builtin_return_addr vs frame_pointer_needed vs -O3

[Ian Dall]

Richard Henderson writes:
 > Good enough.  Just add some whitespace to pretty it up.


Here 'tis then. First a Changelog entry:

2002-07-03  Ian Dall  <ian@sibyl.beware.dropbear.id.au>

	* execute/20010122-1.x: Do not test with -fomit-frame-pointer
	option since it is unlikely to work on most targets.

Now the patch:

*** testsuite/gcc.c-torture/execute/20010122-1.x        30 Apr 2002 23:39:33 -0000      1.1.1.1
--- testsuite/gcc.c-torture/execute/20010122-1.x        3 Jul 2002 11:06:53 -0000
***************
*** 5,8 ****
--- 5,17 ----
        set torture_execute_xfail "mips*-sgi-irix6*"
  }
  
+ # This test relies on __builtin_return_address(1) returning something
+ # useful. This is not guaranteed to be be so, especially
+ # when the -fomit-frame-pointer flag is used.
+ # Do not test with -fomit-frame-pointer option.
+ set torture_eval_before_compile {
+     if {[string match {*-fomit-frame-pointer*} "$option"]} {
+       continue}
+     }
+ 
  return 0

Ian