# Recursive method and references

**URL:** <https://www.ruby-forum.com/t/recursive-method-and-references/99156>\
**Category:** Ruby\
**Created:** [July 20, 2006, 5:46pm UTC](https://www.ruby-forum.com/t/recursive-method-and-references/99156 "2006-07-20T17:46:30Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![sci000](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/7ba0ec/32.png) [@sci000](https://www.ruby-forum.com/u/sci000)\
**Post date:** [July 20, 2006, 5:46pm UTC](https://www.ruby-forum.com/t/recursive-method-and-references/99156/1 "2006-07-20T17:46:30Z")

</div>

Hello @ all,

I just wanted to realize a small heuristic search algorithm, but it  
doesnâ??t work as I expected it to be. Here is the problem:

The test function f works properly:  
def f(a, i=0)  
(puts “—”; return) if i==10  
(0…3).each do |x|  
b=a.dup  
next if x\>=1  
b[0]=i+x  
f(b.dup, i+1)  
print b, “\n”  
end  
end

a=[1,2,3]  
a.push(4)  
print “#”, a, “\n”  
f(a.dup)  
print “#”, a, “\n”

## The output is: #1234

9234  
8234  
7234  
6234  
5234  
4234  
3234  
2234  
1234  
0234  
#1234

Exactly what I want!

But the real one doesnâ??t:  
def tyra(arr, pos2, tmp2, resul, res=0, i=0)  
(puts “#####”; resul.push(res); return) if i\>=13  
(0…3).each do |y|  
pos=pos2.dup; tmp=tmp2.dup  
#~ tmp[0]=(tmp[0]+1)%7 if y==0  
#~ tmp[0]=(tmp[0]-1)%7 if y==1  
#~ tmp[1]=(tmp[1]+1)%8 if y==2  
#~ tmp[1]=(tmp[1]-1)%8 if y==3  
next if pos[tmp[0]][tmp[1]]==1  
pos[tmp[0]][tmp[1]]=1  
tyra(arr, pos.dup, tmp.dup, resul, res+arr[tmp[0]][tmp[1]], i+1)  
print “\*”, pos2, “\n”, “\_”, pos, “\n”  
end  
end

arr, pos, tmp, resul = [], [], [], []  
arr.push([32,80,19,98,01,90,14,85])  
arr.push([66,22,73,52,72,57,83,31])  
arr.push([30,84,41,73,16,74,45,92])  
arr.push([77,06,70,24,00,28,67,11])  
arr.push([32,99,44,81,27,75,42,98])  
arr.push([68,21,72,56,59,42,75,17])  
arr.push([34,87,19,92,05,99,27,88])

arr[0].length.times{tmp.push(0)}  
arr.length.times{pos.push(tmp.dup)}  
tmp2=[3,4]

print “#”, pos, “\n”  
tyra(arr, pos.dup, tmp2.dup, resul)  
print “#”, pos, “\n”

The output is:  
#00000000000000000000000000000000000000000000000000000000  
\*00000000000000000000000000001000000000000000000000000000  
\_00000000000000000000000000001000000000000000000000000000  
#00000000000000000000000000001000000000000000000000000000

The lines marked with a “#” does not equal each other as with the test  
functionâ?¦  
What goes wrong, whereâ??s the bug?

Thanks to all your answers in advance.

---

<div class="post-metadata">

**Author:** ![sci000](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/7ba0ec/32.png) [@sci000](https://www.ruby-forum.com/u/sci000)\
**Post date:** [July 20, 2006, 9:34pm UTC](https://www.ruby-forum.com/t/recursive-method-and-references/99156/2 "2006-07-20T21:34:23Z")

</div>

Sci000 Lem. wrote:  
…

> def tyra(arr, pos2, tmp2, resul, res=0, i=0)  
> (puts “#####”; resul.push(res); return) if i\>=13  
> (0…3).each do |y|  
> pos=pos2.dup; tmp=tmp2.dup  
> …  
> next if pos[tmp[0]][tmp[1]]==1  
> pos[tmp[0]][tmp[1]]=1  
> tyra(arr, pos.dup, tmp.dup, resul, res+arr[tmp[0]][tmp[1]], i+1)  
> print “\*”, pos2, “\n”, “\_”, pos, “\n”  
> end  
> end
> 
> arr, pos, tmp, resul = [], [], [], []  
> arr.push([32,80,19,98,01,90,14,85])  
> arr.push([66,22,73,52,72,57,83,31])  
> arr.push([30,84,41,73,16,74,45,92])  
> arr.push([77,06,70,24,00,28,67,11])  
> arr.push([32,99,44,81,27,75,42,98])  
> arr.push([68,21,72,56,59,42,75,17])  
> arr.push([34,87,19,92,05,99,27,88])
> 
> arr[0].length.times{tmp.push(0)}  
> arr.length.times{pos.push(tmp.dup)}  
> tmp2=[3,4]
> 
> print “#”, pos, “\n”  
> tyra(arr, pos.dup, tmp2.dup, resul)  
> print “#”, pos, “\n”
> 
> The output is:  
> #00000000000000000000000000000000000000000000000000000000  
> \*00000000000000000000000000001000000000000000000000000000  
> \_00000000000000000000000000001000000000000000000000000000  
> #00000000000000000000000000001000000000000000000000000000
> 
> The lines marked with a “#” does not equal each other as with the test  
> functionâ?¦  
> What goes wrong, whereâ??s the bug?

Your pos array contains references to other arrays,  
and pos.dup is a shallow copy, thus pos2 contains refs  
to the same sub-arrays as pos. Any changes made via  
access to pos2 affects the same sub-arrays as viewed  
from pos.

Clear as mud? 9^)

Cheers  
Chris

---

<div class="post-metadata">

**Author:** ![sci000](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/7ba0ec/32.png) [@sci000](https://www.ruby-forum.com/u/sci000)\
**Post date:** [July 21, 2006, 12:24pm UTC](https://www.ruby-forum.com/t/recursive-method-and-references/99156/3 "2006-07-21T12:24:51Z")

</div>

> Your pos array contains references to other arrays,  
> and pos.dup is a shallow copy, thus pos2 contains refs  
> to the same sub-arrays as pos. Any changes made via  
> access to pos2 affects the same sub-arrays as viewed  
> from pos.
> 
> Clear as mud? 9^)
> 
> Cheers  
> Chris

Thank you for the answer, I didnâ??t know that the subarrays in Ruby are  
just referenced to the whole array, so I was wondering why it donâ??t work  
(normally I solve this kind of problems in Mercury, and thatâ??s  
completely different)  
I rewrote the lines which built pos as following:  
arr.length.times{arr[0].length.times{pos.push(0)}}  
and everything works properly!