This seems like very surprising behavior to me.  Is it a bug?

import std.typecons;
alias NT = Nullable!(Tuple!(int, double));
pragma(msg, isTuple!NT); //prints true!

On Wednesday, 2 December 2020 at 05:25:09 UTC, Ben Jones wrote:
> This seems like very surprising behavior to me.  Is it a bug?
>
> import std.typecons;
> alias NT = Nullable!(Tuple!(int, double));
> pragma(msg, isTuple!NT); //prints true!

No, this is not a bug, because Nullable!T currently has an implicit conversion to T via `alias this`. [1] However, this implicit conversion is deprecated, and will be removed in a future release. Once that happens, `isTuple!NT` will be `false`, as you'd expect.

[1] http://phobos.dpldocs.info/std.typecons.Nullable.1.html#alias-this

On Wednesday, 2 December 2020 at 12:59:52 UTC, Paul Backus wrote:
> No, this is not a bug, because Nullable!T currently has an implicit conversion to T via `alias this`. [1] However, this implicit conversion is deprecated, and will be removed in a future release. Once that happens, `isTuple!NT` will be `false`, as you'd expect.
>
> [1] http://phobos.dpldocs.info/std.typecons.Nullable.1.html#alias-this

I guess it's a moot point because the implicit conversion is deprecated, but it seems strange to me that isTuple behaves like canBeUsedAsTuple.

Seems like the reasonable thing to do with a Nullable!(Tuple!Whatever) would be to check isTuple!(TemplateArgsOf!MyNullable)

Thanks for the response