RUBY:数组和并行赋值的奇怪问题

Question

好吧,我一直在搞乱Ruby中不同的排序算法; 主要是变化快速排序.我有一个双枢轴快速排序的版本,可以选择随机枢轴.因此,当随机枢轴落在阵列的开始或结束时,奇怪的事情开始发生.我做了一些调查,并将其归结为这种奇怪的现象.

//irb output             #using Ruby 1.8.6 and irb 0.9.5
irb> foo = [1,2,3,4]     #create my array, very generic for an example
=> [1, 2, 3, 4]
irb> foo[0],foo[1],foo[2],foo[3] = foo[1],foo[0],foo[3],foo[2]
=> [2, 1, 4, 3]          #array swaps inside values with edge values fine.
irb> foo
=> [2, 1, 4, 3]          #values have changed correctly.
irb> foo = [1,2,3,4]     #reset values
=> [1, 2, 3, 4]          #next I am going to try and swap the element foo[0] with itself
irb> foo[0],foo[0],foo[2],foo[3] = foo[0],foo[0],foo[3],foo[2]
=> [1, 1, 4, 3]          #for some reason elements foo[0] and foo[1] take on the value of foo[0]
irb> foo     #check our array again
=> [1, 2, 4, 3]          #neither value foo[0] nor foo[1] are altered.

任何人都可以解释为什么会这样吗？

为了清楚起见,我不是在寻求实施快速排序的帮助.

编辑: 希望,这里使问题更清楚的是实现的样子:

# error caused when (pseudo-code) pivot1|pivot2 == start|end
foo[start], foo[pivot1], foo[pivot2], foo[end] = 
    foo[pivot1], foo[start], foo[end], foo[pivot2] 

Answer 1

这里没什么奇怪的. 请参阅下面的评论.

=> [1, 2, 3, 4]          
irb> foo[0],foo[0],foo[2],foo[3] = foo[0],foo[0],foo[3],foo[2]

#this is return value from the last expression ,that is the 
#parallel assignment , the value of that expression is [1, 1, 4, 3] , 
#not the value of foo
=> [1, 1, 4, 3]        
irb> foo     

#In the last parallel assignment, you set f[0] to f[0] , did not change f[1] ,
#you swapped f[2],f[3] .This is exactly what you did .
=> [1, 2, 4, 3]