More ifcvt woes

[Richard Henderson]

On Sat, May 13, 2000 at 01:53:12PM -0700, Richard Henderson wrote:
> Ug.  I don't think ifcvt should be forced to deal with this kind
> of nonsense.  I wonder why cleanup_cfg didn't do its job...

Well, actually, cleanup_cfg isn't intended to handle this sort
of thing.  It should, but that code hasn't been written. 

In the meantime, see how much farther you get with this.


r~



Index: toplev.c
===================================================================
RCS file: /cvs/gcc/egcs/gcc/toplev.c,v
retrieving revision 1.335
diff -c -p -d -r1.335 toplev.c
*** toplev.c	2000/05/11 13:52:44	1.335
--- toplev.c	2000/05/14 06:29:59
*************** rest_of_compilation (decl)
*** 3344,3350 ****
    timevar_push (TV_FLOW2);
    open_dump_file (DFI_flow2, decl);
  
!   jump_optimize_minimal (insns);
    find_basic_blocks (insns, max_reg_num (), rtl_dump_file);
  
    /* On some machines, the prologue and epilogue code, or parts thereof,
--- 3344,3351 ----
    timevar_push (TV_FLOW2);
    open_dump_file (DFI_flow2, decl);
  
!   jump_optimize (insns, !JUMP_CROSS_JUMP,
! 		 !JUMP_NOOP_MOVES, !JUMP_AFTER_REGSCAN);
    find_basic_blocks (insns, max_reg_num (), rtl_dump_file);
  
    /* On some machines, the prologue and epilogue code, or parts thereof,