java/1745: compiler segfault on bogus code.

[ptf at ftel dot co dot uk]

>Number:         1745
>Category:       java
>Synopsis:       compiler segfault on bogus code.
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          ice-on-illegal-code
>Submitter-Id:   net
>Arrival-Date:   Tue Jan 23 08:56:02 PST 2001
>Closed-Date:
>Last-Modified:
>Originator:     ptf@ftel.co.uk
>Release:        unknown-1.0
>Organization:
>Environment:
Reading specs from /opt/gcc-snap/lib/gcc-lib/i686-pc-linux-gnu/2.97/specs
Reading specs from /opt/gcc-snap/lib/gcc-lib/i686-pc-linux-gnu/2.97/../../../libgcj.spec
rename spec lib to liborig
rename spec startfile to startfileorig
Configured with: ../gcc/configure --prefix=/opt/gcc-snap --enable-languages=c,c++,java
gcc version 2.97 20010122 (experimental)
>Description:
Java compiler segfault.

In a large project I had gcj segfault when trying to
initialise a member with a constant declared in a
derived class (odd, perhaps, but it works). I'm not
sure if this is the same bug because I couldn't
reproduce it in a small example until I managed to
get the name on the constructor of "c2" wrong.

The code *is* totally bogus but the compiler shouldn't
segfault.
>How-To-Repeat:
unpack the following file, cd to bug/p1 and type "make"

begin 664 bug.tar.gz
M'XL(`(VR;3H``^W6WTK#,!0&\-[:IPA>.<$N)_T34+P0;[Q0$.8+I%E6NK5=
M:#,5Q'?W;.O&5)R"M(J>WTU6NJW9OB;]TD4V]#K&><1E'.,H>!A'."+1CFL>
MER(,0Y'$4GH<>`B)Q^*N)[:T:)RJ&?.LF^Q]7Z9*T_0QH7ZEF+^%;F^!K^<?
M`I?1,O]$",J_#VW^&H*INE?=7`/SY$D4?9@_)&*3/P:/]P*`3+C'>#?3>>V?
MYV^5GJG,,`MGOI^7=EX[9D5PC$=VD1:Y9KI03<,TL">?(0U'@_;EL[\:;#UW
M1CLS9B-7YU7&2E.FIF;GRR_2(EA4>&?EA4H+<^8_^S_]B\FN=OW?J)F9Y(7I
MY!J?K7\\N=W_@>->`$)$DM9_'U11G/H'E]<7H]'MQ=W5>1"PX=RZ8:;U25,I
M.TSS"@^F[`1W@O5#@E;P'[):_^+7]#\NPWC=_ZC_]Z+-'Q_3/]?_A(C?Y`]2
M`NW_?=CV/['3_R#0\*X`"F8>G:G&^[O@^B/XGSH<)GFEBDTKW*F!6`TK\]">
A.#K<.7,XH(I(""&$$$(((8000@@AA'S;"[YA//0`*`
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: