Committed, testsuite: correct description of bug in execute/20020615-1.c, PR 7042

[Hans-Peter Nilsson]

Not willing to take blame for the bug :-) I committed this as
obvious to main trunk and 3.1 branch.

	* gcc.c-torture/execute/20020615-1.c: Correct bug-description.

Index: 20020615-1.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/testsuite/gcc.c-torture/execute/20020615-1.c,v
retrieving revision 1.1
diff -p -c -r1.1 20020615-1.c
*** 20020615-1.c	15 Jun 2002 18:14:50 -0000	1.1
--- 20020615-1.c	17 Jun 2002 23:24:37 -0000
***************
*** 1,8 ****
! /* PR target/7042.  In cris.md the return pattern and its condition
!    had a bug uncovered by the combination of saved registers and
!    filled delay slot, where the delay-slot insn set the
!    return-value register.  Derived from ghostscript-6.52 (GPL)
!    by hp@axis.com.  */
  
  typedef struct font_hints_s {
    int axes_swapped;
--- 1,8 ----
! /* PR target/7042.  When reorg.c changed branches into return insns, it
!    completely forgot about any current_function_epilogue_delay_list and
!    drops those insns.  Uncovered on cris-axis-elf, where an insn in an
!    epilogue delay-slot set the return-value register with the test-case
!    below.  Derived from ghostscript-6.52 (GPL) by hp@axis.com.  */
  
  typedef struct font_hints_s {
    int axes_swapped;

brgds, H-P