garnett@mcs.anl.gov (08/09/90)
The following error dump was produced by invoking xlc with -DNAME1="..."
-DNAME2="..." etc. flags with the ... replaced by a large number of strings
(but not an unreasonable number since other machines can handle this
e.g. sun3, sun4, sequent, etc.). This error was produced on a model 320
running AIX 3.1 9021.
----------message starts here----------------
*** EXCEPTION in exception handler -- SIGNAL 11, AIX Signal - not a TRAP at: 100018d0
Registers:
3d227273 2ff7f290 200649f4 7375622c 2ff7f2c8 2ff7f2cc 2ff7f2c8 7375622c
00000000 00310000 82028000 4000d0b0 00000000 2ff97de8 0000d0b0 2ff7f834
deadbeef deadbeef 2003ea00 deadbeef deadbeef deadbeef deadbeef deadbeef
deadbeef 2007a440 00000083 7375622c 00000000 00000000 2ff7f3a0 2ff7f4f8
cr: lr: ctr: xer: mq fpscr: msr:
24226034 100012e8 00000000 00000010 00000000 82028000 0000d0b0
Traceback:
Procedure name line offset address backptr dbtable
$LOCATE 0 00000008 100018d0 2ff7f290 10001b04
RTETRP 9999 0000270f 100012e8 2ff7f360 2007a428
Original exception was SIGNAL 11, AIX Signal - not a TRAP at: 100018d0
Procedure name line offset address backptr dbtable
xlc: 1501-224 fatal error in /usr/lpp/xlc/bin/xlcentry: signal 11 received
-----------message ends here------------------
The same error (but with less extensive diagnostics) is produced when
cc is used instead of xlc. The error goes away when fewer -DNAME="..."
flags are used. I am hoping that there is some compiler flag
(I didn't see one when looking through info) that will increase the
amount of #defined data the C compiler can handle. Thanks for any
assistance.
--
John Garnett
Argonne National Laboratory
garnett@mcs.anl.gov Mathematics and Computer Science Division
Argonne, Illinois