For the normalize function in std.numeric, I think an example of how to use bool normalize(R) where there is a sum value provided should be included.

The reason for this is that the examples given are as follows, and none show how the second sum parameter is given.

double[] a = [];
assert(!normalize(a));
a = [ 1.0, 3.0 ];
assert(normalize(a));
writeln(a); // [0.25, 0.75]
a = [ 0.0, 0.0 ];
assert(!normalize(a));
writeln(a); // [0.5, 0.5]

When one tries to do something like "normalize(a,10)", dmd reports this error.

> Error: template instance normalize!(no, 100) does not match template declaration normalize(R)(R range, ElementType!R sum = 1)

One must actually specify the template parameter, thus

> assert(normalise!(typeof(a)(a,100));

I guess for those more familiar with the language, this would be clear, but for those not as familiar just wanting to use a standard library function, they may be wondering why simply adding the second 'sum' parameter causes an error, or why IFTI isn't working.  I know I got a little stuck.




I also wonder why this second parameter means that DMD cannot deduce the type that (R) is.

On Saturday, 28 December 2019 at 03:28:33 UTC, BoraxMan wrote:
> For the normalize function in std.numeric, I think an example of how to use bool normalize(R) where there is a sum value provided should be included.
>
> The reason for this is that the examples given are as follows, and none show how the second sum parameter is given.
>
> double[] a = [];
> assert(!normalize(a));
> a = [ 1.0, 3.0 ];
> assert(normalize(a));
> writeln(a); // [0.25, 0.75]
> a = [ 0.0, 0.0 ];
> assert(!normalize(a));
> writeln(a); // [0.5, 0.5]
>
> When one tries to do something like "normalize(a,10)", dmd reports this error.
>
>> Error: template instance normalize!(no, 100) does not match template declaration normalize(R)(R range, ElementType!R sum = 1)
>
> One must actually specify the template parameter, thus
>
>> assert(normalise!(typeof(a)(a,100));
>
> I guess for those more familiar with the language, this would be clear, but for those not as familiar just wanting to use a standard library function, they may be wondering why simply adding the second 'sum' parameter causes an error, or why IFTI isn't working.  I know I got a little stuck.
>
>
>
>
> I also wonder why this second parameter means that DMD cannot deduce the type that (R) is.

Pull requests for documentation are usually handled quickly, and they're welcome if you've identified a hole in the existing documentation. Simply adding a new example can be handled by clicking "Improve This Page" in the upper right corner. I just did that recently and it took maybe an hour or two for my changes to be merged. If for some reason you don't want to do that, create a new issue.

On Saturday, 28 December 2019 at 03:28:33 UTC, BoraxMan wrote:
> When one tries to do something like "normalize(a,10)", dmd reports this error.
>
>> Error: template instance normalize!(no, 100) does not match template declaration normalize(R)(R range, ElementType!R sum = 1)
>
> One must actually specify the template parameter, thus
>
>> assert(normalise!(typeof(a)(a,100));

This looks like an instance of issue 1807 [1]. The fix is to move the alias template (in this case, ElementType) out of the argument list and into a template constraint:

bool normalize(R, S)(R range, S sum = 1)
    if (isForwardRange!R && is(S == ElementType!R))

You can work around this particular issue in your own code with the following wrapper function:

auto normalize(R, S)(R range, S sum)
    if (is(S == ElementType!R ))
{
    import std.numeric: normalize;
    return normalize!R(range, sum);
}

[1] https://issues.dlang.org/show_bug.cgi?id=1807

On Saturday, 28 December 2019 at 03:55:18 UTC, Paul Backus wrote:
> On Saturday, 28 December 2019 at 03:28:33 UTC, BoraxMan wrote:
>> When one tries to do something like "normalize(a,10)", dmd reports this error.
>>
>>> Error: template instance normalize!(no, 100) does not match template declaration normalize(R)(R range, ElementType!R sum = 1)
>>
>> One must actually specify the template parameter, thus
>>
>>> assert(normalise!(typeof(a)(a,100));
>
> This looks like an instance of issue 1807 [1]. The fix is to move the alias template (in this case, ElementType) out of the argument list and into a template constraint:
>
> bool normalize(R, S)(R range, S sum = 1)
>     if (isForwardRange!R && is(S == ElementType!R))
>
> You can work around this particular issue in your own code with the following wrapper function:
>
> auto normalize(R, S)(R range, S sum)
>     if (is(S == ElementType!R ))
> {
>     import std.numeric: normalize;
>     return normalize!R(range, sum);
> }
>
> [1] https://issues.dlang.org/show_bug.cgi?id=1807

Thanks for that.  I've created a pull request.  I ended up using my own version which was modified to correctly handle arrays of integers.  I think this function was specifically created only to work with floating points.  Either it should work with integers, or have a constraint where it won't accept integers.  Not sure what the original intent was.