Re: Problems with gdb on Sun Blade 1000

["Jay A. St. Pierre" <Jay dot St dot Pierre at Colorado dot EDU>]

On Fri, 18 Oct 2002, Andrew Cagney wrote:

> > I have been completely unsuccessful in getting gdb to work on the
> > 2 Blade 1000's and 2 Blade 2000's we have.  It does work on
> > another Blade 1000 in a different cluster, but when I copy over
> > the entire gnu toolset (/usr/local/gnu) from that cluster to
> > ours, it still doesn't work.  Obviously there is some
> > configuration or environment difference, but I haven't been able
> > to find it.
> > 
> > Please let me know if you have any ideas.
> 
> Can you compile the GDB on that blade for that blade?
> 
> If I undestand you correctly, you've built a GDB on another
> machine and are then trying to use it on the blade.  Problems
> -> something different in your shared libraries or kernel?

I have compiled gcc and gdb on the machine in question, but I 
have the same results (in fact this is what I did first).

I have compared the list of dynamically linked libraries for gdb 
and the "hello world" binaries, and there are a few with 
different checksums between the two systems.  In particular:

/usr/lib/libcurses.so.1
/usr/lib/libsocket.so.1
/usr/lib/libnsl.so.1
/usr/lib/libdl.so.1
/usr/lib/libc.so.1

Although I've never figured out an easy way to figure out what 
sun patches these binaries actually came from, but the machine 
I'm trying to run on had a recommended patch cluster installed 
recently, whereas the other one has had only security patches 
over the past year.

> Another experiment is to build the GDB statically and then try it.

Good idea, I will try that next.

-Jay