Fix false negatives in UnreachableCode rule

[t-kameyama]

Fixes #3219

[codecov]

Codecov Report

Merging #3220 (42ed623) into master (54201cd) will increase coverage by 0.02%.
The diff coverage is 77.14%.

@@             Coverage Diff              @@
##             master    #3220      +/-   ##
============================================
+ Coverage     80.06%   80.09%   +0.02%     
- Complexity     2645     2647       +2     
============================================
  Files           440      440              
  Lines          8012     8012              
  Branches       1535     1531       -4     
============================================
+ Hits           6415     6417       +2     
  Misses          775      775              
+ Partials        822      820       -2     

Impacted Files	Coverage Δ	Complexity Δ
...h/detekt/core/baseline/IndentingXMLStreamWriter.kt	60.37% <ø> (ø)	15.00 <0.00> (ø)
...kt/generator/printer/rulesetpage/TestExclusions.kt	93.75% <ø> (ø)	4.00 <0.00> (ø)
...kotlin/io/gitlab/arturbosch/detekt/DetektPlugin.kt	82.53% <ø> (ø)	11.00 <0.00> (ø)
...urbosch/detekt/rules/performance/ArrayPrimitive.kt	70.37% <0.00%> (-3.71%)	17.00 <2.00> (+2.00)	⬇️
...lab/arturbosch/detekt/rules/style/UnusedImports.kt	85.91% <75.00%> (+0.62%)	4.00 <0.00> (ø)
...urbosch/detekt/rules/style/UselessCallOnNotNull.kt	83.78% <81.81%> (+7.31%)	17.00 <4.00> (+3.00)
...ab/arturbosch/detekt/rules/bugs/MissingWhenCase.kt	87.50% <100.00%> (-1.39%)	12.00 <3.00> (+1.00)	⬇️
...ab/arturbosch/detekt/rules/bugs/UnreachableCode.kt	100.00% <100.00%> (+7.69%)	5.00 <4.00> (-4.00)	⬆️

---

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 54201cd...c079192. Read the comment docs.

[schalkms]

Thanks for submitting this PR! The rule is now more sophisticated than ever before with the help of type and symbol solving. 🙂
Consequently, this change is breaking. We would need to add the require annotation to the rule comment.
Furthermore, the check for an empty bindingContext is missing in my opinion.