From: Lyle Johnson <lyle@...>
Date: 2003-02-13T09:06:36+09:00
Subject: Re: Should substrings be modifiable in-place??

Hal E. Fulton wrote:

> I've been thinking about this for
> a half hour or so.
> 
> Rather than writing this kind of thing:
> 
>   str[x,y] = str[x,y].reverse
>   str[x..z] = str[x..z].reverse
> 
> I'd like to be able to do this:
> 
>   str[x,y].reverse!
>   str[x..z].reverse!
> 
> There is probably some good reason this
> won't work, though. ;)
> 
> Someone care to comment?

The current semantics are that this:

     str[x, y] = ... some string ...

modifies str in place, but this:

     str[x, y]

returns a *new* String object (and that's the clincher) which is a slice 
from 'str'.

So if str[x, y] returns a new String, then I would *expect* that

     str[x, y].reverse!

instantiates a new String, then calls String#reverse! to perform an 
in-place reverse on that new string -- but still leaves the original 
string (str) unmodified.