Java failures [Re: 75 GCC HEAD regressions, 0 new, with your patch on 2005-04-20T14:39:10Z.]

[Andrew Haley]

Andrew Pinski writes:
 > 
 > On Apr 22, 2005, at 6:25 AM, Andrew Haley wrote:
 > 
 > > GCC regression checker writes:
 > >> With your recent patch, GCC HEAD has some regression test failures,
 > >> which used to pass.  There are 0 new failures, and 75
 > >> failures that existed before and after that patch; 0 failures
 > >> have been fixed.
 > >>
 > >> The old failures, which were not fixed or introduced by your patch, 
 > >> are:
 > >> native g++.sum g++.dg/other/static11.C
 > >> native gcc.sum gcc.c-torture/compile/20030921-1.c
 > >> native gcc.sum gcc.c-torture/compile/20040909-1.c
 > >> native gcc.sum gcc.dg/uninit-1.c
 > >> native libjava.sum ArrayClass
 > >> native libjava.sum ArrayStore
 > >> native libjava.sum ArrayStore2
 > >> native libjava.sum Array_1
 > >> native libjava.sum Array_3
 > >> native libjava.sum Class_1
 > >
 > > ...
 > >
 > > This seems to come and go.  I have been unable to duplicate this on
 > > any machine to which I have access.  I spent quite some time last week
 > > working on a Darwin build, and I got a clean set of test results.  I
 > > don't know if this is a problem of the compiler or the test suite.
 > 
 > the libjava failures on the mainline like the above are the ones which
 > happen at -O3 and are the ICE in postreload.c.

Ah, thank you.  As far as I can tell this assertion is untrue when
using non-call exceptions, and should be removed.  If someone thinks
it's useful, we can conditionalize it on ! non_call_excetptions.

OK?

Andrew.


2005-04-22  Andrew Haley  <aph@redhat.com>

	* postreload-gcse.c (hash_scan_set): Removve bogus assertion.

Index: gcc/postreload-gcse.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/postreload-gcse.c,v
retrieving revision 2.13
diff -u -r2.13 postreload-gcse.c
--- gcc/postreload-gcse.c	12 Apr 2005 12:23:49 -0000	2.13
+++ gcc/postreload-gcse.c	22 Apr 2005 11:46:49 -0000
@@ -737,9 +737,6 @@
   if (JUMP_P (insn) || set_noop_p (pat))
     return;
 
-  /* We shouldn't have any EH_REGION notes post reload.  */
-  gcc_assert (!find_reg_note (insn, REG_EH_REGION, NULL_RTX));
-
   if (REG_P (dest))
     {
       if (/* Don't CSE something if we can't do a reg/reg copy.  */

Andrew.