gnat question...

gnat question...

[gregfisher]

we are porting from the Sun Workshop C/C++ compiler and the Verdix Ada
compiler to GCC C/C++ and GNAT.

Platforms include Solaris Sparc and Solaris X86.

According to the GNAT readme files, the GNAT binaries are built on
GCC281.  We want to use GCC295 for our C/C++ code.

Our code links heavily between Ada and C using Ada pragmas.

Does anyone see a problem using the GNAT GCC281 binaries when linking
with GCC295 compiled objects?

Or, should we recompile GNAT binaries using GCC295?

Thanks in advance

Greg Fisher


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: gnat question...

[Tarjei T. Jensen]

gregfisher@my-deja.com wrote in message <8muf6l$m0t$1@nnrp1.deja.com>...
>we are porting from the Sun Workshop C/C++ compiler and the Verdix Ada
>compiler to GCC C/C++ and GNAT.
>
>Platforms include Solaris Sparc and Solaris X86.
>
>According to the GNAT readme files, the GNAT binaries are built on
>GCC281.  We want to use GCC295 for our C/C++ code.
>
>Our code links heavily between Ada and C using Ada pragmas.
>
>Does anyone see a problem using the GNAT GCC281 binaries when linking
>with GCC295 compiled objects?


You should seach for previous posts on this topic in this newsgroup at
http://www.deja.com/

>Or, should we recompile GNAT binaries using GCC295?


This is not supposed to work.

I would expect that it would be a good idea to become a support customer and
get early access to the version of gnat which will use gcc 3.x or gcc.2.9x
(don't remember which one they will be using).


Greetings,

Re: gnat question...

[Robert Dewar]

In article <8n0851$78u6@news.kvaerner.com>,
  "Tarjei T. Jensen" <tarjei.jensen@kvaerner.com> wrote:
>
> gregfisher@my-deja.com wrote in message
> >Our code links >heavily between Ada and C using Ada pragmas.
> >
> >Does anyone see a problem using the GNAT GCC281 binaries when
> >linking with GCC295 compiled objects?
>
> You should seach for previous posts on this topic in this
newsgroup at
> http://www.deja.com/
>
> >Or, should we recompile GNAT binaries using GCC295?
>
> This is not supposed to work.

It definitely does NOT work, significant changes are required
to both GNAT and to GCC for GNAT to compile successfully with
GCC 2.9x. You won't get even the simplest programs running
successfully without a lot of fiddling, and to get a reliable
port, the only practical thing is to wait for the work to get
completed at Ada Core Technologies.

> I would expect that it would be a good idea to become a
> support customer

Well I certainly will not disagree with that suggestion :-) and
in fact this is typical of the sort of problem we can help
with very effectively.

> and get early access to the version of gnat
> which will use gcc 3.x or gcc.2.9x
> (don't remember which one they will be using).

But not quite for the above readon :-)
Although that might be the preferred approach in the future,
it is premature to suggest that right now, and the problem
is easily solved in the context of the current release of
GNAT Professional.

Robert Dewar
Ada Core Technologies


Sent via Deja.com http://www.deja.com/
Before you buy.

GNAT Question

[chris.danx]

Hi,
    Does anyone know if i can use GNAT to do OS work?  I don't mean writing
apps but the OS itself.  Currently just starting work on it and we're
implementing virtual FS support as a file on disk.  We're doing it in C via
DJGPP but i hate C, like many on this group i gather.  I know of AdaOS and
have been to the site for it.  The info seems to suggest that you can't use
GNAT, and the development of an Ada compiler for this kinda thing is needed.

Why can't i use Ada?  Everything would be great and falt free, well free
from C anyway.  Can I link DJGPP with Ada?  That might be a temp solution.
I thought GNAT was modified C backend so what would get in my way.

I wouldn't have to edit autoexec.bat, reboot, do some C work, edit
autoexec.bat and reboot, do some Ada work, etc...


Thanks
Chris Campbell

Re: GNAT Question

[Ted Dennison]

In article <7wLt6.19470$bL.1855137@news6-win.server.ntlworld.com>, chris.danx
says...
>    Does anyone know if i can use GNAT to do OS work?  I don't mean writing
..
>from C anyway.  Can I link DJGPP with Ada?  That might be a temp solution.

There was a DOS/DJGPP version of Gnat one upon a time. Its probably still
floating around somewhere. The most recent version I could find was 3.10p at
ftp://cs.nyu.edu/pub/gnat/private/old/ez2load/3.10/ . I don't know if it can
build programs that don't need DOS though. I'd be really interested in finding
out myself. If you find out the answer, please tell me here or at my telepath
address. 

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT Question

[Gisle Sælensminde]

In article <EYNt6.1986$94.2845@www.newsranger.com>, Ted Dennison wrote:
>In article <7wLt6.19470$bL.1855137@news6-win.server.ntlworld.com>, chris.danx
>says...
>>    Does anyone know if i can use GNAT to do OS work?  I don't mean writing
>..
>>from C anyway.  Can I link DJGPP with Ada?  That might be a temp solution.
>
>There was a DOS/DJGPP version of Gnat one upon a time. Its probably still
>floating around somewhere. The most recent version I could find was 3.10p at
>ftp://cs.nyu.edu/pub/gnat/private/old/ez2load/3.10/ . I don't know if it can
>build programs that don't need DOS though. I'd be really interested in finding
>out myself. If you find out the answer, please tell me here or at my telepath
>address. 

The GNAT for DOS team at www.gnuada.org seems to have newer ports around.
See their homepage. These ports are not supported or maintained by ACT. 
They seems to have an alpha release of 3.13p and a stable relese of 3.12p.

--
Gisle S�lensminde ( gisle@ii.uib.no )   

With sufficient thrust, pigs fly just fine. However, this is not
necessarily a good idea. It is hard to be sure where they are going
to land, and it could be dangerous sitting under them as they fly
overhead. (from RFC 1925)

Re: GNAT Question

[Stefan Skoglund]

Gisle S�lensminde wrote:
> The GNAT for DOS team at www.gnuada.org seems to have newer ports around.
> See their homepage. These ports are not supported or maintained by ACT.
> They seems to have an alpha release of 3.13p and a stable relese of 3.12p.

GRUB supports loading of different binaries including Linux
and also GNUMach. It should be possible to generate some generic binary
which GRUB can bootstrap for you.

Look for info about GRUB.

Re: GNAT Question

[Hans-Olof Danielsson]

I would suggest a look at the MaRTE OS, an OS mostly written in Ada at
http://ctrpc17.ctr.unican.es/marte.html . Maybe you can reuse some of it or
better, add a virtual FS to it  %>P

Regards/HOD

----- Original Message -----
From: "chris.danx" <chris.danx@ntlworld.com>
Newsgroups: comp.lang.ada
To: <comp.lang.ada@ada.eu.org>
Sent: Tuesday, March 20, 2001 5:25 PM
Subject: GNAT Question


> Hi,
>     Does anyone know if i can use GNAT to do OS work?  I don't mean
writing
> apps but the OS itself.  Currently just starting work on it and we're
> implementing virtual FS support as a file on disk.  We're doing it in C
via
> DJGPP but i hate C, like many on this group i gather.  I know of AdaOS and
> have been to the site for it.  The info seems to suggest that you can't
use
> GNAT, and the development of an Ada compiler for this kinda thing is
needed.
>
> Why can't i use Ada?  Everything would be great and falt free, well free
> from C anyway.  Can I link DJGPP with Ada?  That might be a temp solution.
> I thought GNAT was modified C backend so what would get in my way.
>
> I wouldn't have to edit autoexec.bat, reboot, do some C work, edit
> autoexec.bat and reboot, do some Ada work, etc...
>
>
> Thanks
> Chris Campbell
>
>
> _______________________________________________
> comp.lang.ada mailing list
> comp.lang.ada@ada.eu.org
> http://ada.eu.org/mailman/listinfo/comp.lang.ada

Re: GNAT Question

[(null)]

In article <7wLt6.19470$bL.1855137@news6-win.server.ntlworld.com>,
chris.danx <chris.danx@ntlworld.com> wrote:
>Hi,
>    Does anyone know if i can use GNAT to do OS work?  

Yes.  

Since you'll be working underneath the application/kernel boundary
you won't be able to use any parts of the gnats's built-in runtime
libraries that depend on the underlying OS.  You can either write
your own code to implement those parts of the runtime, or do without.
If you choose to do without (a reasonable choice for many situations)
you'll want to use 'pragma No_Run_Time'.  Look in the "GNAT User's 
Manual".



>Currently just starting work on it and we're
>implementing virtual FS support as a file on disk.  We're doing it in C via
>DJGPP but i hate C, 

BTW, If I were you, I'd use something other than DOS to do
the development work on.  Compiling in Unix and testing on a PC
simulator (such as Bochs) should help things go a lot smoother



>The info seems to suggest that you can't use
>GNAT, and the development of an Ada compiler for this kinda thing is needed.

Actually the reason they are building a compiler from scratch is
because the founder of that project (Nick Roberts) wants to write
an Ada compiler.  The reason he wants to write an operating system
is because he doesn't think any existing operating system nicely
supports Ada compilers.

In other words, he is creating the operating system to support the compiler; 
his compiler is not needed to write an operating system.  


Just curious, does anyone out there know about how many man-years
it might take to write an Ada compiler?  How long did it take for gnat?


-- 
=======================================================================
 Life is short.                  | Craig Spannring 
      Bike hard, ski fast.       | cts@internetcds.com
 --------------------------------+------------------------------------

Re: GNAT Question

[chris.danx]

> >Currently just starting work on it and we're
> >implementing virtual FS support as a file on disk.  We're doing it in C
via
> >DJGPP but i hate C,
>
> BTW, If I were you, I'd use something other than DOS to do
> the development work on.  Compiling in Unix and testing on a PC
> simulator (such as Bochs) should help things go a lot smoother

I would do if i could.  My unfamiliarity with Unix means i can't really do
things this way.  Plus my OS development partner seems to like windows.  I'd
really like to do the OS in Ada, i'm afraid i might be stuck with C.  Don't
know if i can persuade him to change to Ada.  Don't know if he knows any
Ada, come to think of it.

I'm implementing the BFS file system as a file on disk {maybe i mentioned it
before} and was going to do it in Ada first, then once i'm sure it works
re-write the thing in C.  Maybe not a good idea.  BFS is a nightmare anyway!
Does GNAT support writing raw binary data to disk?  I really need to know
this!

> >The info seems to suggest that you can't use
> >GNAT, and the development of an Ada compiler for this kinda thing is
needed.
>
> Actually the reason they are building a compiler from scratch is
> because the founder of that project (Nick Roberts) wants to write
> an Ada compiler.  The reason he wants to write an operating system
> is because he doesn't think any existing operating system nicely
> supports Ada compilers.
>
> In other words, he is creating the operating system to support the
compiler;
> his compiler is not needed to write an operating system.

I must have picked things up wrong.  Maybe i'm also picking this up wrong
but i looked at the project and saw that some of it wasn't planned to be
written in Ada.  I know the bootstrap needs Assembly, but i'd think (almost)
everything else could be done in Ada.  I'm probably wrong about this aswell.


Thanks for the info,
Chris Campbell

p.s. now we just got to figure out what to call the OS!  ARES was suggested
but it sounds too much like ARSE as has already been pointed out on
alt.os.development.  We need suggestions.  Something clever or funny.  Any
suggestions would be appreciated.

Re: GNAT Question

[Ted Dennison]

In article <eIPt6.13980$Q4.2571687@news2-win.server.ntlworld.com>, chris.danx
says...
>p.s. now we just got to figure out what to call the OS!  ARES was suggested
>but it sounds too much like ARSE as has already been pointed out on
>alt.os.development.  We need suggestions.  Something clever or funny.  Any
>suggestions would be appreciated.

The toy OS we used in my OS class was called XINU, as that is UNIX spelled
backwards. If yours is going to be DOS-like, perhaps you should call it SOD. ;-)

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT Question

[chris.danx]

Like it, but it's going to be full blown distributed micro-kernel with nifty
GUI and stuff.

One thing i think we're going to implement is pie-menus.  Has anyone
implemented them in Ada?  I've only used one pie-menu system in Birth of the
Federation -- maybe we could call the OS 'Kirk' or 'Mad Womanizing
Chauvanistic Captain in Yellow who I'm really quite jealous of' -- and it
was really good.  I want to replace linear menus with pie menus in the
system, maybe we should provide the choice.  Don't want 'how does it work
syndrome'.


Live long and prosper,
Chris Campbell

-- anyone know the proper reply to 'Live long and prosper'.  I've forgotten
but i think it's 'Kah'plagh'.

Re: GNAT Question

[Ted Dennison]

In article <qzQt6.14157$Q4.2630301@news2-win.server.ntlworld.com>, chris.danx
says...
>
>One thing i think we're going to implement is pie-menus.  Has anyone
>implemented them in Ada?  I've only used one pie-menu system in Birth of the
>Federation -- maybe we could call the OS 'Kirk' or 'Mad Womanizing

After the reception that BotF's interface got, I'm amazed that anyone is willing
to repeat any part of it. 

As for hand-writing widgets and directly accessing disks, Ada really shouldn't
be an issue. There's no super-duper featue in C that Ada doesn't have an analog
to, and anything in an external C library can be accessed via a "pragma import
(C,...)". The biggest issue is typcially deciphering complex C macros.

>-- anyone know the proper reply to 'Live long and prosper'.  I've forgotten
>but i think it's 'Kah'plagh'.
If that's the rough Vulcan equivalent to "Up your nose with a rubber hose,
Kotter", then its probably about right. :-)

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT Question

[chris.danx]

I didn't say the BotF interface was great i just said it was my first
experience with pie menus.  They seem much better than linear menus.

Kah'plagh is Klingon for 'goodbye' or 'hello' or something Klingon.  I was
*trying* to be funny. Unsuccessfully i might add


Chris Campbell


p.s. BotF had a bad inteface but what about New Worlds? -- I'm ashamed to
have that in my house.  Armada is good except that it don't like my Athlon,
bummer!




"Ted Dennison" <dennison@telepath.com> wrote in message
news:l0Rt6.2077$94.3148@www.newsranger.com...
> In article <qzQt6.14157$Q4.2630301@news2-win.server.ntlworld.com>,
chris.danx
> says...
> >
> >One thing i think we're going to implement is pie-menus.  Has anyone
> >implemented them in Ada?  I've only used one pie-menu system in Birth of
the
> >Federation -- maybe we could call the OS 'Kirk' or 'Mad Womanizing
>
> After the reception that BotF's interface got, I'm amazed that anyone is
willing
> to repeat any part of it.
>
> As for hand-writing widgets and directly accessing disks, Ada really
shouldn't
> be an issue. There's no super-duper featue in C that Ada doesn't have an
analog
> to, and anything in an external C library can be accessed via a "pragma
import
> (C,...)". The biggest issue is typcially deciphering complex C macros.
>
> >-- anyone know the proper reply to 'Live long and prosper'.  I've
forgotten
> >but i think it's 'Kah'plagh'.
> If that's the rough Vulcan equivalent to "Up your nose with a rubber hose,
> Kotter", then its probably about right. :-)
>
> ---
> T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
>           home email - mailto:dennison@telepath.com

Re: GNAT Question

[Tarjei T. Jensen]

chris.danx wrote
>I'm implementing the BFS file system as a file on disk {maybe i mentioned
it
>before} and was going to do it in Ada first, then once i'm sure it works
>re-write the thing in C.  Maybe not a good idea.  BFS is a nightmare
anyway!
>Does GNAT support writing raw binary data to disk?  I really need to know
>this!

It is just a matter of API. Just like C, C++, assembler, etc.
Take a look at the pragma and interface section of the ada reference manual.
It is available on the net somewhere.

Pragmas to look out for are import, export, convention and linker_options

It will be a good idea to have a look at the Gnat source code. It should
contain quite a lot of what you need to know.

>I must have picked things up wrong.  Maybe i'm also picking this up wrong
>but i looked at the project and saw that some of it wasn't planned to be
>written in Ada.  I know the bootstrap needs Assembly, but i'd think
(almost)
>everything else could be done in Ada.  I'm probably wrong about this
aswell.

No you are not wrong.

Gnat can be used on bare boards. That means: no runtime stuff.

What you need to do is to create a kernel environment so that gnat works
like you want. This includes creating your own kernel version of gnat (it
means you will have to do something to the run-time library).


Don't forget to turn on run time checking!


Greetings,

Re: GNAT Question

[Robert A Duff]

cts@kampong.aedinc.net ((null)) writes:

> Just curious, does anyone out there know about how many man-years
> it might take to write an Ada compiler?

If you have people who really know Ada well, and know how to write
compilers, I'd guess about 10 person-years to get something minimal
working (not much optimization, not many optional extras).

- Bob

Re: GNAT Question

[Randy Brukardt]

Robert A Duff wrote in message ...
>cts@kampong.aedinc.net ((null)) writes:
>
>> Just curious, does anyone out there know about how many man-years
>> it might take to write an Ada compiler?
>
>If you have people who really know Ada well, and know how to write
>compilers, I'd guess about 10 person-years to get something minimal
>working (not much optimization, not many optional extras).


I would have said double that. Janus/Ada 83 took about 18 person-years,
and Ada 95 is a lot larger. Sure, *I* could write an Ada 95 compiler
from scratch in less time than that (having done it already), but why
would I want to? I'd rather improve the current one.

            Randy.

GNAT question

[Fred Stihler]

Im starting to learn Ada, I dont know how to configure the compiler to
work the way I want. Does anybody knows how do I do to be able
generate apps that output special characters? I tried to compile a
simple program that prints "Al� Mundo" but when I run it, the
character � doesnt appear. It prints a strange character. what must I
do?

Re: GNAT question

[Robert Dewar]

lab.stihler@unidavi.edu.br (Fred Stihler) wrote in message news:<c5dec41e.0208111459.40b26938@posting.google.com>...
> Im starting to learn Ada, I dont know how to configure the compiler to
> work the way I want. Does anybody knows how do I do to be able
> generate apps that output special characters? I tried to compile a
> simple program that prints "Al� Mundo" but when I run it, the
> character � doesnt appear. It prints a strange character. what must I
> do?

The first thing you must do is learn how to ask questions
on the newsgroup properly. Start a new thread with a proper
subject, don't append a reply to a completely irrelevant
thread about how to write OS level stuff using GNAT.

The second thing you must do is to say exactly what system
you are using, both the machine and operating system and
compiler.

The third thing is you should read the documentation. For
instance if you are on a PC using GNAT, you need to read
what the users manual has to say about character sets in
this environment.

Re: GNAT question

[Fred Stihler]

> The first thing you must do is learn how to ask questions
> on the newsgroup properly. Start a new thread with a proper
> subject, don't append a reply to a completely irrelevant
> thread about how to write OS level stuff using GNAT.
hey sorry, I didnt see that there was a thread with this name. I didnt
know that I must first search for thread names before posting a
thread.

thank you

Re: GNAT question

[tmoran]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 265 bytes --]

> simple program that prints "Al� Mundo" but when I run it, the
> character � doesnt appear. It prints a strange character.
  Are you saying that your source file contains a string containing
a hex F4, but when you print that you get something other than F4?

Re: GNAT question

[Robert Dewar]

tmoran@acm.org wrote in message news:<nYG59.4229$5%1.157129951@newssvr21.news.prodigy.com>...
>   Are you saying that your source file contains a string 
> containing a hex F4, but when you print that you get 
> something other than F4?


No, of course he is not saying this! The issue is most 
likely a code page issue on the PC, but how can we tell
since we do not have the information we need.

Re: GNAT question

[Fred Stihler]

dewar@gnat.com (Robert Dewar) wrote in message news:<5ee5b646.0208120437.1ee3de4@posting.google.com>...
> tmoran@acm.org wrote in message news:<nYG59.4229$5%1.157129951@newssvr21.news.prodigy.com>...
> >   Are you saying that your source file contains a string 
> > containing a hex F4, but when you print that you get 
> > something other than F4?
> 
> 
> No, of course he is not saying this! The issue is most 
> likely a code page issue on the PC, but how can we tell
> since we do not have the information we need.
yes, its about code page. Im running win98, configured to display
portuguese specific characters. I dont know where do I look for
documentation about such configurations. Im running gnat 3.14p. I have
an old turbo pascal 7 compiler and its able to display these
characters just fine. I dont know, maybe my compiler needs a patch, or
something.

Re: GNAT question

[tmoran]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 657 bytes --]

> > simple program that prints "Al� Mundo" but when I run it, the
> > character � doesnt appear. It prints a strange character.
>   Are you saying that your source file contains a string containing
> a hex F4, but when you print that you get something other than F4?
  I notice that hex F4 appears differently when displayed in text mode
(eg: type f.txt) or in Windows mode, for instance using Notepad.  Did
you perhaps enter a character that looks right, via a graphics mode
editor, and then look at the result by running the program in text/command
line mode?  What happens if you pipe the output to a file, then look at
that file with your editor?

Re: GNAT question

[Robert Dewar]

tmoran@acm.org wrote in message news:<feT59.1623$4P6.69518450@newssvr14.news.prodigy.com>...
> I notice that hex F4 appears differently when displayed 
> in text mode
> (eg: type f.txt) or in Windows mode

Well yes of course. The result of displaying the character
F4 of course depends on the code page or font used.