Subject: segfaults on jmw

(This message is being sent to all CPSC 102 Sec 1 students )

Two diligent students have reported to me that after submitting
their mp1, they tried to compile and run it on jmw.  In both
cases it compiled ok, but then when they tried to run it it
segfaulted.

Being diligent students, when it segfaulted, they tried to use
gdb to debug it but were dismayed to find that gdb didn't work
due to a problem with missing libcurses.so.5 or thereabouts.

I have verified that gdb doesn't work and have requested that
the sysadmins fix it. I don't know if I will have any luck any
time soon with that since they are working on the more important
problem of making the file server work :-(

Nevertheless, if this happens to YOU it DOES mean that your program is
BROKEN.  The cause is almost certainly a "loose" pointer because
the Sun Solaris (jmw and spider systems) environment is much better
at segfaulting broken pointers than is Linux.

Because gdb is broken, the ONLY way to track this down is with
diagnostic fprintf(stderr, ) which is something of a pain but is necessary
in many real world domains.  To do this in a sane manner change the
output image size (pixeldim) to something sane like (80 x 60) or
even (8 x 6).

Hopefully, the sysadmins will fix gdb soon. Until they do I will
inflict minimal penalty on programs that don't run on Solaris
but you need to find and fix the problem because in Linux its
more likely to produce BROKEN IMAGES which are even HARDER to
debug.

Most likely causes are:

(1) uninitialized pointer variables:
    use
    tplane_t *tpln = NULL;
    and not
    tplane_t *tpln;

(2) bad pointer arithmetic (as in camera_store_pixel)
    being "off by one row"