🚀 go-pugleaf

As I have mentioned in previous notes, I am working on porting gdb to the
Ncube.   Specifically, I am using gdb as a cross debugger.  It runs on a sparc
but debugs programs out on the Ncube.

In order to implement this, I simply had to rewrite the ptrace() interface to
sends messages to the OS running on the Ncube.  The Ncube OS allows full
debugging of processes in this manner.  That much was easy.

Currently, the biggest road block I have hit is the differing endian-ness of
the two machines.  There doesn't seem to be much support internal to gdb 3.5 to
allow this.   I have had to go into routines like read_var_value(),
value_from_register(), and value_at() and put macros into the code which
do byte-swapping as appropriate for the data type being read.

Still, there are subtle problems.  So much so that I am considering porting gdb
to run on the Ncube flat out.  Ncubes do not run straight Unix, so this project
is a challenge in itself.

My questions are:

1) Are you interested in my changes to support cross-debugging in a
heterogeneous environment?

2) Do you have plans to do the above yourself?  Is it in my interest to wait?
Timeframe?

3) Do you think the whole cross-debugging effort is a crock and I would be
better to bite the bullet and port gdb to run on Ncbue directly?

thanks
--bill