This PR contains a suggestion to use vm_commit(); in place of janet_stack_frame(stack)->pc = pc;.

The vm_commit(); construct is used elsewhere within the JOP_CALL handling in which this instance of janet_stack_frame(stack)->pc = pc; appears:

• janet/src/core/vm.c

  Line 975 in 3a4d56a

  vm_commit();

• janet/src/core/vm.c

  Line 993 in 3a4d56a

  vm_commit();

• janet/src/core/vm.c

  Line 1002 in 3a4d56a

  vm_commit();

Perhaps it makes sense to use vm_commit(); on line 983 as well?

Though I can see that if one wanted to leave the JANET_FUNCTION-handling parts of JOP_CALL and JOP_TAILCALL easier to compare, perhaps not using vm_commit(); might be preferrable.