Refactor Squeak Test Execution

[f4lco]

Major selling points:

• first and foremost, colored and folded Travis logs
• Mark individual Squeak test selectors as (un)stable
• Inside IDEs, be able to execute a single Squeak test class, and make use of rerun all / rerun failed / rerun single test selector:

The Squeak test class is chosen via system property:

# Sample VM argument
-DsqueakTestClass=ObjectTest

# Example MX invocation with individual test case names and timing
$ mx unittest -DsqueakTestClass=UTF8EdgeCaseTest SqueakSUnitTest --very-verbose --enable-timing

I've come to the conclusion that the Travis build log visualization is unreliable. I cannot explain artifacts like these:

Given that the raw build logs look OK to me.

I cannot meet the goals for coverage - I added test code for handling test case failures, which are naturally not part of a successful build.

[codecov]

Codecov Report

Merging #14 into dev will increase coverage by 0.76%.
The diff coverage is 81.76%.

@@             Coverage Diff              @@
##                dev      #14      +/-   ##
============================================
+ Coverage     51.95%   52.72%   +0.76%     
- Complexity     2432     2513      +81     
============================================
  Files           182      188       +6     
  Lines         16189    16645     +456     
  Branches       2265     2315      +50     
============================================
+ Hits           8411     8776     +365     
- Misses         6918     6992      +74     
- Partials        860      877      +17

Impacted Files	Coverage Δ	Complexity Δ
...t/src/de/hpi/swa/graal/squeak/test/Statistics.java	100% <100%> (ø)	3 <3> (?)
.../swa/graal/squeak/test/AbstractSqueakTestCase.java	94.44% <100%> (+0.1%)	20 <1> (+1)	⬆️
.../de/hpi/swa/graal/squeak/test/SqueakSUnitTest.java	66.66% <62.5%> (-15.11%)	10 <10> (-39)
...test/src/de/hpi/swa/graal/squeak/test/TestLog.java	72.91% <72.91%> (ø)	0 <0> (?)
...l/squeak/test/AbstractSqueakTestCaseWithImage.java	80.73% <80.73%> (ø)	30 <30> (?)
....test/src/de/hpi/swa/graal/squeak/test/Travis.java	83.78% <83.78%> (ø)	9 <9> (?)
.../src/de/hpi/swa/graal/squeak/test/SqueakTests.java	84.44% <84.44%> (ø)	11 <11> (?)
...pi/swa/graal/squeak/test/SqueakBasicImageTest.java	84.78% <84.78%> (ø)	12 <12> (?)
...i/swa/graal/squeak/test/SqueakConsistencyTest.java	91.89% <91.89%> (ø)	8 <8> (?)
...raal/squeak/nodes/plugins/MiscPrimitivePlugin.java	86.36% <0%> (-2.03%)	2% <0%> (ø)
... and 26 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c0744c4...eabef6a. Read the comment docs.

[fniephaus]

Thanks, will have a look soon. Happy holidays!

[fniephaus]

Thanks a lot, this looks good! I've merged it in as is :)