Alternate Random Sort?

[Brian Yennie]

By my tests this is at least as fast as the array kludge I came up 
with. Egads, is that a bug or a feature? I know for repeat loops Rev is 
"smarter" than that and will only evaluate the expression once, but it 
does in fact appear that it recalculates the value over and over again 
with the sort command!

Bugzilla? Or is there a reason why it works this way? I guess maybe the 
"of each" constructs cause this, but it still seems like it could be 
optimized.

Either way, I think Sarah's solution is probably a lot cleaner than my 
workaround.

> Instead of asking it to compute the length of your list for every
> line, how about just using:
>   sort lines of temp by random(10000000)
> or some suitably large number?