Re: Potential problem in diadic iota (⍳) or in Pick (⊃)

2020-08-07 Thread Hans-Peter Sorge 

Hi, this boils down to:
 i←63
  t←? 1000 ⍴ i
  t = 2÷⍨ {⍵×2} fasteach t
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

  i←64
  t←? 1000 ⍴ i
  t = 2÷⍨ {⍵×2} fasteach t
0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
1 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 0 1 0 0 
0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 
0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 
0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 
0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 
0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 
0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0


Regards, hans-Peter

Am 07.08.20 um 02:50 schrieb Christian Robert:

Can someone help me debug this function (fasteach). I'm not the author
but from what I've read it may be a faster than the "each" Operator 
especially

when a few unique numbers in a large array


  ∇z←test;⎕RL;t;i;r
⎕RL←0
i←20
Loop:
t←? 1000 ⍴ i
i,r←t ≡ 2÷⍨ {⍵×2} fasteach t
→(r=0)/Fail
i←i+1
→Loop
Fail:
''
'Failed with t←? 1000 ⍴', (⍕i)
''
z←t
∇

  ∇z←(F fasteach) A;shp;vec;nub;res;idx
shp←⍴A   ⍝ Shape and ...
vec←,A   ⍝ ... ravel of arg.
nub←∪ vec    ⍝ Vector of unique elements.
nub←∪ vec    ⍝ Vector of unique elements.
res←F¨nub    ⍝ Result for unique elts.
idx←nub ⍳ vec    ⍝ Indices of arg in nub ...
z←shp ⍴ idx ⊃¨⊂ res  ⍝ ... distribute result.
∇


 

Re: Potential problem in diadic iota (⍳) or in Pick (⊃)

2020-08-07 Thread Kacper Gutowski 

On Thu, Aug 06, 2020 at 08:50:27PM -0400, Christian Robert wrote:

 ∇z←(F fasteach) A;shp;vec;nub;res;idx
shp←⍴A   ⍝ Shape and ...
vec←,A   ⍝ ... ravel of arg.
nub←∪ vec⍝ Vector of unique elements.
nub←∪ vec⍝ Vector of unique elements.
res←F¨nub⍝ Result for unique elts.
idx←nub ⍳ vec⍝ Indices of arg in nub ...
z←shp ⍴ idx ⊃¨⊂ res  ⍝ ... distribute result.
∇


The problem is indeed in iota and it reminds me a similar problem we had 
with an unique before.  For sufficiently big arrays, a different 
algorithm is used where the left argument is sorted beforehand to speed 
up searching; and something goes wrong there.


Smaller reproducer:

  A←⌽⍳64
  B←⍳6
  B≡A[A⍳B]
0

Should be 1, of crouse.

-k

Re: Potential problem in diadic iota (⍳) or in Pick (⊃)

2020-08-07 Thread Dr . Jürgen Sauermann 

  
  
Gentlemen,
  
  thanks, fixed in SVN 1334.

Best Regards,
Jürgen


On 8/7/20 4:33 PM, Kacper Gutowski
  wrote:

On Thu, Aug
  06, 2020 at 08:50:27PM -0400, Christian Robert wrote:
  
   ∇z←(F fasteach) A;shp;vec;nub;res;idx

shp←⍴A   ⍝ Shape and ...

vec←,A   ⍝ ... ravel of arg.

nub←∪ vec    ⍝ Vector of unique elements.

nub←∪ vec    ⍝ Vector of unique elements.

res←F¨nub    ⍝ Result for unique elts.

idx←nub ⍳ vec    ⍝ Indices of arg in nub ...

z←shp ⍴ idx ⊃¨⊂ res  ⍝ ... distribute result.

∇

  
  
  The problem is indeed in iota and it reminds me a similar problem
  we had with an unique before.  For sufficiently big arrays, a
  different algorithm is used where the left argument is sorted
  beforehand to speed up searching; and something goes wrong there.
  
  
  Smaller reproducer:
  
  
    A←⌽⍳64
  
    B←⍳6
  
    B≡A[A⍳B]
  
  0
  
  
  Should be 1, of crouse.
  
  
  -k

Re: Potential problem in diadic iota (⍳) or in Pick (⊃)

2020-08-07 Thread Christian Robert 

It fix the "Smaller reproducer:" but does not fix the whole problem:

$ apl -v
BUILDTAG:
-
Project:GNU APL
Version / SVN:  1.8 / 1334M
Build Date: 2020-08-07 19:19:29 UTC
Build OS:   Linux 5.7.11-200.fc32.x86_64 x86_64
config.status:  'DYNAMIC_LOG_WANTED=no' 'CXXFLAGS=-Werror'
Archive SVN:1319



  ∇z←test;⎕RL;t;i;r
⎕RL←0
i←20
Loop:
t←? 1000 ⍴ i
i,r←t ≡ 2÷⍨ {⍵×2} fasteach t
→(r=0)/Fail
i←i+1
→Loop

Fail:
''
'Failed with t←? 1000 ⍴ ', (⍕i)
''
z←t
∇

  ∇test2;z;t;d;shpz;⎕RL;r
⎕RL←0
d←3
Loop:
z←(d,d) ⍴ (d×d)?(d×d)
shpz←⍴z
t←shpz ⍴ ((∪,z) ⍳ ,z) ⊃ ¨ {⍵×2} ¨⊂ ,z
''
d,("False" "True")[(⎕io+r←z≡t÷2)]
z (t÷2)
→(r≠1)/Fail
d←d+1
→(d<10)/Loop
Fail:
∇

  ∇z←(F fasteach) A;shp;vec;nub;res;idx
shp←⍴A   ⍝ Shape and ...
vec←,A   ⍝ ... ravel of arg.
nub←∪ vec⍝ Vector of unique elements.
res←F¨nub⍝ Result for unique elts.
idx←nub ⍳ vec⍝ Indices of arg in nub ...
z←shp ⍴ idx ⊃¨⊂ res  ⍝ ... distribute result.
∇

  test
20 1
21 1
22 1
23 1
24 1
25 1
26 1
27 1
28 1
29 1
30 1
31 1
32 1
33 1
34 1
35 1
36 1
37 1
38 1
39 1
40 1
41 1
42 1
43 1
44 1
45 1
46 1
47 1
48 1
49 1
50 1
51 1
52 1
53 1
54 1
55 1
56 1
57 1
58 1
59 1
60 1
61 1
62 1
63 1
64 0

Failed with t←? 1000 ⍴ 64

31 12 9 10 50 26 19 11 47 34 58 45 46 9 64 54 4 4 38 25 12 54 29 24 45 44 32 13 
57 54 48 40 27 43 50 51 37 11 14 15 3 32 5 31 55 64 36 11 6 5 13 21 5 47 37 39 
58 30 39 46 51 46 12 58 43 28 52 59 2
  4 56 10 18 11 48 60 36 49 10 14 54 56 63 9 18 52 37 4 5 33 63 61 1 31 19 
9 28 12 17 56 59 35 46 57 60 29 22 52 15 23 44 2 36 27 7 52 22 56 41 63 64 32 
33 35 30 15 49 29 45 12 59 52 55 55 16
  17 3 24 12 63 37 14 20 13 4 24 40 55 30 10 28 13 64 41 21 52 32 23 11 50 
8 35 16 19 20 61 38 51 14 39 29 8 11 63 61 47 56 63 9 50 5 13 62 24 58 39 12 39 
26 51 57 57 22 6 3 13 52 56 54 10 52
  18 20 12 49 14 4 13 23 37 27 50 37 60 11 54 49 10 54 47 26 45 50 7 22 54 
5 18 30 2 17 64 17 29 11 36 62 37 6 17 57 63 27 44 46 27 43 20 27 14 52 20 39 
12 49 29 44 39 26 40 2 61 13 36 39 58 12
  62 50 19 27 4 6 27 5 63 29 33 59 38 48 54 12 43 41 26 37 20 50 50 43 27 
11 59 29 21 56 9 49 44 21 63 31 6 63 55 45 33 53 51 41 51 45 40 44 3 45 56 34 
14 62 44 54 13 15 43 12 24 32 11 44 27 14
  44 27 49 36 57 21 64 45 16 45 61 55 38 8 37 33 27 25 34 19 60 9 27 46 50 
42 55 4 19 29 13 21 53 22 64 8 41 56 2 10 17 14 25 27 35 64 18 44 61 17 49 15 
40 63 27 29 25 28 19 34 3 43 9 47 25 25
  25 7 33 61 15 60 55 13 30 64 46 5 23 48 44 45 12 62 36 18 9 48 46 60 59 
28 37 42 62 48 35 60 35 28 18 24 62 14 7 12 44 37 62 13 61 51 47 22 19 45 17 42 
39 34 11 52 13 11 63 19 34 19 62 7 35
  44 19 46 54 19 27 50 60 30 17 46 34 13 14 24 53 61 52 44 23 64 7 10 11 52 
10 11 26 25 12 40 17 31 25 57 7 31 25 48 55 57 39 45 60 11 12 19 8 53 6 8 22 54 
39 52 50 61 2 32 29 60 3 22 19 45 49
  38 55 46 23 55 28 15 20 49 6 63 13 51 20 27 57 60 61 30 35 54 40 13 7 49 
44 62 1 43 59 47 34 36 27 1 6 10 12 38 60 7 6 1 16 46 47 22 15 49 35 41 29 20 
30 3 39 50 22 49 46 9 36 20 32 23 12 63
  25 51 53 34 3 18 52 52 38 23 37 9 59 64 1 37 60 39 11 17 4 42 19 17 40 28 
38 45 19 63 9 5 26 37 62 30 43 17 63 38 25 42 55 53 10 5 33 34 48 35 2 55 20 20 
60 52 63 59 44 18 2 50 37 22 9 29 12
  2 55 3 15 28 30 37 16 2 10 46 43 59 60 26 62 8 54 30 15 11 33 53 15 42 44 
8 26 55 12 57 3 1 54 20 61 60 17 50 59 30 48 53 49 55 46 31 38 24 34 6 15 35 40 
17 14 47 38 25 23 53 15 40 22 2 18 15
  54 4 56 38 14 64 12 46 60 8 37 62 48 16 22 51 31 43 41 36 50 63 19 22 17 
37 38 51 4 32 7 25 44 49 17 42 9 26 57 60 49 24 2 5 29 59 35 42 47 10 61 9 11 
59 9 14 24 18 49 25 56 20 29 31 36 25 53
  42 48 60 9 56 64 31 55 42 27 47 57 6 15 6 33 54 47 11 37 63 63 46 42 1 61 
22 8 5 2 62 61 5 5 2 59 62 50 63 45 21 5 28 40 58 2 50 50 4 28 8 26 29 4 21 43 
60 24 45 39 23 52 40 56 60 40 19 21 55
  7 37 5 53 26 9 55 25 13 46 62 13 51 40 17 10 49 25 57 35 55 18 5 45 26 19 
38 45 2 17 48 45 19 45 43 64 10 28 17 16 20 54 20 13 56 49 21 62 56 44 35 46 52 
24 35 18 63 53 30 45 8 37 63 37 19 16
  33 64 27 31 7 41 15 7 2 28 44 31 54 43 25 26 43 23 2 15 27 44 23 50 29 19 
11 31 49 36 51 17 15 47 57 27 51 11 58 55 28 3 19 59 55 53 16 25 28 25 21 3 28 
42 2 52 60 13 11 47 37 39 2 3 60 10 17

  test2

 3 True
 3 6 2   3 6 2
 4 8 7   4 8 7
 1 9 5   1 9 5

 4 True
  2  7 15 142  7 15 14
  5  3 10 135  3 10 13
  8 12  6  18 12  6  1
 11  4  9 16   11  4  9 16

 5 True
 16 20 23 13 12   16 20 23 13 12
 14 15  8 18  5   14 15  8 18  5
  3  6  2  7  43  6  2  7  4
 10 21 24 19 25   10 21 24 19 25
  9 17 11  1 229 17 11  1 22

 6 True
  4 28 14 23 16 124 28 14 23 16 12
 26 29 17  2 10  9   26 29 17  2 10  9
 15 19 20 18 27 33   15 19 20 18 27 33
  5 30  6  8  1 255 30  6  8  1 25
 31 21 36 11 34 22   31 21 36 11 34 22
  7  3 32 24 35 137  3 32 24 35 13

 7 True
 45 11 25 24 33 39  4   45 11 25 24 33 39  4
  1 38 17 10  7 46 401 38 17 10  7 46 40
 28 19  6 48 22 44 4

Re: Potential problem in diadic iota (⍳) or in Pick (⊃)

2020-08-07 Thread Hans-Peter Sorge 

This is some intermittent bug..

with less than 255 elements the all elements are always equal
  t←? 255 ⍴ i
  +/t = 2÷⍨ {⍵×2} fasteach t
255
  t←? 512 ⍴ i
  +/t = 2÷⍨ {⍵×2} fasteach t
50
  t←? 300 ⍴ i
  +/t = 2÷⍨ {⍵×2} fasteach t
9

  256 elements  random number of elements are equal ...
  t←? 256 ⍴ i
  +/t = 2÷⍨ {⍵×2} fasteach t
5
  the again all are identical.
  t←? 256 ⍴ i
  +/t = 2÷⍨ {⍵×2} fasteach t
256

Regards Hans-Peter

Am 07.08.20 um 02:50 schrieb Christian Robert:

Can someone help me debug this function (fasteach). I'm not the author
but from what I've read it may be a faster than the "each" Operator 
especially

when a few unique numbers in a large array


  ∇z←test;⎕RL;t;i;r
⎕RL←0
i←20
Loop:
t←? 1000 ⍴ i
i,r←t ≡ 2÷⍨ {⍵×2} fasteach t
→(r=0)/Fail
i←i+1
→Loop
Fail:
''
'Failed with t←? 1000 ⍴', (⍕i)
''
z←t
∇

  ∇z←(F fasteach) A;shp;vec;nub;res;idx
shp←⍴A   ⍝ Shape and ...
vec←,A   ⍝ ... ravel of arg.
nub←∪ vec    ⍝ Vector of unique elements.
nub←∪ vec    ⍝ Vector of unique elements.
res←F¨nub    ⍝ Result for unique elts.
idx←nub ⍳ vec    ⍝ Indices of arg in nub ...
z←shp ⍴ idx ⊃¨⊂ res  ⍝ ... distribute result.
∇


  test
20 1
21 1
22 1
23 1
24 1
25 1
26 1
27 1
28 1
29 1
30 1
31 1
32 1
33 1
34 1
35 1
36 1
37 1
38 1
39 1
40 1
41 1
42 1
43 1
44 1
45 1
46 1
47 1
48 1
49 1
50 1
51 1
52 1
53 1
54 1
55 1
56 1
57 1
58 1
59 1
60 1
61 1
62 1
63 1
64 0

Failed with t←? 1000 ⍴64

31 12 9 10 50 26 19 11 47 34 58 45 46 9 64 54 4 4 38 25 12 54 29 24 45 
44 32 13 57 54 48 40 27 43 50 51 37 11 14 15 3 32 5 31 55 64 36 11 6 5 
13 21 5 47 37 39 58 30 39 46 51 46 12 58 43 28 52 59 2
  4 56 10 18 11 48 60 36 49 10 14 54 56 63 9 18 52 37 4 5 33 63 61 
1 31 19 9 28 12 17 56 59 35 46 57 60 29 22 52 15 23 44 2 36 27 7 52 22 
56 41 63 64 32 33 35 30 15 49 29 45 12 59 52 55 55 16
  17 3 24 12 63 37 14 20 13 4 24 40 55 30 10 28 13 64 41 21 52 32 
23 11 50 8 35 16 19 20 61 38 51 14 39 29 8 11 63 61 47 56 63 9 50 5 13 
62 24 58 39 12 39 26 51 57 57 22 6 3 13 52 56 54 10 52
  18 20 12 49 14 4 13 23 37 27 50 37 60 11 54 49 10 54 47 26 45 50 
7 22 54 5 18 30 2 17 64 17 29 11 36 62 37 6 17 57 63 27 44 46 27 43 20 
27 14 52 20 39 12 49 29 44 39 26 40 2 61 13 36 39 58 12
  62 50 19 27 4 6 27 5 63 29 33 59 38 48 54 12 43 41 26 37 20 50 
50 43 27 11 59 29 21 56 9 49 44 21 63 31 6 63 55 45 33 53 51 41 51 45 
40 44 3 45 56 34 14 62 44 54 13 15 43 12 24 32 11 44 27 14
  44 27 49 36 57 21 64 45 16 45 61 55 38 8 37 33 27 25 34 19 60 9 
27 46 50 42 55 4 19 29 13 21 53 22 64 8 41 56 2 10 17 14 25 27 35 64 
18 44 61 17 49 15 40 63 27 29 25 28 19 34 3 43 9 47 25 25
  25 7 33 61 15 60 55 13 30 64 46 5 23 48 44 45 12 62 36 18 9 48 
46 60 59 28 37 42 62 48 35 60 35 28 18 24 62 14 7 12 44 37 62 13 61 51 
47 22 19 45 17 42 39 34 11 52 13 11 63 19 34 19 62 7 35
  44 19 46 54 19 27 50 60 30 17 46 34 13 14 24 53 61 52 44 23 64 7 
10 11 52 10 11 26 25 12 40 17 31 25 57 7 31 25 48 55 57 39 45 60 11 12 
19 8 53 6 8 22 54 39 52 50 61 2 32 29 60 3 22 19 45 49
  38 55 46 23 55 28 15 20 49 6 63 13 51 20 27 57 60 61 30 35 54 40 
13 7 49 44 62 1 43 59 47 34 36 27 1 6 10 12 38 60 7 6 1 16 46 47 22 15 
49 35 41 29 20 30 3 39 50 22 49 46 9 36 20 32 23 12 63
  25 51 53 34 3 18 52 52 38 23 37 9 59 64 1 37 60 39 11 17 4 42 19 
17 40 28 38 45 19 63 9 5 26 37 62 30 43 17 63 38 25 42 55 53 10 5 33 
34 48 35 2 55 20 20 60 52 63 59 44 18 2 50 37 22 9 29 12
  2 55 3 15 28 30 37 16 2 10 46 43 59 60 26 62 8 54 30 15 11 33 53 
15 42 44 8 26 55 12 57 3 1 54 20 61 60 17 50 59 30 48 53 49 55 46 31 
38 24 34 6 15 35 40 17 14 47 38 25 23 53 15 40 22 2 18 15
  54 4 56 38 14 64 12 46 60 8 37 62 48 16 22 51 31 43 41 36 50 63 
19 22 17 37 38 51 4 32 7 25 44 49 17 42 9 26 57 60 49 24 2 5 29 59 35 
42 47 10 61 9 11 59 9 14 24 18 49 25 56 20 29 31 36 25 53
  42 48 60 9 56 64 31 55 42 27 47 57 6 15 6 33 54 47 11 37 63 63 
46 42 1 61 22 8 5 2 62 61 5 5 2 59 62 50 63 45 21 5 28 40 58 2 50 50 4 
28 8 26 29 4 21 43 60 24 45 39 23 52 40 56 60 40 19 21 55
  7 37 5 53 26 9 55 25 13 46 62 13 51 40 17 10 49 25 57 35 55 18 5 
45 26 19 38 45 2 17 48 45 19 45 43 64 10 28 17 16 20 54 20 13 56 49 21 
62 56 44 35 46 52 24 35 18 63 53 30 45 8 37 63 37 19 16
  33 64 27 31 7 41 15 7 2 28 44 31 54 43 25 26 43 23 2 15 27 44 23 
50 29 19 11 31 49 36 51 17 15 47 57 27 51 11 58 55 28 3 19 59 55 53 16 
25 28 25 21 3 28 42 2 52 60 13 11 47 37 39 2 3 60 10 17


I don't know why it fail here.

I suspect an error either in "idx←nub ⍳ vec" or
in "idx ⊃¨⊂ res"

eg: in diadic iota (⍳) or in Pick (⊃)

will continue to search,
If someone can confirm it may help...

Xtian.

Re: Potential problem in diadic iota (⍳) or in Pick (⊃)

2020-08-07 Thread Kacper Gutowski 
Indeed, as Christian wrote, the fix is not sufficient.  But I still 
believe it's in the branch I previously mentioned as commenting out the 
whole optimization by flipping #if 1 to 0 at PrimitiveFunction.cc:319 
fixes it as far as I can see (at cost of performance of course).


Other examples in the same vein still failing at r1334:

  A←64?64 ⋄ B←⍳6 ⋄ B≡A[A⍳B] ⍝ this should be 1
0
  A←3×⍳64 ⋄ B←⍳6 ⋄ B≡A[A⍳B] ⍝ this should be INDEX ERROR
0
  A←⎕IO+64|3×⍳64 ⋄ B←⍳6 ⋄ B≡A[A⍳B]  ⍝ this should be 1
0

-k

Question for the list...

2020-08-07 Thread Chris Moller 
I've just discovered that you get a syntax error if you try to localise 
a system variable in a named lambda:


   lll←{⍳⍵⊣⎕io←0;⎕io}
   lll 8
   SYNTAX ERROR
   lll[1]  λ←⍳⍵⊣⎕IO←0;⎕IO
 ^  ^

Is it supposed to work that way?  Or is that a bug?