www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - About default parameters in variadic templates

reply Andrej Mitrovic <andrej.mitrovich gmail.com> writes:
Related: http://d.puremagic.com/issues/show_bug.cgi?id=8676

This is something to think about for D3 (or maybe even D2 if it
doesn't break code): it should be possible to inject default
parameters in the function header rather than the body when using
variadic templates. It would make auto-documentation possible and
would make it easier to write template constraints. E.g. the current
function header for lockstep() in std.range is:

auto lockstep(Args...)(Args args)
if (allSatisfy!(isInputRange, staticMap!(Unqual, Args)) || (
    allSatisfy!(isInputRange, staticMap!(Unqual, Args[0..$ - 1])) &&
    is(Args[$ - 1] == StoppingPolicy))

and the body:

static if (is(Args[$ - 1] == StoppingPolicy))
{
    alias args[0..$ - 1] ranges;
    alias Args[0..$ - 1] Ranges;
    alias args[$ - 1] stoppingPolicy;
}
else
{
    alias Args Ranges;
    alias args ranges;
    auto stoppingPolicy = StoppingPolicy.shortest;
}

A lot of code could be stripped if there was a way to set an optional
last parameter for variadics. lockstep could then be written as:

auto newLockstep(Args...)(Args args, StoppingPolicy stoppingPolicy =
StoppingPolicy.shortest)
    if (allSatisfy!(isInputRange, staticMap!(Unqual, Args)))
{
    static if (Args.length > 1)
    {
        return Lockstep!(Args)(args, stoppingPolicy);
    }
    else
    {
        return args[0];
    }
}

So why doesn't this currently work? Using this template does compile:

    int[] arr1 = [1, 2, 3];
    int[] arr2 = [1, 2, 3, 4];

    foreach (x, y; newLockstep(arr1, arr2, StoppingPolicy.shortest))
    {
        writeln(x, " ", y);
    }

but if you don't specify the last StoppingPolicy argument at the call
site it won't compile:

foreach (x, y; newLockstep(arr1, arr2)) { }

test.d(32): Error: template test.newLockstep does not match any
function template declaration
test.d(14): Error: template test.newLockstep(Args...) if
(allSatisfy!(isInputRange,staticMap!(Unqual,Args))) cannot deduce
template function from argument types !()(int[],int[])
Sep 17 2012
next sibling parent Don Clugston <dac nospam.com> writes:
On 17/09/12 14:42, Andrej Mitrovic wrote:
 Related: http://d.puremagic.com/issues/show_bug.cgi?id=8676

 A lot of code could be stripped if there was a way to set an optional
 last parameter for variadics. lockstep could then be written as:

 auto newLockstep(Args...)(Args args, StoppingPolicy stoppingPolicy =
 StoppingPolicy.shortest)
      if (allSatisfy!(isInputRange, staticMap!(Unqual, Args)))
 {
      static if (Args.length > 1)
      {
          return Lockstep!(Args)(args, stoppingPolicy);
      }
      else
      {
          return args[0];
      }
 }

 but if you don't specify the last StoppingPolicy argument at the call
 site it won't compile:

 foreach (x, y; newLockstep(arr1, arr2)) { }

 test.d(32): Error: template test.newLockstep does not match any
 function template declaration
 test.d(14): Error: template test.newLockstep(Args...) if
 (allSatisfy!(isInputRange,staticMap!(Unqual,Args))) cannot deduce
 template function from argument types !()(int[],int[])

That just looks like a bug to me. Please enter it in Bugzilla.
Sep 17 2012
prev sibling parent Andrej Mitrovic <andrej.mitrovich gmail.com> writes:
On 9/18/12, Don Clugston <dac nospam.com> wrote:
 That just looks like a bug to me. Please enter it in Bugzilla.

Ok: http://d.puremagic.com/issues/show_bug.cgi?id=8687 Hope Walter agrees with it.
Sep 18 2012