IfType type guards

[ZachHaber]

The readme mentions that there should be a corresponding IfType guard for every IsType check.

For every IsT type (e.g. IsAny), there is an associated IfT type that can help simplify conditional types.

I'm only seeing a few IfType guards in the package (IfNever, IfAny, and IfUnknown).

Potential fixes (that I can think of at the moment):

1. Add the missing IfType guards
2. Update the readme to specifically call out those three guards

Upvote & Fund

• We're using Polar.sh so you can upvote and help fund this issue.
• The funding will be given to active contributors.
• Thank you in advance for helping prioritize & fund our backlog.

[sindresorhus]

PR welcome for types starting with Is that are missing an If type.

[sindresorhus]

I'm starting to question whether it was the correct choice to add If equivalents. It seemed nice at first, but there is now so many extra types and all of them require extra docs and maintenance. Maybe we should instead have supported the extra parameters on Is* or just referred people to a If type. Thoughts?

// @ikelax

[sindresorhus]

// @Emiyaaaaa @voxpelli

[ikelax]

Sorry for the late response.

I also think that the If types create a lot of extra work. For me, it would be fine if we support the extra parameters on Is* or just refer people to a If type. However, I would prefer later because the name helps a lot.

[ikelax]

// @sindresorhus

[voxpelli]

I'm starting to question whether it was the correct choice to add If equivalents. It seemed nice at first, but there is now so many extra types and all of them require extra docs and maintenance. Maybe we should instead have supported the extra parameters on Is* or just referred people to a If type. Thoughts?

I'm 👍 on this, and eg. the IfAny already works as IsAny if no second or third value are given:

type IfAny<T, TypeIfAny = true, TypeIfNotAny = false>

Though internally it uses IsAny:

type-fest/source/if-any.d.ts

Lines 22 to 24 in 3d473d1

	export type IfAny<T, TypeIfAny = true, TypeIfNotAny = false> = (
	IsAny<T> extends true ? TypeIfAny : TypeIfNotAny
	);

I would deprecate the Is* and refer people to the If*

[sindresorhus]

I would deprecate the Is* and refer people to the If*

Why not the other way around? Is instead of If?

[voxpelli]

No strong opinion, but IfAny<T, Foo, Bar> is more self-explanatory than IsAny<T, Foo, Bar> I think – the If* indicates that there is some kind of if-statement and as such that the two other arguments are probably depending on the first one, whereas Is* doesn't really give any indication about that, IsAny<T, Foo, Bar> sounds like the Foo and the Bar are part of the comparison