Not as simple as I would like, but simple enough I hope:
G =:0 1 2 2 3 4 4 5 5 6 7 7 8 9 10 10 11 11 11 12 12 12
G =: G ,. 1 2 3 4 5 6 5 7 8 7 9 10 10 11 12 11 13 14 15 16 15 14
G =: <"1 G
joinroutes =: 4 : 0&.>
l =. >: (y i. {: x)
f =. 0:^:((#y)&=) (y i. {. x)
z =. (f {. y) , x , (l }. y)
NB. this is the culprit --> y E. z
y
)
jtree =: ([: ~.@; joinroutes)/
If I uncomment the indicated line, I get a loop. That implicates
numeric dyad E. .
Henry Rich
On 4/25/2012 9:54 PM, Joey K Tuttle wrote:
> Henry,
>
> I get different errors from different versions of j (e.g. 32 vs 64 bit
> J7) and, in j602/Linux a Segmentation fault - but corrupted versions of
> G I haven't seen... Did you derive a simpler demonstration of the problem?
>
> On 2012/04/25 13:51 , Henry Rich wrote:
>> I have tracked this down to numeric dyad E. which seems to cause wild
>> stores to memory, sometimes wiping out data, sometimes scripts.
>>
>> In our new open-source environment, how do bugs get fixed and released?
>>
>> Henry Rich
>>
>> On 4/24/2012 9:52 PM, Henry Rich wrote:
>>> Working on R. E.'s latest problem, I hit an interpreter bug. The
>>> session log is
>>>
>>> load 'C:\JPrograms\temp\interpbug.ijs'
>>> jtree G
>>> |domain error: jtree
>>> | jtree G
>>> X
>>> 10 12
>>> Y
>>> 12 15
>>> (<10 12) joinroutes (<12 15)
>>> ┌──────────┐
>>> │┌────────┐│
>>> ││10 12 15││
>>> │└────────┘│
>>> └──────────┘
>>> jtree G
>>> |domain error: joinroutes
>>> | y(+./@:E.}.;)(1{.y) ,x,(l}.y)
>>> jtree G
>>> |length error: joinroutes
>>> | 'f l'=.0 1+(0 1+#y) |y i.x
>>> G
>>> +---+-+---+---+---+---+--+---+---+---+---+----+----+----+-----+-----+-----+-----+-----+-----+-----+-----+
>>> |0 1||2 3|2 4|3 5|4 6| |5 7|5 8|6 7|7 9|7 10|8 10|9 11|10 12|10 11|11
>>> 13|11 14|11 15|12 16|12 15|12 14|
>>> +---+-+---+---+---+---+--+---+---+---+---+----+----+----+-----+-----+-----+-----+-----+-----+-----+-----+
>>>
>>> You can see that the variable G has been corrupted; also, the errors
>>> running jtree are bogus, I think. The script file follows. The first
>>> line is wrapped, and the name 'joinroutes' is defined and then overwritten:
>>>
>>> G =:<"1 ] 0 1 2 2 3 4 4 5 5 6 7 7 8 9 10 10 11 11 11 12 12 12 ,. 1 2 3
>>> 4 5 6 5 7 8 7 9 10 10 11 12 11 13 14 15 16 15 14
>>> NB. x is new pair
>>> NB. y is old route
>>> NB. Result is merged routes
>>> joinroutes =: 4 : 0&.>
>>> select. (#y)> 'f l' =. y i. x
>>> case. 1 1 do.
>>> y ; (f {. y) , x , ((>:l) }. y)
>>> case. 0 1 do.
>>> y ;^:(l~:0)< x , ((>:l) }. y)
>>> case. 1 0 do.
>>> y ;^:(f~:<:#y)< (f {. y) , x
>>> case. do.
>>> y ; x
>>> end.
>>> )
>>> joinroutes =: 4 : 0&.>
>>> X =: x
>>> Y =: y
>>>
>>> 'f l' =. 0 1 + (0 1 + #y) | y i. x
>>> y (+./@:E. }. ;) (f {. y) , x , (l }. y)
>>> )
>>> jtree =: ([: ~.@; joinroutes)/
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
