🚀 go-pugleaf

> The gdb-3.91.4 available from labrea has a broken port to the MC88100 (it has
> several typos in the code, and probably doesn't interface to the new stuff...

The gdb-3.91.xx versions have broken ports to just about EVERYTHING.
These are "pre-alpha-test" releases being sent out to people who are
doing specific work (features or ports) on them.  It's free software,
so you are free to pick it up and see what it is, but don't expect it
to work or anything...

If you want to sign up to port it to a particular machine, please send
me mail (gnu@cygnus.com).  I'd love for gdb-4.0 to come out with
tested, working support for all the same machines gdb-3.5 handled.

A few things to know about the new gdb if you're porting it:

    *	Config files have been separated by host and target machines.
	GDB now supports cross-debugging.  It's often hard to tell which
	things sit on the host side and which on the target side.
	Certainly disassembly, registers, and such are target side.
	Certainly ptrace and include file (e.g. tty) issues are host side.
	For other stuff, we're still figuring it out.

	The config files for hosts are in the xconfig subdirectory.
	Target configs are in tconfig.  Both of these files are basically
	pieces of Makefiles, that get glommed onto Makefile.dist.
	They contain the names of files that should be used for this
	host or target system (e.g. coffread.o if the target is a COFF
	system, solib.o if the target supports shared libraries, etc).

    *	gdb.texinfo is not up to date.  The ChangeLog is your best friend.
	If you send in ChangeLog entries that match your changes, I'll be
	even happier.

    *	GDB now only reads registers from the inferior or target when it
	needs to.  This speeds things up significantly, especially when
	single-stepping continuously (watchpoints) or when running over
	a network or serial link.  Your machine interface code that
	handles fetch_inferior_registers() or remote_fetch_registers()
	should support reading only the register requested, if it's
	cheaper than getting all the registers.

    *	I will be reorganizing the code that reads executables, core
	files, and symbols soon.  I will also be rehacking the
	inferior-versus-remote stuff into a transfer vector.  Please
	keep your changes in these areas to a minimum.

    *	I am trying to get gdb to pass lint (and Saber C, and gcc
	-Wall), though it is still far from that now.  Please cause
	your changes to move it closer, rather than further, from
	these goals.  Examples are declaring function return types
	in the appropriate .h file so that the type is known when
	the function is called; or casting NULLs passed to functions
	to the proper pointer type.

If you encounter any other surprises between gdb-3.5 and gdb-3.91.xx,
let me know so I can document them.

	John Gilmore