[RFA] Avoid macro expansion side-effects w/DEBUG_LOCALS_INSN

[Keith Seitz]

Hi,

The attached patch fixes all users of the DEBUG_LOCALS_INSN macro (in 
the debug interpreter) which are currently experiencing macro side 
effects in certain insns. An example is op_storea:

    op_storea:
      STOREA(GET1U ());
      NEXT_INSN;

In the DEBUG interpreter, this is expanded to:

    op_storea:
      do
        {
          DEBUG_LOCALS_INSN (GET1U (), 'o');
          locals[GET1U ()].o = (--sp)->o;
        }
      while (0);

GET1U is defined as "(pc++)->int_val" (via INTVAL macro).

Obviously, this causes really bad things to happen in the DEBUG 
interpreter with the double-increment of the PC.

I have fixed these using temporary variables in the various STORE* 
macros to hold the passed value before using it in DEBUG_LOCALS_INSN and 
the rest of the macro.

Since the code doesn't diff particularly well, I would appreciate it if 
more eyes would double-check the translation, since I could very well 
have made a typo. [Although I have double-checked it forty times...]

QCC?

Keith

ChangeLog
2007-05-08  Keith Seitz  <keiths@redhat.com>

         * interpret.cc (STOREA): Rewrite using temporary variable to
         avoid double-macro expansion side-effects.
         (STOREI): Likewise.
         (STOREF): Likewise.
         (STOREL)[SIZEOF_VOID_P == 8]: Likewise.
         (STORED)[SIZEOF_VOID_P == 8]: Likewise.
         (STOREL)[SIZEOF_VOID_P != 8]: Likewise.
         (STORED)[SIZEOF_VOID_P != 8]: Likewise.
         (POKEI): Likewise.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: DEBUG_LOCALS_INSN-macro-side-effects.patch
Type: text/x-patch
Size: 3682 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/java-patches/attachments/20070509/3cced6d8/attachment.bin>