Re: /⍨

[Dr . Jürgen Sauermann]

    Hi Peter,
      
      coming back to your proposal below, I have added a document
      
      HOWTOs/GNU-APL-Troubleshooting.html
      
      SVN 1563.
      
    Best Regards,
    Jürgen

On 6/7/22 7:39 PM, Peter Teeson wrote:

Hi Jurgen:

If it’s not already there please add the debugging tips you give below as an item to the documentation.

respect….

Peter

On Jun 7, 2022, at 12:56 PM, Dr. Jürgen Sauermann <mail@xn--jrgen-sauermann-zvb.de> wrote:

Hi Hudson,

I believe that I fixed the double execution for executable scripts,
Looks like the OS handles executable scripts differently than
non-executable ones. SVN 1560.

The F ← {⍵ × (?⍨⍵) ,¨ ⍳⍵} bug is something that I cannot reproduce.
I wonder if it happens always or lnly sometimes. I have some suggestions
for you and others that make my life easier:

* run "make develop" in the top-level directory.

That enables some more internal checks that may be useful for troubleshooting.
Most importantly it enables dynamic logging.

* run "make apl.lines" in the src directory,

That makes GNU APL's stack traces more readable, i.e.whowing  line numbers rather
than hex addresses. After that only use the apl binary in the src directory.

* do "ulimit -c unlimited" go get a core file when apl crashes silently,

It does not hurt and after that you can "gdb ./apl core" to obtain more information
about where a fault has occured (gdb command bt)

* in apl: do

      ]log 25
      ]log 26

That gives more details about where APL errors were thrown. In particular with
your function F because I cannot quite see why it would give a DOMAIN ERROR at all
and therefore the location where that happens would be interesting.

Thanks,
Jürgen

On 6/6/22 7:32 PM, hudson@hudsonlacerda.com wrote:

Hi Jürgen,

----- Dr. Jürgen Sauermann <mail@jürgen-sauermann.de> escreveu:
[...]
> 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).
[….]

Here it is:
  g++ (Debian 11.3.0-3) 11.3.0  

⌹2 is now fixed, but there are strange behaviours in other cases.
Please see the attachments.

Best,
Hudson