looks like the problem likes in
https://github.com/openj/core/blob/18fd23bbdc2f50770eb3047e978cd5e4e3b47039/vrand.c#L726
It looks like it checks to see if the first number in the list is 2
and if so, it uses the roll2 implementation
RZ(w=vi(w)); m=*AV(w);
if( 2==m)RZ(z=roll2 (w,&b));
It also appears to be a 64-bit bug
32-bit:
?. 16 # 2 2 2 16
1 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1
1 1 1 1 0 1 1 1 0 0 1 1 1 13 12 13 4 9 15 5 9 6 3 9 1 2 11 1 8
64-bit
?. 16 # 2 2 2 16
0 1 0 1 1 0 0 1 1 1 0 0 0 0 1 0 0 1 0 1 1 0 0 0 0 0 1 0 0 1 0 1 0 0 0
0 0 0 1 0 1 1 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 1 1 1 0 1 1
The simplest 'fix' would be to disable the roll 2 optimization by
commenting it out, but that may not be ideal since presumably it was
put in place for a reason to speed up random booleans on large arrays.
The next fix would be to check to see if all numbers in the ravel are
2 and only then use roll2. The last, probably more complicated fix
would be to support 2 and other numbers in roll2 safely in 64-bit, but
even that is pretty simple
just copy and paste the block of shifts (since there is only 8), to
add 8 more for 64-bit... need to do this for the random and fixed seed
part
}else for(j=0;j<q;++j){
t=NEXT;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t;
//start new code
#if SY_64
t>>=1;
if(!(2==*v++&&2==*v++&&2==*v++&&2==*v++))R mark; *zv++=mk&t; t>>=1;
if (!(2 == *v++ && 2 == *v++ && 2 == *v++ && 2 == *v++))R mark;
*zv++ = mk&t; t >>= 1;
if (!(2 == *v++ && 2 == *v++ && 2 == *v++ && 2 == *v++))R mark;
*zv++ = mk&t; t >>= 1;
if (!(2 == *v++ && 2 == *v++ && 2 == *v++ && 2 == *v++))R mark;
*zv++ = mk&t; t >>= 1;
if (!(2 == *v++ && 2 == *v++ && 2 == *v++ && 2 == *v++))R mark;
*zv++ = mk&t; t >>= 1;
if (!(2 == *v++ && 2 == *v++ && 2 == *v++ && 2 == *v++))R mark;
*zv++ = mk&t; t >>= 1;
if (!(2 == *v++ && 2 == *v++ && 2 == *v++ && 2 == *v++))R mark;
*zv++ = mk&t; t >>= 1;
if (!(2 == *v++ && 2 == *v++ && 2 == *v++ && 2 == *v++))R mark; *zv++ = mk&t;
#endif
}
On Thu, Dec 17, 2015 at 10:29 PM, Linda A Alvord
<[email protected]> wrote:
> These are the results you should get in a fresh version of JHS or jqt. If
> you use ?. you will insure repeated exact results in a session. In all cases
> the length of the list is 64 elements so there should be enough 2's.
>
> ?. 16 # 2 2 2 16
> 1 0 0 1 0 1 0 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1 1 1
> 1 0 1 1 1 0 0 1 1 1 13 12 13 4 9 15 5 9 6 3 9 1 2 11 1 8
> ?. 16 # 3 2 2 16
> 0 2 2 1 2 0 2 1 1 2 0 2 1 1 1 2 0 0 0 0 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0 1 1 1
> 1 1 1 0 1 1 1 0 0 1 13 11 13 12 13 4 9 15 5 9 6 3 9 1 2 11
> ?. 16 # 16 2 2 2
> 2 11 15 4 6 7 4 1 4 2 6 14 1 14 7 12 0 0 0 0 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0 1
> 1 1 1 1 1 0 1 1 1 0 0 1 1 1 1 0 1 0 1 1 1 1 0 1 1 1 0 1
>
> $?. 16 # 2 2 2 16
> 64
> $?. 16 # 3 2 2 16
> 64
> $?. 16 # 16 2 2 2
> 64
>
> Linda
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of 'Pascal
> Jasmin' via Programming
> Sent: Thursday, December 17, 2015 12:41 PM
> To: [email protected]
> Subject: Re: [Jprogramming] bug in ?
>
> ?. 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
> 2 2 2 2 2 2 2 2 2 2 2 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16
> 0 0 0 0 1 1 0 0 0 1 1 1 0 0 0 0 1 0 0 0 0 1 1 1 1 1 1 0 1 1 1 0 1 0 1 1 1 1
> 0 1 1 1 1 0 0 1 1 1 8 9 6 8 14 7 12 9 1 15 2 15 3 13 3
>
> ?. 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
> 2 2 2 2 2 2 2 2 2 2 2 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16 16
> 0 1 0 1 1 0 0 1 1 1 0 0 0 0 1 0 0 1 0 1 1 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0
> 1 0 1 1 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 1 1 1 0 1 1
>
>
> one less leading 2 in first example. 47 leading 2s, correct result. 48
> leading 2s and everything is 2.
>
>
> ----- Original Message -----
> From: Raul Miller <[email protected]>
> To: Programming forum <[email protected]>
> Sent: Thursday, December 17, 2015 12:28 PM
> Subject: Re: [Jprogramming] bug in ?
>
> Consider
> ?.2 2 2 32
> 0 1 0 18
>
> Thanks,
>
> --
> Raul
>
>
> On Thu, Dec 17, 2015 at 12:13 PM, 'Pascal Jasmin' via Programming
> <[email protected]> wrote:
>>
>>
>>
>> ?. 16 # 2 2 2 16
>>
>>
>> result is same as
>>
>> ?. 16 # 2 2 2 2
>>
>> ignores the request for 0-15 range random numbers.
>>
>>
>> ----- Original Message -----
>> From: Raul Miller <[email protected]>
>> To: Programming forum <[email protected]>
>> Sent: Thursday, December 17, 2015 12:10 PM
>> Subject: Re: [Jprogramming] bug in ?
>>
>> On Thu, Dec 17, 2015 at 11:49 AM, 'Pascal Jasmin' via Programming
>> <[email protected]> wrote:
>>> ?. 16 # 2 2 2 16
>>> 0 1 0 1 1 0 0 1 1 1 0 0 0 0 1 0 0 1 0 1 1 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0
> 0 1 0 1 1 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 1 1 1 0 1 1
>>>
>>>
>>> ?. 16 # 3 2 2 16
>>> 0 2 1 0 2 1 1 1 2 1 0 2 1 2 1 2 1 0 1 1 1 1 1 1 0 0 1 0 0 1 1 0 0 1 1 1 0
> 1 0 0 0 1 1 1 1 0 1 0 8 4 14 3 1 6 14 10 15 13 5 6 2 14 0 5
>>>
>>> The last 16 numbers in the first example should match the 2nd.
>>
>> Why do you say that?
>>
>> Thanks,
>>
>> --
>> Raul
>>
>>
>>
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/forums.htm
>
>
>
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/forums.htm
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
