[comp.lang.modula2] Fitted Software Tools Modula-2 Compiler Question

gnat@kauri.vuw.ac.nz (Nathan Torkington) (05/20/91)

I have been experimenting with the Fitted Software Tools Modula-2
compiler which is available from the SIMTEL-20 archives (and associated
mirror sites) for the IBM PC and clones.  I have a question, however:

	Are the libraries, as distributed, intact?

There only appears to be the "huge" memory model in the M2LIB and UTL files.

Can anyone point out what I'm doing wrong?

Nat.--
Nathan Torkington                           Contracting to CSC,
Internet: gnat@kauri.vuw.ac.nz              Victoria University Of Wellington,
WetComms: +64 4 853 881 (post-chronos)      P.O. Box 600, New Zealand.
This is not an official communication of any part of Victoria University.

"The less you know about home computers
 the more you'll want the new IBM PS/1."
 - Advertisment in the Edmonton Journal, Thursday, December 13, 1990'

jmh@coyote.datalog.com (John Hughes) (05/21/91)

In article <GNAT.91May20135152@kauri.kauri.vuw.ac.nz> gnat@kauri.vuw.ac.nz (Nathan Torkington) writes:
>I have been experimenting with the Fitted Software Tools Modula-2
>compiler which is available from the SIMTEL-20 archives (and associated
>mirror sites) for the IBM PC and clones.  I have a question, however:
>
>	Are the libraries, as distributed, intact?
>
>There only appears to be the "huge" memory model in the M2LIB and UTL files.
>
>Can anyone point out what I'm doing wrong?
>
>Nat.--
>Nathan Torkington                           Contracting to CSC,
>Internet: gnat@kauri.vuw.ac.nz              Victoria University Of Wellington,
>WetComms: +64 4 853 881 (post-chronos)      P.O. Box 600, New Zealand.
>This is not an official communication of any part of Victoria University.
>
>"The less you know about home computers
> the more you'll want the new IBM PS/1."
> - Advertisment in the Edmonton Journal, Thursday, December 13, 1990'

As far as I know, FST does include obj's for both memory models in the
distribution. The obj's have the same names, and the compiler is told
which to use with a DOS environment variable. At least that's how it
works for the later revs (2.0 and up). One does that by placing the lib
objects in seperate directories, and then putting something like:
SET M2LIB=C:!\M2LIB;C;!\M2LIBH (or 'L', as the case may be)
in the autoexec.bat file.

...hope this helps...

You can get a recent rev of the compiler from the FST conference on BIX.
I think that uunet and gatekeeper.dec.com also have it available for
FTP access.


-- 
|     John M. Hughes      | "...unfolding in consciousness at the            |
| datalog.com!moondog!jmh | deliberate speed of pondering."  - Daniel Dennet |
| jmh%coyote@noao.edu     |--------------------------------------------------|
| jmh%moondog@datalog.com | P.O.Box 43305, Tucson, AZ 85733    602-624-8008  |

Jon.Guthrie@p25.f506.n106.z1.fidonet.org (Jon Guthrie) (05/22/91)

 On a message of 19-May-91, Nathan Torkington (1:105/42.0) Said:

 > I have been experimenting with the Fitted Software Tools Modula-2

 > Are the libraries, as distributed, intact?

Yes.

 > There only appears to be the "huge" memory model in the M2LIB and UTL
 > files.

 > Can anyone point out what I'm doing wrong?

The "huge" model is a C-language aberration.  The FST Modula-2
compiler doesn't offer different memory models because Modula-2
doesn't need different memory models even for segmented architectures.
(C needs different models because it treats pointers like funny
integers so you need the option to trade generality for speed.)



--  
uucp: uunet!m2xenix!puddle!106!506.25!Jon.Guthrie
Internet: Jon.Guthrie@p25.f506.n106.z1.fidonet.org

Martin.Scharnke@comp.vuw.ac.nz (Martin Scharnke) (05/27/91)

In article <8184.283BD989@puddle.fidonet.org>,
Jon.Guthrie@p25.f506.n106.z1.fidonet.org (Jon Guthrie) writes:
|> 
|>  On a message of 19-May-91, Nathan Torkington (1:105/42.0) Said:
|> 
|>  > I have been experimenting with the Fitted Software Tools Modula-2
|> 
|>  > Are the libraries, as distributed, intact?
|> 
|> Yes.

Not really - see below

|> 
|>  > There only appears to be the "huge" memory model in the M2LIB and UTL
|>  > files.

That's right - if you want to use large you must recompile the source,
which is only available (or should only be available) to registered
users.

|> 
|>  > Can anyone point out what I'm doing wrong?
|> 

Then you can use the enviroment variable M2MODEL to set which model to
compile to, or specify it each time you compile something.

|> The "huge" model is a C-language aberration.  The FST Modula-2
|> compiler doesn't offer different memory models because Modula-2
|> doesn't need different memory models even for segmented architectures.
|> (C needs different models because it treats pointers like funny
|> integers so you need the option to trade generality for speed.)
|> 

Wrong - it is brought about by the PC architecture. It is applicable to
a wide variety of programming langauges - all that don't just execute in
the tiny model (like BASIC).

So in Modula 2, C, ASM, etc the model makes one heck of a difference.
 
|> 
|> --  
|> uucp: uunet!m2xenix!puddle!106!506.25!Jon.Guthrie
|> Internet: Jon.Guthrie@p25.f506.n106.z1.fidonet.org

 _                                 ||  martin@comp.vuw.ac.nz
' )--,--,         _/_              ||  Victoria University
 /  /  / __   __  / . _            ||  Wellington, New Zealand
/  /  (_(_/|_/ (_(_(_/ )           || 
-- 
 _                                 ||  martin@comp.vuw.ac.nz
' )--,--,         _/_              ||  Victoria University
 /  /  / __   __  / . _            ||  Wellington, New Zealand
/  /  (_(_/|_/ (_(_(_/ )           || 

jmh@coyote.datalog.com (John Hughes) (05/28/91)

In article <1991May26.222219.23677@comp.vuw.ac.nz> Martin.Scharnke@comp.vuw.ac.nz (Martin Scharnke) writes:
>
>In article <8184.283BD989@puddle.fidonet.org>,
>Jon.Guthrie@p25.f506.n106.z1.fidonet.org (Jon Guthrie) writes:
>|> 
>|>  On a message of 19-May-91, Nathan Torkington (1:105/42.0) Said:
>|> 
>|>  > I have been experimenting with the Fitted Software Tools Modula-2
>|> 
>|>  > Are the libraries, as distributed, intact?
>|> 
>|> Yes.
>
>Not really - see below
>
>|> 
>|>  > There only appears to be the "huge" memory model in the M2LIB and UTL
>|>  > files.
>
>That's right - if you want to use large you must recompile the source,
>which is only available (or should only be available) to registered
>users.

Well... er, no. The latest rev's of the compiler package (2.0 and up) come
with the pre-compiled objects in two (2) archived files, which are them-
selves embedded in the library object archive file. They're called, naturally,
huge.arc and large.arc. If you want the library sources, then that's were
the registration fee comes in - but a *complete* distribution comes with
everything one needs to set the model selection variable to either huge or
large and have things compile correctly right out of the box.

[stuff about the model selection variable deleted...]

It would appear that what SIMTEL has is not a complete distribution package.
Anyone within the conUS (or anyone with a hefty phone budget) can call the
FST bulletin-board system (no sign-up silliness) and get the latest
"shareware" distribution. The files areas of the system which contain the
library sources are restricted to registered owners. The phone number is:
(214) 517-4629, and 300/1200/2400 N81 will work.


-- 
|     John M. Hughes      | "...unfolding in consciousness at the            |
| datalog.com!moondog!jmh | deliberate speed of pondering."  - Daniel Dennet |
| jmh%coyote@noao.edu     |--------------------------------------------------|
| jmh%moondog@datalog.com | P.O.Box 43305, Tucson, AZ 85733    602-624-8008  |