Re: /⍨

[Dr . Jürgen Sauermann]

Hi Hudson,

see below.

Best Regards,
Jürgen


On 6/5/22 12:06 AM, hudson@hudsonlacerda.com wrote:
> Hi Jürgen,
>
> About ⌹2  , the exception also occurs at CLEAR WS (see attached log).
Thanks. I believe this is a compiler error (which does not happen with 
mine -
g++ (Ubuntu 9.4.0-1ubuntu1~18.04) 9.4.0).

Apparently your compiler over-optimizes:

*       Cell cZ;**
**       B->get_cscalar().bif_reciprocal(&cZ);**
**       Z->next_ravel_Cell(cZ);**
*
in file *Bif_F12_DOMINO.cc* and does not notice that *bif_reciprocal* 
changes cZ.

I have changes the code a little and maybe that fixes it. *SVN 1558*.
> The )help for ⎕ uses ⌷ in the examples.
fixed.
> The )help for ⍞ describes input twice. (Perhaps mention A←⍞←B ?)
fixed.
> Regards,
> Hudson
>
> ----- Em 4 de Jun de 2022, em 14:38, Dr. Jürgen Sauermann 
> mail@jürgen-sauermann.de escreveu:
>
> > Hi Hudson,
> >
> > thanks, see below ( SVN 1557 ).
> >
> > On 6/2/22 4:52 PM, [ mailto:hudson@hudsonlacerda.com | hudson@hudsonlacerda.com
> > ] wrote:
> >
> > > Hi Jürgen,
> > > Many thanks.
> > > 1) The reference to the ISO "PDF" file without the ZIP-format warning is in the
> > > preamble of doc/apl.html .
> > fixed.
> >
> > > 2) ⌹
> > > ⌹ ,2 (as vector) is OK
> > > ⌹ 2 (as scalar) raises the exception
> > > ==============================================================================
> > > Assertion failed: 0 && "Cell::init_other() called on base class"
> > > in Function:      init_other
> > > in file:          Cell.hh:64
> > Still having problems to reproduce this:
> >
> > ⌹,
> > 0.5
> > ⌹2
> > 0.5
> >
> > Does it also happen in a CLEAR WS ?
> >
> > > 3) Still missing help for ⎕ and ⍞ — not listed by ")help TAB"
> > > )help ⎕
> > > )help ⍞
> > fixed.
> >
> > > 4) ⎕DL cannot be stopped by ATTENTION (Control+C) ­— is it so by design?
> > Actually you need INTERRUPT (= double ^C within short time) and not
> > ATTENTION (= single ^C).
> >
> > > 5) /⍨ now works — but not /¨ :
> > >       (1 0 1)(0 1 1) /¨ ⊂'abc'
> > > DOMAIN ERROR
> > >        (1 0 1) (0 1 1)/¨ ⊂'abc'
> > >        ^                ^
> > >        (1 0 1)(0 1 1) /¨¨ ⊂'abc'
> > > [INFINITE LOOP]
> > > Alternative to /¨ (besides {⍺/⍵}¨ ):
> > >        ,/⊃(1 0 1)(0 1 1) ⍴¨¨ ⊂'abc'
> > >   ac bc
> > fixed. I should mention that some invariants in the IBM APL2 language reference
> > manual are incorrect (at least my PC APL2 differs from the manual). The
> >
> > (0≠L)/R ←→ ⊃,/L⊂R on page 188
> >
> > seems to hold only if L⊂R is not empty.
> >
> > > (Maybe it is not without reason that IBM APL2 does not support operators along
> > > with / ).
> > > Other issues are fixed.
> > > Regards,
> > > Hudson
> > > P.S.: Is it in your TODO list support for Unicode string manipulation (like
> > > toupper(), tolower() etc. )?
> > > ----- Dr. Jürgen Sauermann <mail@jürgen-sauermann.de> escreveu:
> > > > Hi Hudson,
> > > >
> > > > thanks for reporting your issues. I believe that I fixed (most of) them
> > > > in *SVN 1556*.
> > > >
> > > > Regarding the ISO PDF file, it is mentioned below the link in
> > > > README-7-more-info
> > > > that the file needs to be unzipped.
> > > >
> > > > Regarding *⌹{⍺+÷⍵}⍣1000⍨ 1* this works just fine on my machine:
> > > >
> > > > *      ⌹{⍺+÷⍵}⍣1000⍨ 1**
> > > > **0.6180339887**
> > > > *
> > > > If you can reproduce the fault, then please check with *gdb* (after
> > > > *make apl.lines*
> > > > in directory *src* as to get better line numbers in *gdb*).
> > > >
> > > > Best Regards,
> > > > Jürgen
> > > [...]