IA64 Java build problem

[Steve Ellcey]

> To see what is happening in jcf_parse(), run jc1 under gdb, break on
> fatal_error, and then go up and print the JCF.
> 
> Tom

Ok, I ran jc1 and I stopped at fatal_error.  Here is a backtrace and
then the value of jcf after going up to jcf_parse.  Hope it makes some
sense to you.

Steve Ellcey
sje@cup.hp.com

Breakpoint 1, fatal_error (
    gmsgid=0x4000000000cc4a60 "not a valid Java .class file")
    at /hpadl888/sje/svn.reg/src/trunk/gcc/diagnostic.c:577
577     {
(gdb) bt
#0  fatal_error (gmsgid=0x4000000000cc4a60 "not a valid Java .class file")
    at /hpadl888/sje/svn.reg/src/trunk/gcc/diagnostic.c:577
#1  0x400000000009b1e0 in jcf_parse (jcf=0x2000000000041280)
    at /hpadl888/sje/svn.reg/src/trunk/gcc/java/jcf-parse.c:1457
#2  0x40000000000a1860 in java_parse_file (set_yydebug=3704904)
    at /hpadl888/sje/svn.reg/src/trunk/gcc/java/jcf-parse.c:2104
#3  0x4000000000546f20 in toplev_main (argc=Variable "argc" is not available.
)
    at /hpadl888/sje/svn.reg/src/trunk/gcc/toplev.c:1043
#4  0x40000000000be9a0 in main (argc=31, argv=0x60000fffffffb6c8)
    at /hpadl888/sje/svn.reg/src/trunk/gcc/main.c:35
(gdb) up
(gdb) p jcf
$1 = (JCF *) 0x2000000000041280
(gdb) p *jcf
$2 = {buffer = 0x60000000001682b0 "", buffer_end = 0x6000000000168705 "",
  read_ptr = 0x60000000001682b8 "rmic.Messages",
  read_end = 0x6000000000168705 "", right_zip = 1, finished = 0,
  filbuf = @0x4000000000dcd400: 0x4000000000091930 <jcf_unexpected_eof>,
  read_state = 0x600000000014d720,
  filename = 0x6000000000132e80 "sun/rmi/rmic/Messages.class",
  classname = 0x6000000000132d50 "sun.rmi.rmic.Messages",
  zipd = 0x600000000014dab0, access_flags = 44975, this_class = 44975,
  super_class = 44975, cpool = {capacity = 0, count = 0, tags = 0x0,
    data = 0x0}}