On Sat, 10 Oct 2009 02:28:42 +0400, Andrei Alexandrescu <[email protected]> wrote:

Christopher Wright wrote:
Don wrote:
I don't understand why runtime-determined array literals even exist.
They're not literals!!!
They cause no end of trouble. IMHO we'd be *much* better off without them.
You don't see the use. I do. I would go on a murderous rampage if that feature were removed from the language. For example, one thing I recently wrote involved creating a process with a large number of arguments. The invocation looked like: exec("description", [procName, arg1, arg2] ~ generatedArgs ~ [arg3, arg4] ~ moreGeneratedArgs);
 There were about ten or fifteen lines like that.
 You'd suggest I rewrite that how?
char[][] args;
args ~= procName;
args ~= arg1;
args ~= arg2;
args ~= generatedArgs;
args ~= arg3;
Just fucking shoot me. Or better yet, whoever removed array literals with non-constant elements from the language.

Relax. It's a condition known as literalitis. :o)

Literals only have you write [ a, b, c ] instead of toArray(a, b, c). I wouldn't see it a big deal one way or another, but the issue is that the former is a one-time decision that pretty much can't be changed, whereas toArray can benefit of the hindsight of experience.


Andrei

I believe it's okay, but compiler should be able to return static arrays from a function (I'll call it "array", but feel free to substitute any other function name):

foreach (i; [a, b, c]) {  --->   foreach (i, array(a, b,c )) {
    // ...                           // ...
}                                }

int[3] x = [a, b, c];     --->   int[3] x = array(a, b, c);

The first example could work with any range returned, but can you initialize a static array with a range?

Well, some language feature could allow that and translate the latter case into something like this:

int[3] x = void;
foreach (index, value; array(a, b, c)) {
    x[index] = value;
}

But then, what would you do with an array overflow? Disallow it? I.e., make it impossible to create a static array and initialize it at the same time?

If range design is essential in D, perhaps the following could be allowed:

int[] someArray = ...;
someArray[] = someRange(); // initializes an array with elements of a range. Throws an exception if boundaries don't match

And then,

int[3] x = array(a, b, c);

could be translated into

int[3] x = void;
x[] = array(a, b, c);

(unless array doesn't meet range criteria, of course).

You loose compile-time boundaries checking, but that's the best you can do if you drop [a, b, c] feature, I'm affraid.

Reply via email to