[M3devel] m2tom3

[Daniel Alejandro Benavides D.]

Hi all:
and I see this:
http://labs.oracle.com/forest/COM.Sun.Labs.Forest.doc.see_97.paper_pdf.pdf

indicated that it was possible to either compile and cross-assemble or compile and run over its DEC-SRC Microkernel VMS code, at least at user-level.

Anyway, the main thing here is that this cross-ports like SOLGnu or Darwin are actually possible to run in one compilation and two code generations. That is better functionality or at least I guess so for those platforms as well. Of course the idea would be run in SPIN, but in some cases like SPINE it can run actually NT386GNU and NT386 or for instance run in a LAN a NT_ALPHA_GNU (SPINE) and ALPHA_SPIN code almost instantly (like if it would an user-level server) alternatively or somehow a crazy thing like that in IX86 likewise setting.
Thanks in advance  

--- El mié, 16/11/11, Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es> escribió:

> De: Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es>
> Asunto: Re: [M3devel] m2tom3
> Para: "Mark Wickens" <mark at wickensonline.co.uk>, "Mika Nystrom" <mika at async.caltech.edu>
> CC: m3devel at elegosoft.com
> Fecha: miércoles, 16 de noviembre, 2011 14:56
> Hi all:
> I think this is the way to get out of this
> "moral-technical" issue with the Unix interfaces, so get
> back to the old style (of course not brutal cross product),
> though we could that for the sake of further compatibility
> (it should be transparent ideally, like an Universal VM),
> but compiled natively (DEC SRC had the Ultrix VAX API as
> well, so for vax, openVMS you could do that surely).
> And at the end just EXPORT Unix.System() or UnixV, UnixVI,
> etc and that's all.
> OK, let me know, thanks in advance
> 
> --- El mar, 15/11/11, Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es>
> escribió:
> 
> > De: Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es>
> > Asunto: Re: [M3devel] m2tom3
> > Para: "Mark Wickens" <mark at wickensonline.co.uk>,
> "Mika Nystrom" <mika at async.caltech.edu>
> > CC: m3devel at elegosoft.com
> > Fecha: martes, 15 de noviembre, 2011 20:22
> > Hi all:
> > just forget that, it's done, the only thing we need is
> the
> > SUN Modula-2 compilers, and that's all folks :)
> > http://homepage.mac.com/dr2chase/resume.html
> > 
> > How enough hard this men worked but how much good
> fellows
> > they were (I can't see anything like this today in
> > industry).
> > 
> > Thanks in advance
> > 
> > --- El mar, 15/11/11, Daniel Alejandro Benavides D.
> <dabenavidesd at yahoo.es>
> > escribió:
> > 
> > > De: Daniel Alejandro Benavides D. <dabenavidesd at yahoo.es>
> > > Asunto: Re: [M3devel] m2tom3
> > > Para: "Mark Wickens" <mark at wickensonline.co.uk>,
> > "Mika Nystrom" <mika at async.caltech.edu>
> > > CC: m3devel at elegosoft.com
> > > Fecha: martes, 15 de noviembre, 2011 19:22
> > > Hi all:
> > > yeah my fault, but as I recall although I could
> > compile
> > > this years ago, the main issue is that it didn't
> even
> > work
> > > to feed the file to compile or so failed in
> therefore
> > the
> > > first character was read. So it never worked for
> real
> > for
> > > me.
> > > My best bet would export a SUN Unix syscall
> interface
> > > (since it tailored the SUN Modula-2 library)
> spec
> > like
> > > Sphinx in SPIN OS so an export of it for every
> > Modula-3
> > > program understands itself right (yeah it coudl
> take
> > the
> > > years spent from when I compiled until today, or
> so).
> > > If I were to start a new project perhaps would
> start
> > making
> > > Unix legacy and updated replacements.
> > > The other approach perhaps the most indicated is
> just
> > to
> > > wrap C on SPIN code to make it user level server.
> This
> > could
> > > make the thing (inf act there is FreeBSD server
> > already
> > > there just miss several calls, but nonetheless
> is
> > something
> > > there).
> > > BTW, I don' know which platform this compiler
> used to
> > work,
> > > so the main point would be emulate the real sys
> calls,
> > since
> > > they just emulated in Sun OS and Solaris,
> nothing
> > more. This
> > > has a potential utility for another project that
> > depends on
> > > this m2tom3 tool.
> > > Thanks in advance
> > > 
> > > --- El mar, 15/11/11, Mika Nystrom <mika at async.caltech.edu>
> > > escribió:
> > > 
> > > > De: Mika Nystrom <mika at async.caltech.edu>
> > > > Asunto: Re: [M3devel] m2tom3
> > > > Para: "Mark Wickens" <mark at wickensonline.co.uk>
> > > > CC: m3devel at elegosoft.com
> > > > Fecha: martes, 15 de noviembre, 2011 18:24
> > > > What I would do is:
> > > > 
> > > > 1. delete IMPORT TextF
> > > > 2. see what breaks, figure out the intent of
> the
> > > code,
> > > > recode it using Text
> > > > 3. check in the result to the CM3 repository
> so
> > no one
> > > has
> > > > to do it again!
> > > > 
> > > >    Mika
> > > > 
> > > > Mark Wickens writes:
> > > > >Guys,
> > > > >
> > > > >I found the m2tom3 translation utility
> here:
> > 
> > > > >http://freepages.modula2.org/downloads/m2tom3-2.03.tar.gz
> > > > >I attempted to compile it using cm3, but
> I
> > get an
> > > > unknown interface:
> > > > >
> > > > >msw at x60:~/m2tom3/m2tom3/src$ cm3
> > > > >--- building in ../LINUXLIBC6 ---
> > > > >
> > > > >unsupported m3_option value: "-O"
> > > > >new source -> compiling Standard.m3
> > > > >"../src/Analyzer/Standard.m3", line 25:
> > unable to
> > > find
> > > > interface (TextF)
> > > > >1 error encountered
> > > > >compilation failed => not building
> > program
> > > "m2tom3"
> > > > >Fatal Error: package build failed
> > > > >
> > > > >
> > > > >Has anyone heard of interface TextF?
> > > Alternatively, is
> > > > there a port of 
> > > > >this utility to cm3?
> > > > >
> > > > >I recently uncovered a machine-readable
> copy
> > of
> > > my
> > > > degree final year 
> > > > >project, a Meta Assembler, written in
> > Modula-2,
> > > and
> > > > wondered if I could 
> > > > >get it to compile using the utility to
> work
> > with
> > > > Modula-3.
> > > > >
> > > > >Regards,
> > > > >
> > > > >Makr.
> > > >
> > >
> >
>