Detection thresholds and numerical detection

[giograno]

I implemented several improvements to the original detection. There is a new runner implemented in Kotlin. Have a look at the README file to see which kind of parameters does it accept. For each smell, there are now two kind of thresholds that work as detection rule:

• Default thresholds: the old ones. For instance, a default threshold for Eager Test is > 1 calls to production class methods;
• Spadini thresholds: more severe thresholds proposed by Spadini et.al. in their paper; e.g., for Eager Test the threshold is 4 calls.

The user can now specify at which granularity he wants to perform the detection. A boolean granularity return true of false is in a test suite there is at least 1 method that is smelly. A numerical granularity returns the number of smelly elements (i.e., test methods) in a test file.

All these options can be specified by the user via command line (again, have a look at the README).

[mkaouer]

Dear @giograno , thank you for your hard work. We are in the middle of implementing the V2 of this tool.

We plan on taking into account your thresholds in the new release. So we will be in touch with you in this regard.

@shehan it is about time for us make sure this is also taken care of in the new release

[giograno]

Many thanks for your comment @mkaouer! In case you need it, I'd be happy to help in any form :)

[shehan]

Hi @giograno! Thanks for your interest in the tool. We are currently gathering feedback from users to determine the improvements and features that should be incorporated into the next version of the tool.

We would really appreciate it if you can complete this feedback survey based on your experience with the tool: https://docs.google.com/forms/d/16Z2H0r8OjtC0gwxnZGCoV7T32kynQuivXWnFGIr3nHc

Thanks again for your contribution!

[mkaouer]

Thank you @giograno for the hard work and thank you @shehan for the review.

We are now making sure to see how we can incorporate these changes in our new tool.