On Wednesday, 1 October 2025 at 20:00:35 UTC, monkyyy wrote:

float[10] x = [42, ...];   // [42, NaN, ..., NaN]
float[10] y = [1:42, ...]; // [NaN, 42, NaN, ..., NaN]

you can always create a mixin

On Wednesday, 1 October 2025 at 21:01:28 UTC, not you wrote:

float[10] x = [42, ...];   // [42, NaN, ..., NaN]
float[10] y = [1:42, ...]; // [NaN, 42, NaN, ..., NaN]

replacing 1 line of code with 30 isnt great

On Wednesday, 1 October 2025 at 21:28:00 UTC, monkyyy wrote:

float[10] x = [42, ...];   // [42, NaN, ..., NaN]
float[10] y = [1:42, ...]; // [NaN, 42, NaN, ..., NaN]

how often do you need this 'feature'? can you show a real world usecase?

On Monday, 8 September 2025 at 15:44:27 UTC, Walter Bright wrote:

I've been programming in D for many years because of such wonderful language features. But, in my opinion, we need to focus on stabilizing and fixing what we already have. Speaking of static arrays, I’d like to remind that we've had a long-standing bug with a segmentation fault: type inference doesn't work for a static array of type real[1].

void main() {
    import std.stdio : writeln;
    real[1] arr = 0;
    writeln(arr); // SEGFAULT
}

https://github.com/dlang/dmd/issues/21323

Correction: array output to screen does not work.

On Monday, 8 September 2025 at 15:42:27 UTC, Walter Bright wrote:

Please make this footgun an error. The compiler can tell you in the error message to use = x instead of = [x], since it’s a common pattern in C. My best bet is that no-one uses [0] being in fact [0, 0xFF …] intentionally.

On Wednesday, 15 October 2025 at 11:51:05 UTC, Quirin Schroll wrote:

Only because of the insane idea of intentionally invalid initialization

int[N]=[1,2,3] being trailing zeros makes more sense

On Monday, 8 September 2025 at 15:42:27 UTC, Walter Bright wrote:

import core.stdc.stdio;

__gshared char[10] xxx = [0]; // initialize xxx to all zeros

void main()
{
    foreach (c; xxx)
        printf("%d\n", c);
}

__gshared char[10] xxx = 0;

At least the compiler is smart enough to know that the following is an error.

void main()
{
	char[10] xxx = [0];
}

On Wednesday, 15 October 2025 at 15:22:14 UTC, monkyyy wrote:

D not copying that C behavior is sane. But what makes it bad is that D still allows you to not define all elements of the array, and implicitly initiliszes them in a non intuative way.

Here is what I'd change it to:

float[10] a = [42];   // This should be a error.
float[10] b = [0:42, 5: 3.14]; // This also should be an error.
float[10] c = [0..$: 0]; // Sets all to zero.
float[10] d = [42, 1..$: 0]; // First element is 42, everything else is zero.
float[10] e = [1..$: 42]; // Error, first element is not defined.

This should be familiar due to the slicing syntax D already has.

On Wednesday, 22 October 2025 at 16:35:47 UTC, 6622 wrote:

float[10] a = [42];   // This should be a error.
float[10] b = [0:42, 5: 3.14]; // This also should be an error.
float[10] c = [0..$: 0]; // Sets all to zero.
float[10] d = [42, 1..$: 0]; // First element is 42, everything else is zero.
float[10] e = [1..$: 42]; // Error, first element is not defined.

float[10] b = [0:42, 5: 3.14, $:13.37];//[0,13.37,13.37,13.37,13.37,3.14,13.37...