[Info-vax] Fortran

[Dave Froble]

On 12/3/2018 3:21 PM, Arne Vajhøj wrote:
> On 12/3/2018 8:12 AM, Jan-Erik Söderholm wrote:
>> Den 2018-12-03 kl. 14:08, skrev Simon Clubley:
>>> On 2018-12-03, Phillip Helbig (undress to reply)
>>> <helbig at asclothestro.multivax.de> wrote:
>>>> In article <g6jcegFmjliU1 at mid.individual.net>, Steve Lionel
>>>> <steve at seesignature.invalid> writes:
>>>>> It's official! ISO published Fortran 2018 on November 28! Online
>>>>> copies
>>>>> are not available due to ISO copyright rules - you can buy a copy from
>>>>> ISO - but J3 has a working document as a source for interpretations at
>>>>> https://j3-fortran.org/doc/year/18/18-007r1.pdf .
>>>>
>>>> When will we see this on VMS?
>>>
>>> Alpha (and probably Itanium): Approximately never.
>>>
>>> x86-64: After the following overall sequence of events:
>>>
>>> 1) Integration of Fortran 2018 into LLVM is complete,
>>> 2) x86-64 VMS is released,
>>> 3) John ports CMake to x86-64 VMS,
>>> 4) John ports a recent version of LLVM to x86-64 VMS.
>>>
>>> My best guess is approximately lunchtime on Friday. Maybe teatime
>>> instead if the schedule slips. :-)
>>>
>>>> I remember a statement from VSI that one of their goals was to have
>>>> up-to-date compilers.
>>>
>>> I thought that was only for LLVM based compilers and that LLVM
>>> was only going to be available for x86-64 VMS. Are there any
>>> plans I have missed to port LLVM to earlier VMS architectures ?
>>
>> Was "earlier VMS architectures" even mentioned in the post from Phillip?
>
> No.
>
> But I think there are two different sort of implicit assumptions:
>
> will XYZ be available on VMS = will XYZ be available on VMS x86-64
>
> will XYZ be available on VMS = will XYZ be available on VMS on all
> supported platforms
>
> Arne
>
>

Personally, my perspective is that the only path forward for VSI is x86, 
for now, and any efforts on VAX, Alpha, and that sinking ship detract 
from what they need to be doing.

And why do so?  Once VMS is running on x86, most should be able to adopt 
that platform and use all the new goodies from VSI.  Forget backward 
compatibility.  Forget clustering with older platforms.  (Looking at you 
Phillip).  It's not worth the effort.  Any efforts in that direction 
lessen the "new goodies" from VSI.

VSI's efforts are to keep a common code base for the OS.  That should be 
more than enough.

Look at it this way.  Back porting LLVM and compilers to older HW 
probably would cost more than just giving those needing it some new x86 
systems.

-- 
David Froble                       Tel: 724-529-0450
Dave Froble Enterprises, Inc.      E-Mail: davef at tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA  15486