Convert DetektJvmSpec to use ProjectBuilder

[3flex]

See #3778. This reduced execution of the DetektJvmSpec test by about 25% on my machine.

[codecov]

Codecov Report

Merging #4075 (ea1274e) into main (cf1d830) will increase coverage by 0.06%.
The diff coverage is n/a.

❗ Current head ea1274e differs from pull request most recent head 3a306c2. Consider uploading reports for the commit 3a306c2 to get more accurate results

@@             Coverage Diff              @@
##               main    #4075      +/-   ##
============================================
+ Coverage     83.45%   83.51%   +0.06%     
- Complexity     3185     3186       +1     
============================================
  Files           463      461       -2     
  Lines          9095     9095              
  Branches       1768     1768              
============================================
+ Hits           7590     7596       +6     
+ Misses          571      570       -1     
+ Partials        934      929       -5     

Impacted Files	Coverage Δ
...detekt/rules/style/UnderscoresInNumericLiterals.kt	81.08% <0.00%> (-0.50%)	⬇️
...etekt/generator/printer/defaultconfig/Exclusion.kt	96.66% <0.00%> (-0.11%)	⬇️
...gitlab/arturbosch/detekt/core/tooling/Lifecycle.kt	100.00% <0.00%> (ø)
...itlab/arturbosch/detekt/core/extensions/Loading.kt	20.00% <0.00%> (ø)
...tlab/arturbosch/detekt/core/reporting/Reporting.kt	94.11% <0.00%> (ø)
...ch/detekt/core/reporting/console/FindingsReport.kt	100.00% <0.00%> (ø)
.../core/reporting/console/FileBasedFindingsReport.kt	100.00% <0.00%> (ø)
...t/generator/printer/defaultconfig/ConfigPrinter.kt	92.10% <0.00%> (ø)
...etekt/core/reporting/console/LiteFindingsReport.kt
...t/core/reporting/console/AbstractFindingsReport.kt
... and 7 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 cf1d830...3a306c2. Read the comment docs.

[chao2zhang]

Pre-merge check currently fails. Is it fair to assume that the improvement comes from changing from groovy interpreter to Kotlin compiler?

[3flex]

Is it fair to assume that the improvement comes from changing from groovy interpreter to Kotlin compiler?

It's from using ProjectBuilder instead of Gradle's TestKit.

TestKit runs a full Gradle build and spins up its own daemon. ProjectBuilder is much more limited, and can't be used for functional tests, because you can't execute tasks. But for checking things like whether task input & output values are set correctly it's suitable, and faster.

[cortinico]

Great work 👍

Just one side note: now we're moving from actually running Gradle tasks to just setting up a project and asserting that the tasks are configured properly.

Not a big deal but if the task blows up during execution, the former task would have found it. This one not. Something to keep in mind 👍