void foo(int[int] bar)
{
    // ...
}



Is it possible to send an empty array literal?

foo( [ 0 : 2 ] ) works
foo( [] ) doesn't

int[int] empty;
foo(empty);
works but it's two lines

On Thursday, 9 July 2020 at 19:53:42 UTC, JN wrote:
> void foo(int[int] bar)
> {
>     // ...
> }
>
>
>
> Is it possible to send an empty array literal?
>
> foo( [ 0 : 2 ] ) works
> foo( [] ) doesn't
>
> int[int] empty;
> foo(empty);
> works but it's two lines

Hmm, foo(null) seems to work, but is it correct way to do it?

On Thursday, 9 July 2020 at 19:53:42 UTC, JN wrote:

>
> foo( [] ) doesn't

Should work in principle, but you can foo(null) to work around.

On Thursday, 9 July 2020 at 19:53:42 UTC, JN wrote:
> void foo(int[int] bar)
> {
>     // ...
> }
>
>
>
> Is it possible to send an empty array literal?
>
> foo( [ 0 : 2 ] ) works
> foo( [] ) doesn't
>
> int[int] empty;
> foo(empty);
> works but it's two lines

I always did foo((int[int]).init);

On 7/9/20 4:04 PM, JN wrote:
> On Thursday, 9 July 2020 at 19:53:42 UTC, JN wrote:
>> void foo(int[int] bar)
>> {
>>     // ...
>> }
>>
>>
>>
>> Is it possible to send an empty array literal?
>>
>> foo( [ 0 : 2 ] ) works
>> foo( [] ) doesn't
>>
>> int[int] empty;
>> foo(empty);
>> works but it's two lines
> 
> Hmm, foo(null) seems to work, but is it correct way to do it?
> 

Yes, that is correct.

-Steve

On Thursday, 9 July 2020 at 20:24:11 UTC, Steven Schveighoffer wrote:

>
> Yes, that is correct.
>
> -Steve

Why isn't [] accepted as an empty AA literal?

On 7/9/20 4:31 PM, Max Samukha wrote:
> On Thursday, 9 July 2020 at 20:24:11 UTC, Steven Schveighoffer wrote:
> 
>>
>> Yes, that is correct.
> 
> Why isn't [] accepted as an empty AA literal?

Because it's an empty dynamic array literal.

If D were to accept an empty AA literal, I'd expect it to be [:].

-Steve

On Thursday, 9 July 2020 at 20:24:11 UTC, Steven Schveighoffer wrote:
> On 7/9/20 4:04 PM, JN wrote:
>> On Thursday, 9 July 2020 at 19:53:42 UTC, JN wrote:
>>> void foo(int[int] bar)
>>> {
>>>     // ...
>>> }
>>>
>>>
>>>
>>> Is it possible to send an empty array literal?
>>>
>>> foo( [ 0 : 2 ] ) works
>>> foo( [] ) doesn't
>>>
>>> int[int] empty;
>>> foo(empty);
>>> works but it's two lines
>> 
>> Hmm, foo(null) seems to work, but is it correct way to do it?
>> 
>
> Yes, that is correct.
>
> -Steve

Interesting. Often in D discussion, an argument pops up that the language should be protecting against hidden breakages from API changes. This would be an example of that happening.

void foo(int[int] bar), someone calls it with a null, suddenly the signature changes to void foo(int* bar) and you will be sending a null pointer and possibly breaking the app.

On 7/9/20 5:13 PM, JN wrote:
> On Thursday, 9 July 2020 at 20:24:11 UTC, Steven Schveighoffer wrote:
>> On 7/9/20 4:04 PM, JN wrote:
>>> Hmm, foo(null) seems to work, but is it correct way to do it?
>>>
>>
>> Yes, that is correct.
>>
> 
> Interesting. Often in D discussion, an argument pops up that the language should be protecting against hidden breakages from API changes. This would be an example of that happening.
> 
> void foo(int[int] bar), someone calls it with a null, suddenly the signature changes to void foo(int* bar) and you will be sending a null pointer and possibly breaking the app.

This is a stretch.

This means you NEVER call it with an actual associative array (Which would fail to compile), and that foo never expects to get a null pointer. Even if it does break, it breaks by segfaulting and not corrupting your program.

All this, plus the author of foo cares nothing for his users, who now suddenly have non-compiling code.

-Steve

On Thursday, 9 July 2020 at 20:08:47 UTC, Anonymouse wrote:
> On Thursday, 9 July 2020 at 19:53:42 UTC, JN wrote:
>> void foo(int[int] bar)
>> {
>>     // ...
>> }
>>
>>
>>
>> Is it possible to send an empty array literal?
>>
>> foo( [ 0 : 2 ] ) works
>> foo( [] ) doesn't
>>
>> int[int] empty;
>> foo(empty);
>> works but it's two lines
>
> I always did foo((int[int]).init);

Isn't that just 'null'.

I want to make note that you cannot pass null, modify the aa, and expect the parent stack to see those changes. Since you aren't using a variable that is null you are fine.