[comp.lang.modula2] is modula2 available

Jason.Kankiewicz@f345.n109.z1.fidonet.org (Jason Kankiewicz) (07/07/90)

Hey Pat, what's the latest version of the Benchmark Compiler?  Have you  
ever gotten any upgrades from Avant-Garde?  Thanks. 
Oh yeah, do you know if there are any library modules for the 
manipulation of 3-dimensional objects in the public domain?



--  
uucp: uunet!m2xenix!puddle!109!345!Jason.Kankiewicz
Internet: Jason.Kankiewicz@f345.n109.z1.fidonet.org

Pat.Speer@f180.n226.z1.fidonet.org (Pat Speer) (07/21/90)

Jason, sorry so long getting back to ya but you know how things go. Anyway I believe there is a new version of the compiler available (don't know its number though). Unfortunately I think the only way to get it is to buy the new (sorta new) debugger. Leon might send it to you if you called him but I guess it is not a major enough update to warrant getting ahold of all the registered users. As far as 3D stuff goes, I am not personally aware of anything but will take a look around (there must be something us




efull in all these old archive files I've got). See ya.


--  
uucp: uunet!m2xenix!puddle!226!180!Pat.Speer
Internet: Pat.Speer@f180.n226.z1.fidonet.org

Jason.Kankiewicz@f345.n109.z1.fidonet.org (Jason Kankiewicz) (07/24/90)

Thanks for replying, Pat!  Someone else responded to me from UseNet and  
told me that the 1.3 libraries for Benchmark are not being released  
because Avant-Garde's next official release will be the 2.0 libraries.   
He has an unofficial release of the 1.3 libraries and offered to E-Mail  
them to me but, alas, I'm not on UseNet.  Would you happen to know of  
another source for these?  I might be able to get them when I go back to 
school, and on the VAX again.  As for 3D libraries, I've managed to  
write  my own based upon the demo program "Bounce".  Take it easy.



--  
uucp: uunet!m2xenix!puddle!109!345!Jason.Kankiewicz
Internet: Jason.Kankiewicz@f345.n109.z1.fidonet.org

a665@mindlink.UUCP (Anthon Pang) (07/27/90)

Jason.Kankiewicz@f345.n109.z1.fidonet.org writes:
> Someone else responded to me from UseNet and
> told me that the 1.3 libraries for Benchmark are not being released
> because Avant-Garde's next official release will be the 2.0 libraries.
> He has an unofficial release of the 1.3 libraries and offered to E-Mail
> them to me but, alas, I'm not on UseNet.  Would you happen to know of
> another source for these?  I might be able to get them when I go back to
> school, and on the VAX again.

Since, the fidonet modula-2 echo is being posted to Usenet, technically, you
are on Usenet.

Another possible source, is to ftp the files from xanth.cs.odu.edu.  Or use the
mail server at bitftp@pucc.princeton.edu.  Or have someone post a uuencoded
archive to this newsgroup (echo).

Jason.Kankiewicz@f345.n109.z1.fidonet.org (Jason Kankiewicz) (07/31/90)

You again!  Thanks for your original response, sorry I could not reply  
directly.  My SysOp was impressed at how you were able to route the UUCP 
message directly to his board.  Was that automatic?  BTW, how is  
Avant-Garde planning on distributing the 2.0 version of the libraries?   
Are we going to have to upgrade to a whole new version of the  
development system?  Do you know if Leon's planning on supporting the 
LONGREAL datatype?  (Im all questions)  I've got the xanth file node  
address in my capture of your original message.  I'll try to get it when 
I get back to school.  Thanks.



--  
uucp: uunet!m2xenix!puddle!109!345!Jason.Kankiewicz
Internet: Jason.Kankiewicz@f345.n109.z1.fidonet.org

a665@mindlink.UUCP (Anthon Pang) (08/03/90)

> Jason.Kankiewicz@f345.n109.z1.fidonet.org writes:
> You again!  Thanks for your original response, sorry I could not reply
> directly.  My SysOp was impressed at how you were able to route the UUCP
> message directly to his board.  Was that automatic?  BTW, how is
> Avant-Garde planning on distributing the 2.0 version of the libraries?
> Are we going to have to upgrade to a whole new version of the
> development system?  Do you know if Leon's planning on supporting the
> LONGREAL datatype?  (Im all questions)  I've got the xanth file node
> address in my capture of your original message.  I'll try to get it when
> I get back to school.  Thanks.

Yes, mail routing was automatic...I'm not a "wizard" yet :)

I'll send Leon Frenkel email and ask.  IMO, I don't think Avant-Garde will send
out upgrade notices for 2.0...simply because it's only "officially" running on
the 3000.  To quote Mr Frenkel, "It's already the best development system
around"...take it for what you will (I call it boasting :).  But, I doubt
there's a new version of the compiler.

LONGREALs?  Good question...I'll ask.
By the time you get back to school, xanth may have undergone some changes...the
moderator has moved the archives (not sure which ones) to abcfd20.larc.nasa.gov
(128.155.23.64)...so try there also.

David.Vazquez@f1.n396.z1.fidonet.org (David Vazquez) (08/08/90)

 AP> running on
 AP> the 3000.  To quote Mr Frenkel, "It's already the best development system
 AP> around"...take it for what you will (I call it boasting :).  But, I doubt
 AP> there's a new version of the compiler.

     Can you ask if there will be a new version of the linker?  The current 
linker insists on importing the entire contents of an OBM file into the final 
executable, even if only one PROCEDURE is used in it (ex:  you IMPORT 
OpenScreen FROM Intuition; the compiler does just fine, but the linker brings 
in everything else).  This adds alot of unnecessary weight to programs.
     A friend of mine has found a way to get around this by writing his own 
bindings for some of the libraries.  He hasn't taught me yet, and this does 
seem a bit inconvinient when TDI's Modula-2 linker handled all this stuff 
correctly.
     Thanks in advance for a reply.
  
  -Dave.



--  
uucp: uunet!m2xenix!puddle!396!1!David.Vazquez
Internet: David.Vazquez@f1.n396.z1.fidonet.org

a665@mindlink.UUCP (Anthon Pang) (08/09/90)

> David.Vazquez@f1.n396.z1.fidonet.org writes:
>      Can you ask if there will be a new version of the linker?  The current
> linker insists on importing the entire contents of an OBM file into the final
> executable, even if only one PROCEDURE is used in it (ex:  you IMPORT
> OpenScreen FROM Intuition; the compiler does just fine, but the linker brings
> in everything else).  This adds alot of unnecessary weight to programs.

TDI has an "optimization" feature in its linker...aka "smart" or "intelligent"
linker...that removes dead code in modules.  This is the main reason the linker
is so slow in this mode.  I've found that the Benchmark compiler doesn't write
enough info into .obj files to make "smart" linking easy.  Anyways, I've
already asked Leon this question (last year)...so I already have a reply :)
> 3) The linker has not been optimized to reduce unused procedures.  However,
> carefuly (sic) testing has revealed that such a feature would not produce
> very signifcant savings in most cases.  For large programs the saving is a
> fraction of a percent and for small programs the saving depends heavily on
> which modules are use, since there is a lot of inter-module depedancies.
> (sic)

So, I guess this means no.  What a load of b.s.!

a665@mindlink.UUCP (Anthon Pang) (08/09/90)

Your friend is probably just typing in the CONST & TYPE declarations listed in
the DEFINITION module listings, in the manual, and ignoring the PROCEDURES.
For an IMPLEMENTATION module, he would use:

     IMPLEMENTATION MODULE modulename;
     BEGIN
     END modulename.

Most of the modules are Amiga specific...ie stubs to library routines. Since
all of the libraries are documented in the RKMs, he can use INLINE code for
calling library routines that he needs.  Since he's using Intuition, it's
already an Amiga specific program...so, there's no need to worry about
portability.

As an alternative, I disassemble the final load file (after adding Symbols to
it), strip or comment out the unused procedures.  Which is quite easy, as
unused procedures will show up as dc.b statements.  Then, I fix the
obvious [missed] optimizations (eg unnecessary tst's).  Choose my favorite
optimizing (eg branches) assembler--A68K--, and link it.

     Examples:           unoptimized         optimized      savings
          MovePointer     7444 bytes        4304 bytes       43 %
          WarpDrive       8704 bytes        2772 bytes       69 %
          Split           3768 bytes        1832 bytes       52 %

The savings will be even more noticeable when the 2.0 libraries come
out...AmigaDOS.DEF will have over 100 exported procedures...a good
reason to abandon "my" method, then.

a665@mindlink.UUCP (Anthon Pang) (08/10/90)

Jason.Kankiewicz@f345.n109.z1.fidonet.org writes:
> BTW, how is Avant-Garde planning on distributing the 2.0 version of the
> libraries? Are we going to have to upgrade to a whole new version of the
> development system?  Do you know if Leon's planning on supporting the
> LONGREAL datatype?

Just got an email response from Leon (on Compuserve).  He is planning a 2.0
version of Benchmark.  No release date, no distribution policy, yet.
And he didn't specifically mention the libraries or the compiler/linker tools.
Asked about LONGREAL support...there was no answer.  He's getting kind of
evasive.  Must have something to do with him being involved with supporting
M2Sprint (as Preferred Technologies), as well.

Oh, and he did mention that all of the files uploaded to his vendor section on
CI$, were freely redistributable.  I might send them to Fred Fish for inclusion
in the library...of course, I'd like to get more material...and make it a full
disk. :)