[comp.soft-sys.andrew] some more info on PL7 on SunOS4.1

jis@mtgzx.att.com (J Mukerji) (10/02/90)

As far as the kernel trap problem goes, a core image from it is on its
way to the Sun PAL. I should hear from him tomorrow. 

After much experimentation I have come to the conclusion that there is a
definite correlation between an error message that gets displayed in the
console and the subsequent coredump. The error message is something like
the following:

<messages>X error BadValue (integer parameter out of range for
operation) (2) major code 51 minor code 0 ignored [From host mtgzflw1] 
( 5:08:46 PM )

The Andrew application crashes soon after producing this message. The
actual point where it crashes is quite inoccuous as mentioned earlier.
That would seem to indicate that something is trashing a pointer or two
somewhere.

Has anyone seen this error message come out of Andrew? Where is it
coming from? Andrew? or the X server?

The X server that we are using is MIT X V11R4.

                            Jishnu Mukerji,  
                           jis@mtgzx.att.com, 
                           +1 201 957 5986,  
                        AT&T Bell Laboratories, 
                      MT 3K-423, 200 Laurel Ave., 
                           Middletown NJ 07748

janssen@parc.xerox.com (Bill Janssen) (10/02/90)

Excerpts from ext.andrew: 1-Oct-90 some more info on PL7 on Su.. J
Mukerji@mtgzx.att.com (1022+0)

> <messages>X error BadValue (integer parameter out of range for
> operation) (2) major code 51 minor code 0 ignored [From host mtgzflw1] 
> ( 5:08:46 PM )

> Has anyone seen this error message come out of Andrew? Where is it
> coming from? Andrew? or the X server?

This message is from the Xlib, called from ATK in either (guessing) xim
or xfont; it indicates a problem with a call to SetFontPath (code 51). 
Andrew has a bit of cleverness built in so that it adds the directory
containing the Andrew fonts to your font path in the X server if it is
not already there.  Of course, this breaks occasionally, as when you run
several different servers with different font binary formats, but it's
usually OK.

The only two integers in this call are the length of the request
(computed automatically by Xlib), and the number of directory strings in
the request.  I'd imagine that the number of strings is either negative,
or very large (zero seems to be a valid value).  You can find out by
running the program xscope between the ATK program and the xserver, and
watching for the SetFontPath call.

Bill