Gnat Ada on OpenVMS is back

Gnat Ada on OpenVMS is back

[gérard Calliet]

Hello,

The OpenVMS OS, after an announced death in 2013, is back with its 
primary engeniering team, organized as a start-up (www.vmssoftware.com).

In this context, we did a new build on OpenVMS / Itanium of the Gnat Ada 
compiler. We needed it for an industrial project (a port from OenVMS 
Alpha to OpenVMS Itanium, an urban transportation control software). The 
project is about to be runned in production, with images compiled by our 
Gnat Ada compiler.

We want to generalize this renewal of Ada on OpenVMS. We offer free of 
charge our compiler (with no support) to downloading. The portal (still 
in work) www.vmsadaall.org can be used to ask for a download. We wish it 
becomes a place for exchanges between all people that are interested by 
the initiative.

If you are interested, please tell us, we need evaluations, advice, ideas.

Gérard Calliet
contact@vmsadaall.org


---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus

Re: Gnat Ada on OpenVMS is back

[Simon Clubley]

On 2017-03-02, gérard Calliet <gerard.calliet@pia-sofer.fr> wrote:
> Hello,
>
> The OpenVMS OS, after an announced death in 2013, is back with its 
> primary engeniering team, organized as a start-up (www.vmssoftware.com).
>
> In this context, we did a new build on OpenVMS / Itanium of the Gnat Ada 
> compiler. We needed it for an industrial project (a port from OenVMS 
> Alpha to OpenVMS Itanium, an urban transportation control software). The 
> project is about to be runned in production, with images compiled by our 
> Gnat Ada compiler.
>

Congratulations Gerard.

I've tried to build GNAT for VMS Alpha from the FSF sources several
times but I never succeeded because various bits appear to be
incomplete or missing for the VMS Alpha target.

I never tried to build an IA64 compiler because I don't have an IA64
box but it's clearly more complete than the Alpha target is (or you
put more work into it than I was willing to do. :-))

What versions of the FSF GCC and binutils source kits did you use ?

How much trouble did you have building gcc against the DEC C headers ?

I have noticed that the DEC C headers required fewer changes in later
versions of FSF gcc in my various build attempts than the earlier
versions did.

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

Re: Gnat Ada on OpenVMS is back

[Dennis Lee Bieber]

On Thu, 2 Mar 2017 13:37:12 +0100, gérard Calliet
<gerard.calliet@pia-sofer.fr> declaimed the following:

>Hello,
>
>The OpenVMS OS, after an announced death in 2013, is back with its 
>primary engeniering team, organized as a start-up (www.vmssoftware.com).
>
	Sounds like it might be the end of the hobbyist release...
-- 
	Wulfraed                 Dennis Lee Bieber         AF6VN
    wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/

Re: Gnat Ada on OpenVMS is back

[gérard Calliet]

Le 02/03/2017 à 20:26, Simon Clubley a écrit :
> On 2017-03-02, gérard Calliet <gerard.calliet@pia-sofer.fr> wrote:
>> Hello,
>>
>> The OpenVMS OS, after an announced death in 2013, is back with its
>> primary engeniering team, organized as a start-up (www.vmssoftware.com).
>>
>> In this context, we did a new build on OpenVMS / Itanium of the Gnat Ada
>> compiler. We needed it for an industrial project (a port from OenVMS
>> Alpha to OpenVMS Itanium, an urban transportation control software). The
>> project is about to be runned in production, with images compiled by our
>> Gnat Ada compiler.
>>
>
> Congratulations Gerard.
>
> I've tried to build GNAT for VMS Alpha from the FSF sources several
> times but I never succeeded because various bits appear to be
> incomplete or missing for the VMS Alpha target.
>
> I never tried to build an IA64 compiler because I don't have an IA64
> box but it's clearly more complete than the Alpha target is (or you
> put more work into it than I was willing to do. :-))
>
> What versions of the FSF GCC and binutils source kits did you use ?
>
> How much trouble did you have building gcc against the DEC C headers ?
>
> I have noticed that the DEC C headers required fewer changes in later
> versions of FSF gcc in my various build attempts than the earlier
> versions did.
>
> Simon.
>
We have built gcc a 4.7. Because, afterwards, gcc needs some c++ 
features we did not find on itanium OpenVMS 8.4.1-h1 (from VSI, 18months 
ago).

We had some troubles with DEC C headers, a issue with prefixes on symbols.

The build was for a big port operation, which is just ending. Now I want 
to take time to generalize the offer, hoping feedback of the community. 
Please connect to our "vms ada alliance", and we could go on on gcc, for 
the moment on Itanium. (You can perhaps find some "for hobbyist" 
platform with itanium,  and anyway every discussion is worth it).

Gérard Calliet

---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus

Re: Gnat Ada on OpenVMS is back

[gérard Calliet]

Le 03/03/2017 à 02:18, Dennis Lee Bieber a écrit :
> On Thu, 2 Mar 2017 13:37:12 +0100, gérard Calliet
> <gerard.calliet@pia-sofer.fr> declaimed the following:
>
>> Hello,
>>
>> The OpenVMS OS, after an announced death in 2013, is back with its
>> primary engeniering team, organized as a start-up (www.vmssoftware.com).
>>
> 	Sounds like it might be the end of the hobbyist release...
>
I don't think so. It seems VSI will continue the hobbyist program.
And, without VSI, no more OpenVMS, so,...

Gérard Calliet

---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus

Re: Gnat Ada on OpenVMS is back

[Simon Clubley]

On 2017-03-03, gérard Calliet <gerard.calliet@pia-sofer.fr> wrote:
> We have built gcc a 4.7. Because, afterwards, gcc needs some c++ 
> features we did not find on itanium OpenVMS 8.4.1-h1 (from VSI, 18months 
> ago).
>

The approach I took was somewhat different as I didn't have a working
Ada compiler on VMS to be able to compile the gcc Ada frontend with.

In some discussions with Adacore several years ago, it appeared that
they were building a cross compiler on Linux and then using that
cross compiler to built a native VMS compiler on Linux, but the
difference if that they were doing this with their own version of
the gcc/binutils toolchain.

Adacore have been pushing various pieces into the FSF toolchain from
their own toolchain which is why I tried to duplicate this approach
with the FSF toolchain.

I don't know if this is the approach you took or if you relied on
having an existing GNAT Pro compiler for VMS Alpha available to be
able to build the FSF sources for an IA64 target.

Building on Linux has the downside that you need the FSF toolchain to be
sufficiently complete for you to be able to build the cross compiler.

When you say missing C++ features on VMS, are you talking about the
run-time environment or do you mean the HP C++ compiler ? If it's the
latter then I don't understand why you simply couldn't use the C++
support in gcc (unless Adacore have disabled it for their GNAT Pro
compiler for VMS Alpha).

> We had some troubles with DEC C headers, a issue with prefixes on symbols.
>

I hit the same problem back when I tried to build a cross compiler
for VMS Alpha with gcc 4.7 as well.

> The build was for a big port operation, which is just ending. Now I want 
> to take time to generalize the offer, hoping feedback of the community. 
> Please connect to our "vms ada alliance", and we could go on on gcc, for 
> the moment on Itanium. (You can perhaps find some "for hobbyist" 
> platform with itanium,  and anyway every discussion is worth it).
>

My interest in VMS is not as strong as it once was and as I don't
have an IA64 box, it's not something I can try out.

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world