Bug J9.7-beta 3: ^: prematurely stopping

[Jan-Pieter Jacobs]

I think I ran into another bug in J 9.7-beta3: basically, the last item of u^:a: is not (always) the same as u^:_  . This example runs correctly in J9.6.3.

1.ijs

mem =: [: (# _1}:@,@:,.~i.@>.@-:@#) "."0

ind =: _1 (i.&1 , i:&0)@:= ]

upd =: (ind (|.@:{)`[`]}^:(</@[) ])

session log:

   load'~temp/1.ijs'
   JVERSION
Engine: j9.7.0-beta3/j64arm/android
Build: commercial/2025-04-16T14:35:35/clang-14-0-7/SLEEF=1
Library: 9.7.1
J Android: 1.4.88/14/34
Platform: Android 64 (arm64-v8a)
Installer: unknown
InstallPath: /storage/emulated/0/Android/data/com.jsoftware.j.beta.android/files
Contact: www.jsoftware.com
   ]t=:mem '12345'
0 _1 _1 1 1 1 _1 _1 _1 _1 2 2 2 2 2
   upd^:a: t
0 _1 _1 1 1 1 _1 _1 _1 _1  2  2  2  2  2
0  2 _1 1 1 1 _1 _1 _1 _1  2  2  2  2 _1
0  2  2 1 1 1 _1 _1 _1 _1  2  2  2 _1 _1
0  2  2 1 1 1  2 _1 _1 _1  2  2 _1 _1 _1
0  2  2 1 1 1  2  2 _1 _1  2 _1 _1 _1 _1
0  2  2 1 1 1  2  2  2 _1 _1 _1 _1 _1 _1
   upd^:_ t
0 2 2 1 1 1 _1 _1 _1 _1 2 2 2 _1 _1
   (upd^:_ -: [: {: upd^:a:) t
0

Best regards,

Jan-Pieter

[Henry Rich]

Fixed for next beta, which should be soon.  The gerund} was executed in place but it shouldn't have been.  The best workaround I can see is to replace

`]}

with

`(>@<@])}

Thank goodness you're running the betas!

Henry Rich

To unsubscribe from this group and stop receiving emails from it, send an email to forum+un...@jsoftware.com.

[R.E. Boss]

JVERSION

Engine: j9.7.0-beta4/j64avx2/windows

Build: commercial/2025-05-05T13:15:47/clang-18-1-8/SLEEF=1

Library: 9.7.1

Qt IDE: 2.5.11/6.8.3(6.8.3)

OS Ver: Windows 10 Version 22H2 10.0.19045

Platform: Win 64

Installer: j9.7 install

InstallPath: c:/program files/j9.7

Contact: www.jsoftware.com

 

   JVERSION

Engine: j9.6.0-beta22/j64avx2/windows

Build: commercial/2024-10-10T20:22:53/clang-18-1-8/SLEEF=1

Library: 9.6.14

Qt IDE: 2.5.6/6.5.3(6.5.3)

OS Ver: Windows 10 Version 22H2 10.0.19045

Platform: Win 64

Installer: j9.6 install

InstallPath: c:/program files/j9.6

Contact: www.jsoftware.com

 

Let t be such that

   |:"2 t

1  1  1  1 1  1 1  1  1  1 1  1  1  1  1  1  1  1  1  1  2 _2

2 _2  2 _2 2 _2 2 _2  2 _2 2 _2  2 _2  2 _2  2 _2  2 _2 _1 _1

 

2 _2 _2  2 2 _2 2 _1 _1 _2 2 _2 _2  2 _2 _1 _1  2 _2  2  2 _2

1  1 _1 _1 1  1 1 _2  2  1 1  1 _1 _1 _1 _2  2 _1 _1 _1  1  1

   $@:~."2 t

4 2

6 2

 

   i.~"2 t                NB. Which is wrong

0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 20 21

0 1 2 3 0 1 0 7 8 1 0 1 2 3 2 7 8 3 2 3  0  1

 

   (~. i. ])"2 t        NB. Which seems to be right

0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 2 3

0 1 2 3 0 1 0 4 5 1 0 1 2 3 2 4 5 3 2 3 0 1

 

R.E. Boss

[Marcin Żołek]

Is the following the correct definition of t (shape 2 22 2)?

t =: |:"2 ] 2 2 22 $ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 _2 2 _2 2 _2 2 _2 2 _2 2 _2 2 _2 2 _2 2 _2 2 _2 2 _2 _1 _1 2 _2 _2 2 2 _2 2 _1 _1 _2 2 _2 _2 2 _2 _1 _1 2 _2 2 2 _2 1 1 _1 _1 1 1 1 _2 2 1 1 1 _1 _1 _1 _2 2 _1 _1 _1 1 1

I used Dissect tool to analyze i.~"2 t

load 'debug/dissect’

dissect 'i.~"2 t'

Clicking on a row in the output field highlights the part of the input data from which that row was created. What is wrong with the output?

Marcin

[Henry Rich]

The result looks right to me.

   |:"2 t
1  1  1  1 1  1 1  1  1  1 1  1  1  1  1  1  1  1  1  1  2 _2
2 _2  2 _2 2 _2 2 _2  2 _2 2 _2  2 _2  2 _2  2 _2  2 _2 _1 _1

2 _2 _2  2 2 _2 2 _1 _1 _2 2 _2 _2  2 _2 _1 _1  2 _2  2  2 _2
1  1 _1 _1 1  1 1 _2  2  1 1  1 _1 _1 _1 _2  2 _1 _1 _1  1  1

     i.~"2 t

0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 20 21
0 1 2 3 0 1 0 7 8 1 0 1 2 3 2 7 8 3 2 3  0  1

As a check,

    t -: ((i.~"2)   {"1 2   ]) t
1

i.~ returns not the index in the nub, but the index in y itself.

Henry Rich

[R.E. Boss]

Thanks.
I knew what i.~ does, but did not realize it.

 

R.E. Boss