First non-VSI x86 application port?

[Jan-Erik Söderholm]

From the WASD mailing list:

Pleased to announce the release of WASD v12.0 for the Alpha,
Itanium and x86-64 (V9.1-A) platforms.

[Ian Miller]

Post by Jan-Erik Söderholm
Pleased to announce the release of WASD v12.0 for the Alpha,
Itanium and x86-64 (V9.1-A) platforms.

Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

[David Jones]

Post by Ian Miller
Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

SQLite was ported in September. Just like the Alpha --> IA64 transition,
the 'port' is mostly re-compiling on the new platform and involves no
where near the effort required for VAX --> Alpha. The X86 cross build
environment is biggest wrinkle I've had to deal with.

[Craig A. Berry]

Post by David Jones

Post by Ian Miller
Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

SQLite was ported in September. Just like the Alpha --> IA64 transition,
the 'port' is mostly re-compiling on the new platform and involves no
where near the effort required for VAX --> Alpha. The X86 cross build
environment is biggest wrinkle I've had to deal with.

And WASD was ported back in May:

https://groups.google.com/g/comp.os.vms/c/2t1vcNS_eYo/m/ZoT-LvtSAAAJ

[Simon Clubley]

Post by Craig A. Berry

Post by David Jones

Post by Ian Miller
Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

SQLite was ported in September. Just like the Alpha --> IA64 transition,
the 'port' is mostly re-compiling on the new platform and involves no
where near the effort required for VAX --> Alpha. The X86 cross build
environment is biggest wrinkle I've had to deal with.

https://groups.google.com/g/comp.os.vms/c/2t1vcNS_eYo/m/ZoT-LvtSAAAJ

Ok, at which point does someone come out claiming they did the earliest
port to x86-64 VMS while walking uphill to work in the snow (both ways) ? :-)

Simon.

[Dave Froble]

Post by Craig A. Berry

Post by David Jones

Post by Ian Miller
Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

SQLite was ported in September. Just like the Alpha --> IA64 transition,
the 'port' is mostly re-compiling on the new platform and involves no
where near the effort required for VAX --> Alpha. The X86 cross build
environment is biggest wrinkle I've had to deal with.

https://groups.google.com/g/comp.os.vms/c/2t1vcNS_eYo/m/ZoT-LvtSAAAJ

Maybe VSI should create a page listing "well known" software already
available for VMS x86-64.
That could create a positive impression of the state of the port.
Arne

Until there are native compilers, I would not go that far. Perhaps ported and tested,
but "available"? Beyond my comfort level.

[John Reagan]

Post by Craig A. Berry

Post by David Jones

Post by Ian Miller
Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

SQLite was ported in September. Just like the Alpha --> IA64 transition,
the 'port' is mostly re-compiling on the new platform and involves no
where near the effort required for VAX --> Alpha. The X86 cross build
environment is biggest wrinkle I've had to deal with.

https://groups.google.com/g/comp.os.vms/c/2t1vcNS_eYo/m/ZoT-LvtSAAAJ

Maybe VSI should create a page listing "well known" software already
available for VMS x86-64.
That could create a positive impression of the state of the port.
Arne

Until there are native compilers, I would not go that far. Perhaps ported and tested,
but "available"? Beyond my comfort level.
--
David Froble Tel: [phone]
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

You didn't hear this from me :) but we already have clang running on x86 and we're just days away from attempting to link a native BLISS compiler using LLVM 10.0.1 as our first landing spot.

[Simon Clubley]

Post by John Reagan
You didn't hear this from me :) but we already have clang running on x86 and we're just days away from attempting to link a native BLISS compiler using LLVM 10.0.1 as our first landing spot.

Does that mean you have CMake running under VMS ?

If so, would CMake build on Alpha as well, or is the port x86-64 specific ?

BTW, what build files are you generating ? Is it "Unix Makefiles" ?

Also BTW, does the "-target" option work for your port of clang and
does it support the usual target triples ? :-)

That last question is strictly just a "just curious" BTW. :-)

Simon.

[John Reagan]

Post by Simon Clubley

Post by John Reagan
You didn't hear this from me :) but we already have clang running on x86 and we're just days away from attempting to link a native BLISS compiler using LLVM 10.0.1 as our first landing spot.

Does that mean you have CMake running under VMS ?
If so, would CMake build on Alpha as well, or is the port x86-64 specific ?
BTW, what build files are you generating ? Is it "Unix Makefiles" ?
Also BTW, does the "-target" option work for your port of clang and
does it support the usual target triples ? :-)
That last question is strictly just a "just curious" BTW. :-)
Simon.
--
Walking destinations on a map are further away than they appear.

I didn't say where we built it. :) Currently built on Linux using a modified clang, linked on OpenVMS x86. Building on OpenVMS is the next step. We only configured the x86-64 target for the build. It is slow enough as it is. And yes, we did the same thing with libcxx and libcxxabi (which is still a work in progress)

We're working on CMake.

[Lawrence D’Oliveiro]

Post by John Reagan
We're working on CMake.

Will your patches be going back upstream?

Because the last thing you want to do is maintain your own proprietary fork.

[John Reagan]

Post by Lawrence D’Oliveiro

Post by John Reagan
We're working on CMake.

Will your patches be going back upstream?
Because the last thing you want to do is maintain your own proprietary fork.

The CMake code already comes with ifdef's for __VMS. For OpenVMS, it uses
vfork()/execvp(). Most of our "work" on CMake is shaking out our header changes
and such. I should say "started porting CMake to help shaking out other issues".
We're currently working on the personality routines in libcxxabi because of it.

[John Reagan]

Post by Lawrence D’Oliveiro

Post by John Reagan
We're working on CMake.

Will your patches be going back upstream?
Because the last thing you want to do is maintain your own proprietary fork.

The CMake code already comes with ifdef's for __VMS. For OpenVMS, it uses
vfork()/execvp(). Most of our "work" on CMake is shaking out our header changes
and such. I should say "started porting CMake to help shaking out other issues".
We're currently working on the personality routines in libcxxabi because of it.

In fairness, that vfork() change might have been make inside VSI. I'm now checking in it.

[Dave Froble]

Post by John Reagan

Post by Craig A. Berry

Post by David Jones

Post by Ian Miller
Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

SQLite was ported in September. Just like the Alpha --> IA64 transition,
the 'port' is mostly re-compiling on the new platform and involves no
where near the effort required for VAX --> Alpha. The X86 cross build
environment is biggest wrinkle I've had to deal with.

https://groups.google.com/g/comp.os.vms/c/2t1vcNS_eYo/m/ZoT-LvtSAAAJ

Maybe VSI should create a page listing "well known" software already
available for VMS x86-64.
That could create a positive impression of the state of the port.
Arne

Until there are native compilers, I would not go that far. Perhaps ported and tested,
but "available"? Beyond my comfort level.
--
David Froble Tel: [phone]
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

You didn't hear this from me :) but we already have clang running on x86 and we're just days away from attempting to link a native BLISS compiler using LLVM 10.0.1 as our first landing spot.

Too late! I heard that.

Congratulations and thanks for what you've done so far. Keep up the excellent
work.

When will I not hear that anyone is thinking about Basic?

[John Reagan]

Post by John Reagan

Post by Craig A. Berry

Post by David Jones

Post by Ian Miller
Perl was announced earlier :-)
https://groups.google.com/g/comp.os.vms/c/yBNni5v_oLk

SQLite was ported in September. Just like the Alpha --> IA64 transition,
the 'port' is mostly re-compiling on the new platform and involves no
where near the effort required for VAX --> Alpha. The X86 cross build
environment is biggest wrinkle I've had to deal with.

https://groups.google.com/g/comp.os.vms/c/2t1vcNS_eYo/m/ZoT-LvtSAAAJ

Maybe VSI should create a page listing "well known" software already
available for VMS x86-64.
That could create a positive impression of the state of the port.
Arne

Until there are native compilers, I would not go that far. Perhaps ported and tested,
but "available"? Beyond my comfort level.
--
David Froble Tel: [phone]
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

You didn't hear this from me :) but we already have clang running on x86 and we're just days away from attempting to link a native BLISS compiler using LLVM 10.0.1 as our first landing spot.

Too late! I heard that.
Congratulations and thanks for what you've done so far. Keep up the excellent
work.
When will I not hear that anyone is thinking about Basic?
--
David Froble Tel: [phone]
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

I'm currently fighting with BASIC MAPs. Fortran generates reasonable GEM CIL and symbol table for COMMON blocks, but BASIC uses a "clever" sequence of symbol tables that the GEM to LLVM converter was unprepared for. I think I grok it.

[Hunter Goatley]

I have started building some of my freeware for OpenVMS X86_64, too. You
can find them in the OpenVMS Resource Center on the Process Software
website:

https://www.process.com/resources/openvms/index.html

This ungainly URL will display the packages that run on OpenVMS X86_64 now.

http://vms.process.com/scripts/fileserv/fileserv_search.exe?package=&description=&author=&system=x86

There are only two right now, but I hope to add more soon.