[Bug optimization/13469] [3.4 Regression] IRIX 6.5 O32 libjava bootstrap failure: ICE in verify_local_live_at_start, at flow.c:557

["ro at techfak dot uni-bielefeld dot de" <gcc-bugzilla at gcc dot gnu dot org>]

------- Additional Comments From ro at techfak dot uni-bielefeld dot de  2004-01-07 19:26 -------
Subject: Re:  [3.4 Regression] IRIX 6.5 O32 libjava bootstrap failure: ICE in verify_local_live_at_start, at flow.c:557

I've now re-bootstrapped current mainline with that patch: at least the
bootstrap finishes, but all the O32 libjava execution tests fail, e.g. 

FAIL: ArrayStore execution - source compiled test

Program received signal SIGSEGV, Segmentation fault.
strlen () at strlen.s:58
58      strlen.s: No such file or directory.
        in strlen.s
Current language:  auto; currently asm
(gdb) where
#0  strlen () at strlen.s:58
#1  0x0fa3e7e0 in _strdup () at strdup.c:29
#2  0x5f921f68 in _Jv_RunMain(java::lang::Class*, char const*, int, char const**, bool) () from ../../mabi=32/libjava/.libs/libgcj.so.5
#3  0x5f9222a0 in JvRunMain () from ../../mabi=32/libjava/.libs/libgcj.so.5
#4  0x00401e40 in main ()

But this seems to be nothing new, it happened also during a
bootstrap/regtest on 20031119, before the current ICE started.  I'll
probably have to bootstrap with GNU as once again (maybe with and without
--enable-sjlj-exceptions), so I can see what's going on here.

Apart from that, there seem to be no new unexpected failures compared to
20031208, so the patch causes no regressions.  The only new failures are

-mabi=n32:

+FAIL: gcc.dg/20031201-1.c (test for excess errors)

known, but not yet fixed:

http://gcc.gnu.org/ml/gcc-regression/2004-01/msg00012.html

-mabi=32:

+FAIL: gcc.dg/20031201-1.c (test for excess errors)
+FAIL: gcc.dg/dwarf-die5.c scan-assembler 3.*DW_AT_inline
+FAIL: gcc.dg/dwarf-die7.c scan-assembler 1.*DW_AT_inline

most likely due to lack of debug information

-mabi=64:

+FAIL: gcc.dg/20031201-1.c (test for excess errors)
+FAIL: gcc.dg/winline-7.c (test for excess errors)

testcase bug:

winline-7.c:7: warning: cast from pointer to integer of different size

	Rainer


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13469