Array literals are weird.

[Q. Schroll]

On Saturday, 1 May 2021 at 19:49:51 UTC, russhy wrote:
> On Saturday, 1 May 2021 at 12:22:50 UTC, evilrat wrote:
>> On Saturday, 1 May 2021 at 11:50:27 UTC, Adam D. Ruppe wrote:
>>>
>>> or you can use the library `.staticArray` thing to expressly 
>>> indicate your intention on the original foo
>>>
>>> foo([1,2,3].staticArray);
>>>
>>>
>>
>> Is there one in Phobos? Anyway this works, but maybe it can be 
>> made more type safe with constraints or just with improved 
>> symtax.
>>
>> ```d
>> // compiles with -vgc and -betterC (weird, also without import 
>> stdc.core.stdio)
>> import std.range : ElementType;
>>
>> template staticArray(alias T)
>> {
>> 	enum ElementType!(typeof(T))[T.length] staticArray = T;
>> }
>>
>> extern(C) void main()
>> {
>> 	import core.stdc.stdio;
>> 	auto arr = staticArray!([1,2,3]);
>> 	pragma(msg, typeof(arr)); // int[3]
>> 	foreach(i; arr)
>> 		printf("%d\n", i);
>> }
>> ```
>
>
> don't you realize something is weird?
>
> someone is asking to be able to do auto a = [1, 2, 3]
>
> and you propose a template?
>
> this is asking people to look for alternative language
>
> ```D
> auto a = new[1,2,3]  <== this should be allocated array
> auto a = [1,2,3] <== this should be static array
> ```
>
> let's fix that, shall we? instead of telling people to bloat 
> their files with templates

I agree that the non-obvious allocation is kind of a problem and 
`new [ 1, 2, 3 ]` would be more explicit, but where I see the 
actual issue isn't where you think it is. In `@nogc` code, the 
compiler tells you that an allocation takes place and you look 
for solutions to avoid it (there are multiple approaches, but no 
general one).