pr83 random failure

[Andrew Haley]

Michael Ritzert writes:
 > Hi all,
 > 
 > "FAIL: pr83 -O3 compilation from bytecode" fails at random for me with two machines.
 > The error mode points to a memory corruption problem:
 > 
 > logs-HEAD-i686-1095654228.tar.bz2:pr83.java: In method `pr83.tryfinally()':
 > logs-HEAD-i686-1095654228.tar.bz2:pr83.java:0: internal compiler error: virtual array basic_block_info[6]: element 4294967295 out of bounds in set_up_bb_rts_numbers, at global.c:2274
 > 
 > logs-HEAD-i686-1095692211.tar.bz2:pr83.java:0: internal compiler error: virtual array basic_block_info[6]: element 4294967295 out of bounds in set_up_bb_rts_numbers, at global.c:2274
 > 
 > logs-HEAD-i686-1095879021.tar.bz2:pr83.java:0: internal compiler error: virtual array basic_block_info[6]: element 65537 out of bounds in set_up_bb_rts_numbers, at global.c:2274
 > 
 > logs-HEAD-i686-1096011201.tar.bz2:pr83.java:0: internal compiler error: virtual array basic_block_info[6]: element 80 out of bounds in set_up_bb_rts_numbers, at global.c:2274
 > 
 > logs-HEAD-i686-1096068427.tar.bz2:pr83.java:0: internal compiler error: virtual array basic_block_info[6]: element 3619816 out of bounds in set_up_bb_rts_numbers, at global.c:2274
 > 
 > Is this a known problem? I searched bugzilla for pr83 and found the closed report
 > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16841
 > 
 > Should this PR be reopened?

Maybe.  However, I didn't see it with yesterday's CVS head.  Exactly
what platform/gcc source version were you using?

Andrew.