Char literals

[Steven Schveighoffer]

On Fri, 06 Nov 2009 06:24:51 -0500, Eldar Insafutdinov  
<e.insafutdinov at gmail.com> wrote:

> I've just written a simple replicate function (the one from std.range  
> does not work in CTFE):
> T[] replicate(T)(int n, T value)
> {
>     T[] ret;
>     if (n > 0)
>     {
>         for(int i = 0; i < n; i++)
>             ret ~= value;
>     }
>     return ret;
> }
> it's not the best implementation, but that's not the thing I want to say  
> today. When I call it
>
> string str = replicate(5, 'a');
>
> I get
>
> moc.d(370): Error: cannot implicitly convert expression  
> (replicate(5,',')) of type char[] to string
>
> That means that char literals are "char", but not "immutable(char)". I  
> remember the whole conversation about array literals, but char literals  
> are closer to those of the string. What's the reasoning behind leaving  
> them as mutable char?

a value type such as char makes no sense as an immutable entity.  First of  
all, an immutable char can be implicitly cast to and from a mutable char.   
Although it might seem incorrect, since it is a value type, assigning it  
to an immutable variable makes a copy anyways, so the  
mutability/immutability of the original variable stays intact.

You bring up a legitimate concern.  However, you are probably never going  
to see 'a' by itself typed as an immutable char, because mutable makes the  
most sense as the default constancy of a value type literal.  It seems  
that string literals being typed as immutable(char)[] is an inconsistency  
in the world of D literals that always needs to be handled specially.   
What you probably want to have is a specialization for your template.

Again, another use case for unique(T)...

BTW, you can save a loop by doing this:

T[] ret = new T[n];
ret[] = value; // assign value to all elements of ret

-Steve