You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To reproduce, unzip, and modify compile.sh so that CF_ROOT points to your local Checker Framework repo. Then run ./compile.sh. The checker exhausts 16GB of RAM analyzing HiveParser.java. It's a large (~30k lines) source file generated by ANTLR I think. But still, 16GB of heap for one file is a lot. Even just running the Tainting Checker on this class needs about 8GB of heap. The RLC is hurt by needing to run the MustCallChecker as a sub checker and keep its state in memory. Doing some memory profiling, it seems a significant amount of memory is consumed by the caching of flow analysis results in the org.checkerframework.framework.type.GenericAnnotatedTypeFactory#flowResult field.
The text was updated successfully, but these errors were encountered:
Test case:
hive_parse_test.zip
To reproduce, unzip, and modify
compile.sh
so thatCF_ROOT
points to your local Checker Framework repo. Then run./compile.sh
. The checker exhausts 16GB of RAM analyzingHiveParser.java
. It's a large (~30k lines) source file generated by ANTLR I think. But still, 16GB of heap for one file is a lot. Even just running the Tainting Checker on this class needs about 8GB of heap. The RLC is hurt by needing to run the MustCallChecker as a sub checker and keep its state in memory. Doing some memory profiling, it seems a significant amount of memory is consumed by the caching of flow analysis results in theorg.checkerframework.framework.type.GenericAnnotatedTypeFactory#flowResult
field.The text was updated successfully, but these errors were encountered: