[ REQ 5090]: Bug report: Geomview.X11, DEC Alpha, OSF V3.0

[brandenberg at wrksys.enet.dec.com]

1.  During installation of the geomview-x11/geomview script in the BINDIR
    directory, the script is sed-edited with some incorrect substitutions.
    Basically, all the scripts target pointers want to be relative to
    BINDIR instead of the current GEOMROOT basing.

2.  On Alpha AXP OSF V3.0 (latest release), I am having problems displaying
    on a 24-plane framebuffer-type system.  When clicking the sample
    WebOOGL 'Still Life With Pear and Dodecahedron,'  gvx activates,
    generates the following two unaligned access messages, and draws
    a window filled with red with no trace of the geometry:

Unaligned access pid=17810 <gvx> va=1400090ac pc=1200a57e4 ra=1200a57a4 type=ldq
Unaligned access pid=17810 <gvx> va=1400090b4 pc=1200a57e4 ra=1200a57a4 type=ldq

    After awhile, the following message appears:

Shutting down weboogl...
chat'open(www.geom.umn.edu,80): Bad file number


    I haven't dove in with dbx to get any more details, as of yet...

3.  I'm using a beta release of perl 5.0 on my 3.0 system.  One result of
    doing this is some weirdness with the system header files processed
    by the h2ph script.  Attempting to require these will produce a
    message about an unknown symbol/function called 'SEC_BASE' due to
    a '#if SEC_BASE' test in the original header files.  My fix for
    perl scripts which attempt to require such header files, e.g. 
    sys/socket.ph, is to add the following subroutine in a suitable
    location:

		sub SEC_BASE {
			return 0;
		}

    This subroutine should be harmless in a non-OSF environment.

4.  I happen to be a consultant in DEC's Open3D graphics group (we
    produce the OpenGL product).  If you are planning an Alpha/OpenGL
    release anytime in the near future, I would be happy to test it
    and, if possible, give you some performance-improving suggestions...

Monty Brandenberg