Current strategy of providing the CompilationPhases straight in the enum fires back when you look at the profiles, when you see these "distinct" versions: jdk.nashorn.internal.codegen.CompilationPhase$1.transform jdk.nashorn.internal.codegen.CompilationPhase$2.transform jdk.nashorn.internal.codegen.CompilationPhase$3.transform jdk.nashorn.internal.codegen.CompilationPhase$4.transform jdk.nashorn.internal.codegen.CompilationPhase$5.transform jdk.nashorn.internal.codegen.CompilationPhase$6.transform jdk.nashorn.internal.codegen.CompilationPhase$7.transform jdk.nashorn.internal.codegen.CompilationPhase$8.transform jdk.nashorn.internal.codegen.CompilationPhase$9.transform jdk.nashorn.internal.codegen.CompilationPhase$10.transform jdk.nashorn.internal.codegen.CompilationPhase$11.transform jdk.nashorn.internal.codegen.CompilationPhase$12.transform Naming these classes properly should help to get more understandable profiles -- that probably explodes the CompilationPhase implementations from the enum into the top-level classes.
|