Hi Seth,
I've been able to reproduce the problem. Will look into it.
I suspect it may an issue with some member variables not being
threadlocal in the inner workings of the StepCreator.
Thanks for your feedback. Always useful and helpful.
Cheers
On 24/10/2011 15:25, Seth Carter wrote:
Using the threads example in /jbehave-core/examples/threads to learn
more about the storyTimeoutInSecs I added
<storyTimeoutInSecs>5</storyTimeoutInSecs>
to properties stanza and
<storyTimeoutInSecs>${storyTimeoutInSecs}</storyTimeoutInSecs>
to the jbehave-maven-plugin configuration in
jbehave-core/examples/threads/pom.xml
I also made Tom count to 15 so his thread would run longer than Huck's
but I get strange results. Only Huck's story gets timeout warning and
Tom seemingly stops counting when Huck does.
Console output:
[INFO] Running stories using embedder
[INFO] Found class names: [org.jbehave.examples.threads.ThreadsStories]
[INFO] Running embeddable org.jbehave.examples.threads.ThreadsStories
[INFO] Processing system properties {}
[INFO] Using 2 threads
[INFO] Using executor service
java.util.concurrent.ThreadPoolExecutor@1df5f59[Running, pool size =
0, active threads = 0, queued tasks = 0, completed tasks = 0]
[INFO] Running story org/jbehave/examples/threads/stories/a_long.story
[INFO] Running story
org/jbehave/examples/threads/stories/another_long.story
Huck says 0 Mississippi
Tom says 0 Mississippi
Huck says 1 Mississippi
Tom says 1 Mississippi
Huck says 2 Mississippi
Tom says 2 Mississippi
Huck says 3 Mississippi
Tom says 3 Mississippi
Huck says 4 Mississippi
Tom says 4 Mississippi
Huck says 5 Mississippi
Tom says 5 Mississippi
Huck says 6 Mississippi
Tom says 6 Mississippi
Huck says 7 Mississippi
Tom says 7 Mississippi
Huck says 8 Mississippi
Tom says 8 Mississippi
Huck says 9 Mississippi
Tom says 9 Mississippi
(org/jbehave/examples/threads/stories/a_long.story)
(org/jbehave/examples/threads/stories/another_long.story)
Scenario: A very long scenario
Scenario: A very long scenario
[32mWhen [1;32mTom[0;32m counts to [1;32m15[0;32m Mississippi
[0m[32mWhen [1;32mHuck[0;32m counts to [1;32m10[0;32m Mississippi
[0m
[WARNING] Story org/jbehave/examples/threads/stories/a_long.story has
timed out after 6 seconds
(AfterStories)
Would someone else confirm this behavior?
Thanks,
Seth
On Friday, October 21, 2011, Seth Carter <[email protected]
<mailto:[email protected]>> wrote:
> Excellent, thanks. First I'll investigate the native story level
timeout for the easy win and post back if/when I can get a finer
grained approach working similar to the junitperf options.
> Thanks again.
> Seth
>
> On Friday, October 21, 2011, Mauro Talevi
<[email protected] <mailto:[email protected]>> wrote:
>> And the time each story takes to complete is reported in the
reports view.
>>
>> If on the other hand you want to have some more fine grained
stop-watch metrics (e.g. at scenario level) you'll need to implement
them yourself.
>>
>> On Fri Oct 21 11:02:14 2011, Mauro Talevi wrote:
>>>
>>> Hi Seth,
>>>
>>> JBehave already supports the concept of timeout for a given story.
You can specify the timeout in secs via the Embedder, Maven or Ant.
>>>
>>> And yes, JBehave can certainly be used for performance testing.
>>>
>>> Cheers
>>>
>>> On 21/10/2011 02:16, Seth Carter wrote:
>>>>
>>>> in another life I worked on a python based test framework using
pyunit and pyunitperf and was able to add a test and run it normally
for pass/fail based on business logic, if that test lent itself nicely
for performance I could run it (same test) inside a timed wrapper for
pass/fail based on a time limit. Further, I could supply tolerance
time, number of users, number of iterations/user, and a delay between
iterations/test/user. I'm now working with a series of java apps and
(of course) jbehave and have come to the same question from a developer:
>>>> "Well what if I want to make sure my test(scenario) runs in under
x seconds?"
>>>> This reminded me of the problem solved with pyUnitPerf. I liked
the idea of writing the test once with the possibility to time, or
load and time it. pyUnitPerf is a port of jUnitPerf so I figured the
same is possible with a jbehave test as it is based on junit?
>>>>
>>>> The real problem is my java skills are laughable (but improving),
I'm wondering if anyone has fooled around with this?
>>>> Also wondering if there is some performance module in jbehave
that I have completely missed.
>>>> Thanks,
>>>> Seth
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe from this list, please visit:
>>>
>>> http://xircles.codehaus.org/manage_email
>>>
>>>
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe from this list, please visit:
>>
>> http://xircles.codehaus.org/manage_email
>>
>>
>>
