Problem: No wait when getting value from queue#2423
Conversation
Solution: Random failure of test `tests/test_events.py::test_event_handler_raises_when_called_after_start` seem to be because of `get_nowait` which returns way to quickly even when the queue has data
Codecov Report
@@ Coverage Diff @@
## master #2423 +/- ##
==========================================
+ Coverage 88.43% 88.48% +0.04%
==========================================
Files 39 39
Lines 2275 2275
==========================================
+ Hits 2012 2013 +1
+ Misses 263 262 -1 |
ttmc
left a comment
ttmc
left a comment
There was a problem hiding this comment.
I don't fully understand what's going on here, but the description is much better now, so it should help others understand the reasoning in the future.
bigchaindb/events.py
Outdated
|
|
||
| try: | ||
| self.started_queue.get_nowait() | ||
| self.started_queue.get(timeout=0.01) |
There was a problem hiding this comment.
A bigger timeout (say, a second) won't hurt - a successful build will take the same amount of time anyway while if there is an issue, waiting a second is not a big deal.
There was a problem hiding this comment.
I really don't know what would be an appropriate value of this timeout i.e. 10ms, 100ms, 1sec? But I do understand the having no timeout at all would lead us back to the situation wherein the test failed randomly. Also one must keep in mind that while the timeout will help the test case it will also impact the code performance.
There was a problem hiding this comment.
Since CI often fails for us here it makes sense to stay on the safe side and set a second. It only impacts the performance of the unsuccessful case. Do you expect failed retrievals to happen a lot? I suppose, we should not expect them to happen often so it's ok to wait for a second if it happens.
There was a problem hiding this comment.
Do you expect failed retrievals to happen a lot?
The only place where get_subscriber_queue is called in the code base is here. So it seems that it might not effect much.
Solution: Increase timeout
5 participants
Solution: Random failure of test
tests/test_events.py::test_event_handler_raises_when_called_after_startseems to be because ofget_nowaitwhich returns way to quickly even when the queue has data i.e. the test expect the queue to have the message'STARTED'which should be returned when.get_nowait()is called. But sinceget_nowait()returns without waiting which causes a queueEmptyexception rather thanRuntimeexception.