CloneDetectionRoutines: Prefer big code clones

This reduces the calculated difference value for pairs of bigger clones. This makes sense as bigger function with the same relative difference are more likely to provide a refactoring opportunity than small functions. As a result of this change small functions are less likely to be recognized as clones which is a good property mirroring the real usefulness in theory. As the testsuite still passes this change does not break existing testcases, as more are added, the usefulness of this change will be evaluated.
coala · Jun 24, 2015 · 657cf78 · 657cf78 · fneu · Jun 24, 2015
1 parent d506bd3
commit 657cf78
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/bears/codeclone_detection/ClangCloneDetectionBear.py b/bears/codeclone_detection/ClangCloneDetectionBear.py
@@ -8,7 +8,7 @@
 class ClangCloneDetectionBear(GlobalBear):
     def run(self,
             dependency_results: dict,
-            max_clone_difference: float=0.42):
+            max_clone_difference: float=0.33):
         '''
         Checks the given code for similar functions that are probably
         redundant.

diff --git a/bears/codeclone_detection/CloneDetectionRoutines.py b/bears/codeclone_detection/CloneDetectionRoutines.py
@@ -113,4 +113,4 @@ def compare_functions(cm1, cm2):  # pragma: no cover
 
     # If max_sum is zero diff_sum should be zero so division by zero can't
     # occur here.
-    return diff_sum/max_sum
+    return (diff_sum/max_sum) * ((3*max_sum+1)/(4*max_sum))