[Info-vax] Pascal question

[John Reagan]

On Monday, January 21, 2019 at 8:18:26 PM UTC-5, Arne Vajhøj wrote:
> On 1/21/2019 4:57 PM, John Reagan wrote:
> > On Monday, January 21, 2019 at 4:20:58 PM UTC-5, Arne Vajhøj wrote:
> >> On 1/21/2019 4:05 PM, Bill Gunshannon wrote:
> >>> On 1/21/19 3:41 PM, Arne Vajhøj wrote:
> >>>> On 1/21/2019 1:53 PM, Dave Froble wrote:
> >>>>> On 1/21/2019 1:07 PM, Arne Vajhøj wrote:
> >>>>>> Maybe it was time to add 25-50 convenience functions to
> >>>>>> VMS Pascal?
> >>>>>>
> >>>>>> People expect more today. And adding them would show some
> >>>>>> progress. And the effort should not be that big aka low hanging
> >>>>>> fruit.
> >>>>>
> >>>>> Or, one can build a library  of often used capabilities ....
> >>>>
> >>>> Of course.
> >>>>
> >>>> But it would be better if the standard language runtime came
> >>>> with such functionality instead of having hundreds of developers
> >>>> each implement their own.
> >>>>
> >>>> It is not like uppercasing is a very special function that
> >>>> only very few need.
> >>>>
> >>>> It is relative common.
> >>>>
> >>>> And many other mostly newer languages certainly has it.
> >>>
> >>> So push the Pascal Standards Committee to add it.  :-)
> >>
> >> Not going to happen.
> >>
> >> But since VMS Pascal already have several extensions compared
> >> to ISO Pascal, then the lack of it in the standard should not
> >> prevent VSI from adding it.
> > 
> > Good point.  Just as soon as a real customer writes that on the back
> > of a $100 bill, I'll get right on it.  I'll call STR$UPCASE for you.
> > :)
> 
> I know you have priorities.
> 
> My thinking is just along the lines of:
> * it will take some time until the VMS Pascal users start
>    running their production on VMS x86-64.
> * many improvements of Pascal would be a lot of work to
>    implement and would certainly be a no go from a
>    priority perspective
> * but maybe just maybe it would make sense to release
>    a new Pascal version with some easy stuff (a few
>    bug fixes, a bunch of new predefined functions, a few
>    extra PEN files) to show that VMS is on the move - throw
>    us a few appetizers while we are waiting for the big meal
> 
> Arne

I did add a bunch of things to Pascal before I took my 5 year assignment to NonStop.  Given the difficulty of updating the documents, all of the new features are in the release notes.  Anybody notice the new SELECT and SELECTONE statements (stolen from BLISS), the %IF conditional compilation, or all the 64-bit pointer support?

Given that I was a Pascal user before I joined Digital (I answered my own SPR when I started) and was on the Pascal committee, I have a strong connection with Pascal.  Don't tell my management but I do have a list of additional things I'm going to add to Pascal.  It needs 64-bit descriptor work and all the RTL data structures need to be 64-bit'ized.  We did that for our Tru64 product, but I wasn't able to make the change on Alpha or Itanium. I'll widen them all for x86.

I really want to improve the SDL Pascal backend to get a much better STARLET.PAS.  The current Pascal output was done back in the VAX days with only 32-bit INTEGERs.  Now that Pascal has lots of 64-bit pointer and 64-bit integer support, I can make it much better.  However, there might be code that has used the traditional STARLET.PAS prototypes that might not like the "better" ones.  However, I won't do the NEWSTARLET that the C headers did.