Thread overview
why does phobos use [0, 5, 8, 9][] instead of [0, 5, 8, 9] in examples?
Apr 08, 2015
Timothee Cour
Apr 08, 2015
Vladimir Panteleev
Apr 08, 2015
Ali Çehreli
April 08, 2015
Eg, code like this in std.algorithm:
assert(equal(setSymmetricDifference(a, b), [0, 5, 8, 9][]));
why not just:
assert(equal(setSymmetricDifference(a, b), [0, 5, 8, 9]));
?


April 08, 2015
On Wednesday, 8 April 2015 at 02:40:14 UTC, Timothee Cour wrote:
> Eg, code like this in std.algorithm:
> assert(equal(setSymmetricDifference(a, b), [0, 5, 8, 9][]));
> why not just:
> assert(equal(setSymmetricDifference(a, b), [0, 5, 8, 9]));
> ?

It's historic. DMD 2.041 changed the type of array literals from int[N] (i.e. fixed-length literals of static arrays) to int[] (dynamic arrays). Since [] means to take a slice of the entire static array, it would tell the compiler to create a static array, but only pass a slice of it to equal() instead of passing the entire static array by-value.
April 08, 2015
On 04/07/2015 07:40 PM, Timothee Cour via Digitalmars-d-learn wrote:
> Eg, code like this in std.algorithm:
> assert(equal(setSymmetricDifference(a, b), [0, 5, 8, 9][]));
> why not just:
> assert(equal(setSymmetricDifference(a, b), [0, 5, 8, 9]));
> ?
>

It must be a leftover from the time when the type of an array literal was a static array (versus a dynamic one). Since static arrays are not ranges, the author of the code apparently has first created a slice to its elements.

However, today array literals are already slices.

Ali