Improve CHA runtime and improve InvokeInterface resolving#734
Improve CHA runtime and improve InvokeInterface resolving#734JonasKlauke merged 24 commits intodevelopfrom
Conversation
Transformed Callgraphs tests to bytecode. Improved Default Method resolving
…the invoke is an interface invoke
…untime # Conflicts: # sootup.core/src/main/java/sootup/core/typehierarchy/MethodDispatchResolver.java
… unnecessary tests. Moved utility methods to the actual class using them. moved tests.
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## develop #734 +/- ##
=============================================
- Coverage 63.64% 63.58% -0.06%
+ Complexity 3360 3338 -22
=============================================
Files 317 317
Lines 15124 15072 -52
Branches 2551 2543 -8
=============================================
- Hits 9625 9583 -42
+ Misses 4608 4601 -7
+ Partials 891 888 -3 ☔ View full report in Codecov by Sentry. |
…method is in the missing class. Changed the test to use the bytecode frontend
|
Which Method do you mean? Sourcecode might not be correct. I wrote this down from mind |
|
But we could think about if we want to provide an option to define a maximum depth for a call graph algorithm. For example a method which just looks at the given method and does not continue further. But for RTA we have the problem that we miss instantiations of classes which result in new targets. But CHA should be okay |
Uh oh!
There was an error while loading. Please reload this page.