Address review comments

 - Clarify definition of From
 - Add section on computed fields
 - Add NamedTuples.scala in appendix

Author: odersky

content/named-tuples.md

@@ -146,6 +146,47 @@ The translation of named tuples to instances of `NamedTuple` is fixed by the spe
  - All tuple operations also work with named tuples "out of the box".
  - Macro libraries can rely on this expansion.
 
+### Computed Field Names
+
+The `Selectable` trait now has a `Fields` type member that can be instantiated
+to a named tuple.
+
+```scala
+trait Selectable:
+  type Fields <: NamedTuple.AnyNamedTuple
+```
+
+If `Fields` is instantiated in a subclass of `Selectable` to some named tuple type,
+then the available fields and their types will be defined by that type. Assume `n: T`
+is an element of the `Fields` type in some class `C` that implements `Selectable`,
+that `c: C`, and that `n` is not otherwise legal as a name of a selection on `c`.
+Then `c.n` is a legal selection, which expands to `c.selectDynamic("n").asInstanceOf[T]`.
+
+It is the task of the implementation of `selectDynamic` in `C` to ensure that its
+computed result conforms to the predicted type `T`
+
+As an example, assume we have a query type `Q[T]` defined as follows:
+
+```scala
+trait Q[T] extends Selectable:
+  type Fields = NamedTuple.Map[NamedTuple.From[T], Q]
+  def selectDynamic(fieldName: String) = ...
+```
+
+Assume in the user domain:
+```scala
+case class City(zipCode: Int, name: String, population: Int)
+val city: Q[City]
+```
+Then
+```scala
+city.zipCode
+```
+has type `Q[Int]` and it expands to
+```scala
+city.selectDynamic("zipCode").asInstanceOf[Q[Int]]
+```
+
 ### The NamedTuple.From Type
 
 The `NamedTuple` object contains a type definition
@@ -154,15 +195,19 @@ The `NamedTuple` object contains a type definition
 ```
 `From` is treated specially by the compiler. When `NamedTuple.From` is applied to
 an argument type that is an instance of a case class, the type expands to the named
-tuple consisting of all the fields of that case class. Here, fields means: elements of the first parameter section. For instance, assuming
+tuple consisting of all the fields of that case class.
+Here, _fields_ means: elements of the first parameter section. For instance, assuming
 ```scala
 case class City(zip: Int, name: String, population: Int)
 ```
 then `NamedTuple.From[City]` is the named tuple
 ```scala
 (zip: Int, name: String, population: Int)
 ```
-The same works for enum cases expanding to case classes.
+The same works for enum cases expanding to case classes, abstract types with case classes as upper bound, alias types expanding to case classes
+and singleton type with case classes as underlying type (in terms of the implementation, the `classSymbol` of a type must be a case class.
+
+`From` is also defined on named tuples. If `NT` is a named tuple type, then `From[NT] = NT`.
 
 ### Pattern Matching with Named Fields in General
 
@@ -183,6 +228,9 @@ This revives SIP 43, with a much simpler desugaring than originally proposed.
 Named patterns are compatible with extensible pattern matching simply because
 `unapply` results can be named tuples.
 
+### Operations on Named Tuples
+
+The operations on named tuples are defined in object `scala.NamedTuple`. The current version of this object is listed in the appendix.
 
 ### Restrictions
 
@@ -298,3 +346,202 @@ This section should list prior work related to the proposal, notably:
 
 ## FAQ
 
+## Appendix: NamedTuple Definition
+
+Here is the current definition of `NamedTuple`. This is part of the library and therefore subject to future changes and additions.
+
+```scala
+package scala
+import annotation.experimental
+import compiletime.ops.boolean.*
+
+@experimental
+object NamedTuple:
+
+  opaque type AnyNamedTuple = Any
+  opaque type NamedTuple[N <: Tuple, +V <: Tuple] >: V <: AnyNamedTuple = V
+
+  def apply[N <: Tuple, V <: Tuple](x: V): NamedTuple[N, V] = x
+
+  def unapply[N <: Tuple, V <: Tuple](x: NamedTuple[N, V]): Some[V] = Some(x)
+
+  extension [V <: Tuple](x: V)
+    inline def withNames[N <: Tuple]: NamedTuple[N, V] = x
+
+  export NamedTupleDecomposition.{Names, DropNames}
+
+  extension [N <: Tuple, V <: Tuple](x: NamedTuple[N, V])
+
+    /** The underlying tuple without the names */
+    inline def toTuple: V = x
+
+    /** The number of elements in this tuple */
+    inline def size: Tuple.Size[V] = toTuple.size
+
+    // This intentionally works for empty named tuples as well. I think NnEmptyTuple is a dead end
+    // and should be reverted, justy like NonEmptyList is also appealing at first, but a bad idea
+    // in the end.
+
+    /** The value (without the name) at index `n` of this tuple */
+    inline def apply(n: Int): Tuple.Elem[V, n.type] =
+      inline toTuple match
+        case tup: NonEmptyTuple => tup(n).asInstanceOf[Tuple.Elem[V, n.type]]
+        case tup => tup.productElement(n).asInstanceOf[Tuple.Elem[V, n.type]]
+
+    /** The first element value of this tuple */
+    inline def head: Tuple.Elem[V, 0] = apply(0)
+
+    /** The tuple consisting of all elements of this tuple except the first one */
+    inline def tail: Tuple.Drop[V, 1] = toTuple.drop(1)
+
+    /** The last element value of this tuple */
+    inline def last: Tuple.Last[V] = apply(size - 1).asInstanceOf[Tuple.Last[V]]
+
+    /** The tuple consisting of all elements of this tuple except the last one */
+    inline def init: Tuple.Init[V] = toTuple.take(size - 1).asInstanceOf[Tuple.Init[V]]
+
+    /** The tuple consisting of the first `n` elements of this tuple, or all
+     *  elements if `n` exceeds `size`.
+     */
+    inline def take(n: Int): NamedTuple[Tuple.Take[N, n.type], Tuple.Take[V, n.type]] =
+      toTuple.take(n)
+
+    /** The tuple consisting of all elements of this tuple except the first `n` ones,
+     *  or no elements if `n` exceeds `size`.
+     */
+    inline def drop(n: Int): NamedTuple[Tuple.Drop[N, n.type], Tuple.Drop[V, n.type]] =
+      toTuple.drop(n)
+
+    /** The tuple `(x.take(n), x.drop(n))` */
+    inline def splitAt(n: Int): NamedTuple[Tuple.Split[N, n.type], Tuple.Split[V, n.type]] =
+      toTuple.splitAt(n)
+
+    /** The tuple consisting of all elements of this tuple followed by all elements
+     *  of tuple `that`. The names of the two tuples must be disjoint.
+     */
+    inline def ++ [N2 <: Tuple, V2 <: Tuple](that: NamedTuple[N2, V2])(using Tuple.Disjoint[N, N2] =:= true)
+      : NamedTuple[Tuple.Concat[N, N2], Tuple.Concat[V, V2]]
+      = toTuple ++ that.toTuple
+
+    // inline def :* [L] (x: L): NamedTuple[Append[N, ???], Append[V, L] = ???
+    // inline def *: [H] (x: H): NamedTuple[??? *: N], H *: V] = ???
+
+    /** The named tuple consisting of all element values of this tuple mapped by
+     *  the polymorphic mapping function `f`. The names of elements are preserved.
+     *  If `x = (n1 = v1, ..., ni = vi)` then `x.map(f) = `(n1 = f(v1), ..., ni = f(vi))`.
+     */
+    inline def map[F[_]](f: [t] => t => F[t]): NamedTuple[N, Tuple.Map[V, F]] =
+      toTuple.map(f).asInstanceOf[NamedTuple[N, Tuple.Map[V, F]]]
+
+    /** The named tuple consisting of all elements of this tuple in reverse */
+    inline def reverse: NamedTuple[Tuple.Reverse[N], Tuple.Reverse[V]] =
+      toTuple.reverse
+
+    /** The named tuple consisting of all elements values of this tuple zipped
+     *  with corresponding element values in named tuple `that`.
+     *  If the two tuples have different sizes,
+     *  the extra elements of the larger tuple will be disregarded.
+     *  The names of `x` and `that` at the same index must be the same.
+     *  The result tuple keeps the same names as the operand tuples.
+     */
+    inline def zip[V2 <: Tuple](that: NamedTuple[N, V2]): NamedTuple[N, Tuple.Zip[V, V2]] =
+      toTuple.zip(that.toTuple)
+
+    /** A list consisting of all element values */
+    inline def toList: List[Tuple.Union[V]] = toTuple.toList.asInstanceOf[List[Tuple.Union[V]]]
+
+    /** An array consisting of all element values */
+    inline def toArray: Array[Object] = toTuple.toArray
+
+    /** An immutable array consisting of all element values */
+    inline def toIArray: IArray[Object] = toTuple.toIArray
+
+  end extension
+
+  /** The size of a named tuple, represented as a literal constant subtype of Int */
+  type Size[X <: AnyNamedTuple] = Tuple.Size[DropNames[X]]
+
+  /** The type of the element value at position N in the named tuple X */
+  type Elem[X <: AnyNamedTuple, N <: Int] = Tuple.Elem[DropNames[X], N]
+
+  /** The type of the first element value of a named tuple */
+  type Head[X <: AnyNamedTuple] = Elem[X, 0]
+
+  /** The type of the last element value of a named tuple */
+  type Last[X <: AnyNamedTuple] = Tuple.Last[DropNames[X]]
+
+  /** The type of a named tuple consisting of all elements of named tuple X except the first one */
+  type Tail[X <: AnyNamedTuple] = Drop[X, 1]
+
+  /** The type of the initial part of a named tuple without its last element */
+  type Init[X <: AnyNamedTuple] =
+    NamedTuple[Tuple.Init[Names[X]], Tuple.Init[DropNames[X]]]
+
+  /** The type of the named tuple consisting of the first `N` elements of `X`,
+   *  or all elements if `N` exceeds `Size[X]`.
+   */
+  type Take[X <: AnyNamedTuple, N <: Int] =
+    NamedTuple[Tuple.Take[Names[X], N], Tuple.Take[DropNames[X], N]]
+
+  /** The type of the named tuple consisting of all elements of `X` except the first `N` ones,
+   *  or no elements if `N` exceeds `Size[X]`.
+   */
+  type Drop[X <: AnyNamedTuple, N <: Int] =
+    NamedTuple[Tuple.Drop[Names[X], N], Tuple.Drop[DropNames[X], N]]
+
+  /** The pair type `(Take(X, N), Drop[X, N]). */
+  type Split[X <: AnyNamedTuple, N <: Int] = (Take[X, N], Drop[X, N])
+
+  /** Type of the concatenation of two tuples `X` and `Y` */
+  type Concat[X <: AnyNamedTuple, Y <: AnyNamedTuple] =
+    NamedTuple[Tuple.Concat[Names[X], Names[Y]], Tuple.Concat[DropNames[X], DropNames[Y]]]
+
+  /** The type of the named tuple `X` mapped with the type-level function `F`.
+   *  If `X = (n1 : T1, ..., ni : Ti)` then `Map[X, F] = `(n1 : F[T1], ..., ni : F[Ti])`.
+   */
+  type Map[X <: AnyNamedTuple, F[_ <: Tuple.Union[DropNames[X]]]] =
+    NamedTuple[Names[X], Tuple.Map[DropNames[X], F]]
+
+  /** A named tuple with the elements of tuple `X` in reversed order */
+  type Reverse[X <: AnyNamedTuple] =
+    NamedTuple[Tuple.Reverse[Names[X]], Tuple.Reverse[DropNames[X]]]
+
+  /** The type of the named tuple consisting of all element values of
+   *  named tuple `X` zipped with corresponding element values of
+   *  named tuple `Y`. If the two tuples have different sizes,
+   *  the extra elements of the larger tuple will be disregarded.
+   *  The names of `X` and `Y` at the same index must be the same.
+   *  The result tuple keeps the same names as the operand tuples.
+   *  For example, if
+   *  ```
+   *     X = (n1 : S1, ..., ni : Si)
+   *     Y = (n1 : T1, ..., nj : Tj)  where j >= i
+   *  ```
+   *  then
+   *  ```
+   *     Zip[X, Y] = (n1 : (S1, T1), ..., ni: (Si, Ti))
+   *  ```
+   *  @syntax markdown
+   */
+  type Zip[X <: AnyNamedTuple, Y <: AnyNamedTuple] =
+    Tuple.Conforms[Names[X], Names[Y]] match
+      case true =>
+        NamedTuple[Names[X], Tuple.Zip[DropNames[X], DropNames[Y]]]
+
+  type From[T] <: AnyNamedTuple
+
+end NamedTuple
+
+/** Separate from NamedTuple object so that we can match on the opaque type NamedTuple. */
+@experimental
+object NamedTupleDecomposition:
+  import NamedTuple.*
+
+  /** The names of a named tuple, represented as a tuple of literal string values. */
+  type Names[X <: AnyNamedTuple] <: Tuple = X match
+    case NamedTuple[n, _] => n
+
+  /** The value types of a named tuple represented as a regular tuple. */
+  type DropNames[NT <: AnyNamedTuple] <: Tuple = NT match
+    case NamedTuple[_, x] => x
+```