[patch] jv-lang.c, dead code

[patch] jv-lang.c, dead code

[msnyder]

[-- Attachment #1: Type: text/plain, Size: 25 bytes --]

noside is tested twice.


[-- Attachment #2: 12.txt --]
[-- Type: text/plain, Size: 697 bytes --]

2007-08-04  Michael Snyder  <msnyder@access-company.com>

	* jv-lang.c (evaluate_subexp_java): Dead code.

Index: jv-lang.c
===================================================================
RCS file: /cvs/src/src/gdb/jv-lang.c,v
retrieving revision 1.48
diff -p -r1.48 jv-lang.c
*** jv-lang.c	4 Aug 2007 02:50:16 -0000	1.48
--- jv-lang.c	5 Aug 2007 04:06:23 -0000
*************** evaluate_subexp_java (struct type *expec
*** 857,864 ****
  	  type = type_from_class (java_class_from_object (arg1));
  	  arg1 = value_cast (lookup_pointer_type (type), arg1);
  	}
-       if (noside == EVAL_SKIP)
- 	goto nosideret;
        return value_ind (arg1);
  
      case BINOP_SUBSCRIPT:
--- 857,862 ----

Re: [patch] jv-lang.c, dead code

[Jim Blandy]

msnyder@sonic.net writes:
> noside is tested twice.

I'd agree that the code is messed up, and that your patch doesn't
change its behavior, but which test is the one that ought to be there?
I don't even know when EVAL_SKIP would be used for Java.

Re: [patch] jv-lang.c, dead code

[msnyder]

>
> msnyder@sonic.net writes:
>> noside is tested twice.
>
> I'd agree that the code is messed up, and that your patch doesn't
> change its behavior, but which test is the one that ought to be there?
> I don't even know when EVAL_SKIP would be used for Java.

In that, I would have to defer to you.

Annotate says that all four instances of EVAL_SKIP in jv-lang.c
have been there from the beginning (1999).

The only other places where it occurs are in eval.c and ada-lang.c.

I would bet that the instances in jv-lang.c were copied from
a common ancestor with the bits in eval.c.  The OP_STRING case,
for instance, looks like a close cousin.

Re: [patch] jv-lang.c, dead code

[msnyder]

>
> msnyder@sonic.net writes:
>> noside is tested twice.
>
> I'd agree that the code is messed up, and that your patch doesn't
> change its behavior, but which test is the one that ought to be there?
> I don't even know when EVAL_SKIP would be used for Java.

Any thoughts on what we ought to do about this?

Re: [patch] jv-lang.c, dead code

[Jim Blandy]

msnyder@sonic.net writes:
>> msnyder@sonic.net writes:
>>> noside is tested twice.
>>
>> I'd agree that the code is messed up, and that your patch doesn't
>> change its behavior, but which test is the one that ought to be there?
>> I don't even know when EVAL_SKIP would be used for Java.
>
> Any thoughts on what we ought to do about this?

I vote leave it alone.