Existence of Ada95 compiler for embedded 80x86?

Existence of Ada95 compiler for embedded 80x86?

[Greg Bond]

The GNAT port for DOS does not work reliably for delay statements, and has
no language support for interrupts. GNAT/RTEMS for 80x86 is vaporware (its
release has been postponed for well over a year now). I seem to recall a
newsgroup or mailing list posting a while back stating that GNAT had been
ported to Lynx (by Lynx?). I can't find any mention of it on their homepage
though. Did it ever happen?

Is there *any* Ada95 port, commercial or otherwise, suitable for embedded
80x86? Any such port should supprt the real-time annex, and the systems
programming annex.

--
* Greg Bond                         * Dept. of Electrical Eng.  
* email: bond@ee.ubc.ca             * Univ. of British Columbia      
* voice: (604) 822 0899             * 2356 Main Mall                 
* fax:   (604) 822 5949             * Vancouver, BC              
* web: http://www.ee.ubc.ca/~bond   * Canada, V6T 1Z4                                       

Re: Existence of Ada95 compiler for embedded 80x86?

[Robert Dewar]

Greg Bond said

<<The GNAT port for DOS does not work reliably for delay statements, and has
no language support for interrupts>>

We have no reports of delay not working in the DOS port, and it seems to
be working in all the tests we have run, so if you have an example of
it not working, you should submit a bug report!

Robert Dewar
Ada Core Technologies

Re: Existence of Ada95 compiler for embedded 80x86?

[Jerry van Dijk]

Robert Dewar (dewar@merv.cs.nyu.edu) wrote:

: We have no reports of delay not working in the DOS port, and it seems to
: be working in all the tests we have run, so if you have an example of
: it not working, you should submit a bug report!

I'm already on record saying I love GNAT, however:

the delay implementation of GNAT may work as intented by the RM, it is really
unsuitable for serious DOS programming.

This has been discussed often enough (and I've spend enough time building
work arounds) that I'm not inclined to yet again post samples.

If you feel differently, please prove this.

Jerry.

-- 
+----------------+-------------------------------+
| Jerry van Dijk | Consultant, Ordina Finance BV |
|    Team Ada    |        Haarlem, Holland       |
+----------------+-------------------------------+

Re: Existence of Ada95 compiler for embedded 80x86?

[Douglas Rupp]

Jerry van Dijk wrote:
> the delay implementation of GNAT may work as intented by the RM, it
> is really unsuitable for serious DOS programming.
> 
> This has been discussed often enough (and I've spend enough time
> building work arounds) that I'm not inclined to yet again post
> samples.
> 
> If you feel differently, please prove this.
> 
> Jerry.

Proof:

A search of DejaNews shows zero posts by you with a subject containing
the word "delay" in the last 18 months in the newsgroups comp.lang.ada
or comp.os.msdos.djgpp. 

20 Matches for Query:  delay

       Date   Scr        Subject              Newsgroup           Author
 
  1. 97/01/27 023 Ada DELAY statement accu comp.lang.ada      
jmotes@sed.redstone
  2. 97/01/28 022 Re: Ada DELAY statement  comp.lang.ada      
jhopper@erinet.com 
  3. 97/01/27 021 Re: Ada DELAY statement  comp.lang.ada      
evans@evans.pgh.pa.
  4. 96/12/17 020 Re: delay() problem      comp.os.msdos.djgpp
kccheng@hep2.phys.s
  5. 96/12/16 020 delay() problem          comp.os.msdos.djgpp Michael
Phelps <mor
  6. 96/12/16 020 Re: delay() problem      comp.os.msdos.djgpp Yamaha /
XYZZ <scri
  7. 96/12/14 020 delay() problem          comp.os.msdos.djgpp
kccheng@hep2.phys.s
  8. 96/05/22 018 Delay in Tasks           comp.lang.ada       Frank
Schneider <sc
  9. 96/04/19 018 Delay Before Programs Lo comp.os.msdos.djgpp
drewnan@pennet.net 
 10. 96/05/23 017 Re: Delay in Tasks       comp.lang.ada      
jsa@organon.com (Jo
 11. 96/05/22 017 Re: Delay in Tasks       comp.lang.ada       "Theodore
E. Dennis
 12. 96/05/01 017 Re: Delay Before Program comp.os.msdos.djgpp
jb3207@eclipse.co.u
 13. 96/04/24 017 Re: Delay Before Program comp.os.msdos.djgpp
drewnan@pennet.net 
 14. 96/04/23 017 Re: Delay Before Program comp.os.msdos.djgpp
eheft@sisko.dnaco.n
 15. 96/04/22 017 Re: Delay Before Program comp.os.msdos.djgpp
nicolas@dsys.ceng.c
 16. 96/02/20 017 GNAT for Amiga - DELAY s comp.lang.ada      
llewell@llewell.us.
 17. 95/10/04 017 Quickie, Delay delays be comp.os.msdos.djgpp
jonathn@teleport.co
 18. 95/09/28 016 problems with delay      comp.os.msdos.djgpp "Petr
Danicek, " <D
 19. 95/09/28 016 Re: problems with delay  comp.os.msdos.djgpp Andreas
Busse <andy
 20. 95/12/15 015 Delay 0.0 causing dispat comp.lang.ada      
dewar@cs.nyu.edu (R

Re: Existence of Ada95 compiler for embedded 80x86?

[Robert Dewar]

Jerry says

<<I'm already on record saying I love GNAT, however:

the delay implementation of GNAT may work as intented by the RM, it is really
unsuitable for serious DOS programming.

This has been discussed often enough (and I've spend enough time building
work arounds) that I'm not inclined to yet again post samples.

If you feel differently, please prove this.>>


Certainly when I say that delay works in DOS, I do mean that it works
according to the description in the RM. Not clear what else it might
mean. Whether or not the particular implementation choices are suitable
for a particular program is of cours not decided by that. Certainly a lot
of ppeople are doing serious programming with the DOS version of GNAT.
However, I don't think that anyone would expect the DJGPP implementation
to be a reasonable candidate for an ebmedded 80x86 (as per the subject
line).

If you simply mean that you need a finer resolution for delay in the
DOS port, that is of course easily achieved, but it is not something
for which we have had any customer demand so far.

GNAT/RTEMS

[csmith]

> GNAT/RTEMS for 80x86 is vaporware (its
> release has been postponed for well over a year now).

Does anybody knows the status of this port?
Thanks.

Re: GNAT/RTEMS

[Joel Sherrill <joel@merlin.gcs.redstone.army.mil>]

In article <3313A11C.2AD0@ccgate.hac.com>, csmith <csmith@ccgate.hac.com> writes:
>> GNAT/RTEMS for 80x86 is vaporware (its
>> release has been postponed for well over a year now).
>
>Does anybody knows the status of this port?
>Thanks.



First the RTEMS project has the same policy on announcement of release
dates that GNAT does.  We do not officially announce them.  I apologize
if there has been a misunderstanding regarding this.

Currently we are testing GNAT/RTEMS on the SPARC.  GNAT 3.09 has some RTEMS
support in it.  The primary deficiency is that 3.09 does not have full
support for building cross applications.   Most of these problems have
been addressed in the current GNAT and RTEMS developers source.  

The PowerPC is expected to be the next processor targetted.  Other processor
families will be added as there is customer interest. 

OAR and ACT are workign together to support this configuration

Joel Sherrill                             Sr. Computer Scientist
joel@OARcorp.com                  On-Line Applications Research
Ask me about RTEMS: a free    RTOS  Huntsville AL 35805
   Support Available                (205) 722-9985