Partial fix for Throughput Shaping Timer

[pbaykalov]

This PR partially fixes erratic behaviour of Throughput Shaping Timer with low requested intensity.

This is the problem:

TST_MCVE.zip

Actual intensity might strongly depend on what exactly the requested intensity is and how many threads there are.

The timer is still not guaranteed to work correctly in specific but realistic situaitons.

[codecov]

Codecov Report

Merging #558 (c4580a8) into master (da6d10c) will increase coverage by 0.00%.
The diff coverage is 0.00%.

@@            Coverage Diff            @@
##             master     #558   +/-   ##
=========================================
  Coverage     68.94%   68.95%           
  Complexity     2574     2574           
=========================================
  Files           230      230           
  Lines         15609    15609           
  Branches       1589     1589           
=========================================
+ Hits          10762    10763    +1     
+ Misses         4057     4055    -2     
- Partials        790      791    +1     

Impacted Files	Coverage Δ
.../kg/apc/jmeter/timers/VariableThroughputTimer.java	82.24% <0.00%> (-0.60%)	⬇️
...ava/kg/apc/jmeter/jmxmon/JMXMonConnectionPool.java	82.81% <0.00%> (+3.12%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[pbaykalov]

What's the main reason for return 0; in delay()?

[undera]

What's the main reason for return 0; in delay()?

It's the default branch, in case we're not delivering enough hits. The meaning of 0 here is "do the next sample as soon as possible, don't add any delay".

[pbaykalov]

What's the main reason for return 0; in delay()?

It's the default branch, in case we're not delivering enough hits. The meaning of 0 here is "do the next sample as soon as possible, don't add any delay".

It's not "default" it's the only return statement in delay.

[undera]

What's the main reason for return 0; in delay()?

It's the default branch, in case we're not delivering enough hits. The meaning of 0 here is "do the next sample as soon as possible, don't add any delay".

It's not "default" it's the only return statement in delay.

Ok, maybe we're talking about different places in code. Can you point to exact file and line that you're talking about?

[undera]

What's the main reason for return 0; in delay()?

It's the default branch, in case we're not delivering enough hits. The meaning of 0 here is "do the next sample as soon as possible, don't add any delay".

It's not "default" it's the only return statement in delay.

Ok, I see what you mean. So, the meaning of it is we tell JMeter that there should be no delay at all on its side, since our code has already did introduce the required delay with wait()