breakAtPossibleNextInstructions should use const InterpreterState

Summary:
`breakAtPossibleNextInstructions()` has the same purpose as
`breakpointCaller()` and `breakpointExceptionHandler(const
InterpreterState &state);`. Their job is to set a temporary breakpoint
for where things should pause next.

This diff just makes the function signature align with
`breakpointExceptionHandler()` by using const. This differentiates the
function from things like `stepInstruction()`.

Reviewed By: avp

Differential Revision: D55273733

fbshipit-source-id: dca8782bd1356d5d80287ab075bef8922fb60cd1

include/hermes/VM/Debugger/Debugger.h

@@ -554,7 +554,7 @@ class Debugger {
   llvh::Optional<uint32_t> findJumpTarget(CodeBlock *block, uint32_t offset);
 
   /// Set breakpoints at all possible next instructions after the current one.
-  void breakAtPossibleNextInstructions(InterpreterState &state);
+  void breakAtPossibleNextInstructions(const InterpreterState &state);
 
   /// Get the actual OpCode produced from the source without being affected by
   /// any user installed breakpoint "Debugger" OpCode overrides.

lib/VM/Debugger/Debugger.cpp

@@ -172,7 +172,7 @@ llvh::Optional<uint32_t> Debugger::findJumpTarget(
 #undef DEFINE_JUMP_LONG_VARIANT
 }
 
-void Debugger::breakAtPossibleNextInstructions(InterpreterState &state) {
+void Debugger::breakAtPossibleNextInstructions(const InterpreterState &state) {
   auto nextOffset = state.codeBlock->getNextOffset(state.offset);
   // Set a breakpoint at the next instruction in the code block if this is not
   // the last instruction.