🚀 go-pugleaf

In reply to Steven Orr's message of the 15th:

With the COFF patch that I supplied for gas-1.38 it's debugging output can be
thought of as encapsulated dbx.  It hides some of the debuging information
in structure entries ( just in case your version of COFF doesn't support
all of the various entries needed for debugging )

I assume that gdb was working ok for you in gas-1.37, and it was
supporting the 'native' debugging output format.

I must admit that I'm working on getting gdb-3.5 running on my 68k V.2 box
at the moment.

The best rule of thumb I can think of as far as choosing an assembler
would be as follows ( in a COFF environment ):

a)  old 68k (or other CPU's) V.x systems with assemblers that don't
    support the 68020 ( like the plexus ) or are just broken go with the
    patch I provided.  This makes getting gcc running *much* easier than
    trying to deal with getting gcc running.  Use the patch kit I made
    available.  It has the ablity to order the symbol table for ld's use,
    and support BSD style relocation information that some COFF systems
    use.

b)  Probably most SystemV.3 systems can use the patch to gas ( although
    the last version it supported was 1.37, not sure, and the changes to
    make it work on 1.38.1 should be minor ) provided by Loic Dachary and
    is available on the dg archive machine.  This supports outputting
    debugging information in the systems 'native' format ( if your system
    supports dbx )

c)  Use the system assembler, probably ok for most '386 platforms and
    newer 68K platforms.

I should have a copy of the patches to gdb-3.5 up on the archive site at
cs.umb.edu hopefully early next week.  Will post when that happens ( or
you can pick up the original COFF patches on archive.ohio-state.edu )

-randy