[Coco] Looking for rma or r63 source.

[Jeff Teunissen]

Unfortunately, the contents of /SOURCECODE/C/RMA in my copy of EOU
beta 4 isn't actual source code, but a disassembly that has been
converted back into C. It appears to have been done by the same person
who "decompiled" c.comp, and looks to be the same origin as the "R63
source" found elsewhere.

Lines like "*(d00e6++) = *reg;" are a dead giveaway that the origin
was a disassembly. Having said that, however, the structure of the
code clearly has some relation to how the original code was
structured.

For example, when I look at the "reconstructed" source for c.comp
found in various "mw-cc.tar.gz" etc. archives around the net, I find
that the files correspond exactly in the number and rough contents to
the original source code, _in the order in which they were linked by
the original makefile_. All of the global (and static) variables were
collapsed into a single block at the beginning of the program, of
course, but the actual code to the functions is identifiable as
belonging to one of the functions from the original source. That is,
if you have the original source you can look at the reconstructed code
and compare how it does things.

The practical upshot of this is that we can tell that while all of the
function and variable names, structures, coding style, etc. in that
RMA source are wrong (and clearly so), we know that the original
source code really _was_ comprised of 9 files, and if we knew the
original filenames we could be sure that the contents of those files
corresponds to the function bodies of all the functions that comprised
the original assembler.

/SOURCECODE/C/RLINK isn't the original rlink source either, it's a
(mostly?) compatible rlink that was written in the 2000s by, I
believe, Boisy Pitre.

On Fri, Dec 6, 2019 at 2:46 PM Bill Nobel <b_nobel at hotmail.com> wrote:
>
> RMA and rlink are direct replacements for c.asm and c.link (they are the same).  I acquired these years ago from an archive I can’t remember (maybe from Princeton archive). The c.comp module preps the source code for these 2 to complete the compile.
>
> Sent from my iPhone
>
> > On Dec 6, 2019, at 1:26 PM, Stephen Fischer <SFischer1 at mindspring.com> wrote:
> >
> > Just those two?
> >
> > Likely as "RMA" and "RLINK" were in the "Development System" and not the original "C" Compiler package.
> >
> > Bill Pierce made a webpage with information about the "C" Compiler, a good part of that webpage is outdated!
> >
> > No Source is listed, where did you get "RMA" and "RLINK" source?
> >
> > Perhaps sending the source to Guillaume Major and suggesting that the "C" Compiler needs a directory of source.
> >
> > SHF
> >
> >
> >> On 12/6/2019 11:05 AM, Bill Nobel wrote:
> >> I have RMA and RLINK sources, but not R63.   It is the full C source code.  They may give you something to work with.  I do know that they need to be updated as they are still pre Y2K so make goofs up on checking dates and times for recompile. I also have the source to a updated make that adds some features the original make did not have.
> >> Sent from my iPhone
> >>> On Dec 6, 2019, at 12:38 PM, Stephen Fischer <SFischer1 at mindspring.com> wrote:
> >>>
> >>> The only possible location of the original "C" Compiler source might be Bill Pierce. But I remember a couple of disasters causing it to be lost.
> >>>
> >>> Try contacting him directly for information.
> >>>
> >>> Other than building a Time Machine and going to Des Moines about the last time I was there (Boy there were lots of signs offering some meat product on the way into town.) and breaking in and stealing the archive.
> >>>
> >>> The one attempt I made to obtain a computer that the source might have been that had passed hands was met here with such violence it was really unbelievable.
> >>>
> >>> People shoot themselves in the heart here sometimes, and innocent bystanders quite often.
> >>>
> >>> SHF
> >>>
> >>>> On 12/5/2019 10:24 PM, Jeff Teunissen wrote:
> >>>> Damn, I was hoping that would contain source instead of a disassembly.
> >>>> I have some compiler source, and once I get it producing "correct"
> >>>> code, I'd like to have the original assembler to put the results
> >>>> together.
> >>>> The code the compiler produces works, but has a slightly different
> >>>> (more efficient, actually) calling convention from the libraries we
> >>>> have available, so the binaries it produces don't run correctly;
> >>>> hacking the compiler to use the old calling convention seems like
> >>>> it'll be easier than modifying and recompiling all the old libraries,
> >>>> so that's my approach.
> >>>>> On Fri, Dec 6, 2019 at 1:04 AM Stephen Fischer <SFischer1 at mindspring.com> wrote:
> >>>>>
> >>>>> I must be on your kill list, I gave Guillaume Major the two programs you
> >>>>> asked for before and posted their location here.
> >>>>>
> >>>>> He placed r63 here:
> >>>>>
> >>>>> http://www.colorcomputerarchive.com/coco/Programming/Source/R63/R63%20(David%20Breeding).zip
> >>>>>
> >>>>> I made a statement that we have source for all of the "C" Compiler, I
> >>>>> will need to think and search my files for "rma".
> >>>>>
> >>>>> It may be Bill Pierce that has modified the single pass "C" Compiler who
> >>>>> has "rma" source.
> >>>>>
> >>>>> Yes, "r63" appears to have been written in "C".
> >>>>>
> >>>>> SHF
> >>>>>
> >>>>>
> >>>>>> On 12/5/2019 9:37 PM, Walter Zambotti wrote:
> >>>>>> Does anyone know if source exists for rma or r63?
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> I'm hoping it is written in C as this will make it more useful to me.
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> Walter
> >>>>>>
> >>>>>>
> >>>>>
> >>>>> --
> >>>>> Coco mailing list
> >>>>> Coco at maltedmedia.com
> >>>>> https://pairlist5.pair.net/mailman/listinfo/coco
> >>>
> >>> --
> >>> Coco mailing list
> >>> Coco at maltedmedia.com
> >>> https://pairlist5.pair.net/mailman/listinfo/coco
> >
> > --
> > Coco mailing list
> > Coco at maltedmedia.com
> > https://pairlist5.pair.net/mailman/listinfo/coco
>
> --
> Coco mailing list
> Coco at maltedmedia.com
> https://pairlist5.pair.net/mailman/listinfo/coco