the waits on
 test_event_handling. I raised them on test_reattaching_listeners...

 https://gitweb.torproject.org/stem.git/commitdiff/a5f8168

 I haven't seen issues in the jenkins test runs so maybe indeed we just
 need to raise test_event_handling's waits from two to four seconds. Giving
 it a try...

 https://gitweb.torproject.org/stem.git/commitdiff/ffd7274

 That said this timing dependency in both tests is definitely still a code
 stink.

 > Also, is it possible to check that we are running under Jenkins during a
 test (e.g. test.runner.is_jenkins())?

 Nope, we don't have a check for that. Thus far though the jenkins issues
 have indicated either a problem with its test environment or a problem
 with our tests when run under load. Ideally both would be fixed rather
 than addressed by an is_jenkins() check.

 > I am looking to use LOG event with DEBUG level as this should be very
 noisy, we can prompt output by sending a signal, and the event type goes
 back to tor 0.1.1.1.

 Why not use CONF_CHANGED? I realize that it's a newer event, but it has
 the advantage of being provided if and only if a specific controller
 action is taken (ie, SETCONF). No doubt we can get the similar behaviour
 from DEBUG, but it risks breaking if tor changes its debug statements.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/8865#comment:2>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs