execution time stats

[malaskowski]

Collecting some execution time stats and present them on the report. Later it may be used to analyse time consumed for processing the suite.

---

I hereby agree to the terms of the AET Contributor License Agreement.

[subiron]

can we also measure time on TEST, and URL level?
I would be helpful to detect bottlenecks without doing any computation (like sum of steps duration)

[subiron]

Is there anything else that we can measure inside Statistics class in the future?
I'm asking because it looks like we have here the whole class to wrap single field with no additional reasons.

[malaskowski]

Yes, my idea was to add more data here in the future, e.g. start/end time if required, maybe some wait/sleep time required to pass.
I'd like to have statistics of how long AET really needed to perform a test (how long loading/sleep took vs real test time).

[malaskowski]

@subiron measuring full statistics at Test and Url level is very trick and I didn't include it in this pull request.
The problem is that test and url processing (I mean both collect and compare phases) is performed in dispersed environment. AET uses (especially during comparison phase) idea similar to divide and conquer. This mean totally separate workers/nodes execute parts of test and url processing. Because of that storing Statistics for the url or test may even not make a sens - it would be better to sum up all statistics of descendant data nodes. But this can be already done with the data that is already stored in those nodes (for url, these nodes are steps, for test it would be the sum of all steps of all urls).