Re: [netrexx-pipelines] Migrate OS/2 pipelines to NetRexx Pipes

[<hp...@we...>]

Hello René!

Am 06.06.2022 um 22:23 schrieb René Jansen:
>> On 4 Jun 2022, at 16:20, hp...@we... wrote:
> […]
>> Back to the subject. […] Instead of
>>
>>> say length("ABC")
>>
>> I have to put it this way:
>>
>>> say "ABC".length
>>
>> What a progress! Re-use of old-style REXX routines impossible.
>
> Glad you said that. This is one of the improvements of 4.04 - there you can use Classic Rexx notation for all built-in-functions (BIFs). In this case, I did that and the original decision to have only oo-invocation is Mike’s.

The saying goes 'the better is the death of the good', but since I
am faster with what I'm used to, it must be much better. So thank
you to bring back the old-style notation, I'm looking forward to
V4.04.

>> OK, back to the subject.
>> [...] Something
>> adequate the GLOBALV concept of VM/CMS? Well, that would require
>> some kind of "pull" in ooRexx after the NetRexx pipe finished, a
>> "push" of the results (variables set directly) would be more
>> appropriate.
>>
> We have to speak to Rony about this, I am sure BSF4ooRexx has a fitting solution here.

With my finding of this WE I made the next step and prepared a DIY
stage, which (hopefully) could serve as beginning of Varload. It
filters and rearranges the content of VilMA.INI to be used by an
interpret() -- no options at the moment:

> /* p4varload_1.nrx: prepare for kind of VARLOAD */
> import org.netrexx.njpipes.pipes.
> class p4varload_1 extends stage
> method run()
>    loop forever
>       line = Rexx peekto()
>       parse line.upper sc +1 vn (sc) sv   /* UPPER to force keyword SYMBOLIC */
> /* First: nice pitfall could be IF expression, expression, expression, ... THEN */
> /* length(sc) > 0 to avoid filter STRIP!L -- whild strip is missing here */
> /* pos(sc, '* ') = 0 is the default COMMENT "* " */
> /* the plan is to put the output to stack or stem and interpret it line by line */
> /* I was unable to split the clause and inject "inline" comments */
>       if sc.length > 0 & '* '.pos(sc) = 0 then output(vn '=' sv)
>       readto()
>    catch StageError
>      rc = rc()
>    end
> exit(rc*(rc<>12))

BTW, to be in conformity with CMS Pipelines the stage LOCATE
should work with all abbreviations down to L.
What I describe in the comment as "nice pitfall" is in fact a
nasty surprise sice it works just in contrary to ooRexx and C.
A commentline like

> /* pos(sc, '* ') = 0 is the default COMMENT /* / */

was rejected by the compiler (or translator?) for "unbalanced /*"
what is kind of overdone.

The output of p4varload_1 looks nice:
> C:\prgm\NetRexx\examples\VilMA>ripe "(sep !) < vilma.ini!p4varload_1!term
> VIMDR = C:\PRGM\REXX\VILMA\
> CTLDR = C:\PRGM\HP-EMULATORS\V41\
> VILDR = C:\PRGM\HP-EMULATORS\VIRTIL\
> DELTA = 3
> ITLVL = STD.VILMA
> CTRLS = VILMA.CONTROLLERS
> TPCPT = TYPICAL.CAPTION
> DIAPOS = 321 23

Now we "only" need to interpret this in ooRexx.

Because I am a little unhappy with the "if sc.length > 0 &..." I
tried to insert something like

> addpipe "(sep !) *in:!strip!l!*in:"
but the compiler doesn't like the way I put it -- and I can't find
sufficing documentation about addpipe.

Now, without help, I'm at the road's end.

Best,
M.