Re: MVS380 on Linux in 10 steps...
- --- In email@example.com, Gerhard Postpischil <gerhardp@...> wrote:
>I had intended to update the C version of the
> kerravon86 wrote:
> > E.g. I'll get the disassemblies working to the
> > point where MVS is built from "source", but have
> > no intention of modifying MVS myself. It is best
> > that those with the hard-won skills do that. I'll
> > just make it easier for them by doing the obvious
> > and straightforward legwork.
> After your suggestion, I modified DISASM to let you specify the
> instruction set (OPTION S/370, etc.). I also added support for
> expanding DC C constants for lower case text (I was getting sick
> and tired of looking at DSSDUMP dumps, and needed a diversion).
> I still want to change the way C constants are handled (support
> continuation and explicit length), so let me know if you want it
disassembler to support that feature before
doing the mass disassemblies, but if you've
update the assembler version, then I think that's
This hasn't reached the top of my priority list
yet though. VM/380 has priority but I'm not
working on it currently. After that comes the C
compiler, as doors have been opened on that to
allow it to push forward more.
Specifically, in the last few hours I've got a
single Linux GCC 3.4.6 executable compiled with
PDPCLIB that is able to compile simple C programs,
which means that I should now be in a position
to do testing on my Windows box and on MVS, free
from Unix nonsense.
Driving forward with GCC to bring it up-to-date
allows other things like PL/1 to potentially be
enabled, and for whatever reason, there is a
paucity of people willing or able to drive GCC
forward, so that remains the most productive
niche for me to concentrate on.
Although I am mindful of the fact that a lack of
source hinders others with rare skills from
working productively too. But diassembling is
not a rare skill. It's an interesting journey. :-)