[dev.typeparams] Don't check typecheck(3) on transform, so no need to export/import it

We have a value typecheck(3) that indicates that a node in a generic
function still needs transformation (via the functions in transform.go).
But it is not very desirable to export/import the value of typecheck(3).
So, I changed the stenciling code to just try to transform all relevant
node types during node copy. Almost all tranform functions were already
idempotent. I only had to add an extra if check before calling
transformAssign() in the OAS case. We still use the typecheck(3) in
noder to determine when higher-nodes have to delay transformation
because one or more of their args are delaying transformation.

Added new test mapsimp.go that required these tranformations after import.

As an additional change, export/import of OINDEX requires exporting the
type using w.exoticType() rather than w.typ(), in order to handle
generic functions. Since generic functions can have pre-transform
operations, the index operation can have a tuple type (multiple return
from a map lookup).

Added printing of imported function bodies in -W=3 debug mode.

Change-Id: I220e2428dc5f2741e91db146f075eb5b6045f451
Reviewed-on: https://go-review.googlesource.com/c/go/+/322191
Trust: Dan Scales <danscales@google.com>
Run-TryBot: Dan Scales <danscales@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>

src/cmd/compile/internal/noder/stencil.go

@@ -374,9 +374,6 @@ func (subst *subster) node(n ir.Node) ir.Node {
 		}
 		ir.EditChildren(m, edit)
 
-		if x.Typecheck() == 3 {
-			// These are nodes whose transforms were delayed until
-			// their instantiated type was known.
 		m.SetTypecheck(1)
 		if typecheck.IsCmp(x.Op()) {
 			transformCompare(m.(*ir.BinaryExpr))
@@ -397,8 +394,12 @@ func (subst *subster) node(n ir.Node) ir.Node {
 
 			case ir.OAS:
 				as := m.(*ir.AssignStmt)
+				if as.Y != nil {
+					// transformAssign doesn't handle the case
+					// of zeroing assignment of a dcl (rhs[0] is nil).
 					lhs, rhs := []ir.Node{as.X}, []ir.Node{as.Y}
 					transformAssign(as, lhs, rhs)
+				}
 
 			case ir.OASOP:
 				as := m.(*ir.AssignOpStmt)
@@ -410,9 +411,6 @@ func (subst *subster) node(n ir.Node) ir.Node {
 			case ir.OSEND:
 				transformSend(m.(*ir.SendStmt))
 
-				default:
-					base.Fatalf("Unexpected node with Typecheck() == 3")
-				}
 			}
 		}
 

src/cmd/compile/internal/typecheck/iexport.go

@@ -1735,7 +1735,7 @@ func (w *exportWriter) expr(n ir.Node) {
 		w.expr(n.X)
 		w.expr(n.Index)
 		if go117ExportTypes {
-			w.typ(n.Type())
+			w.exoticType(n.Type())
 			if n.Op() == ir.OINDEXMAP {
 				w.bool(n.Assigned)
 			}

src/cmd/compile/internal/typecheck/iimport.go

@@ -1025,6 +1025,10 @@ func (r *importReader) funcBody(fn *ir.Func) {
 	fn.Inl.Body = body
 
 	r.curfn = outerfn
+	if base.Flag.W >= 3 {
+		fmt.Printf("Imported for %v", fn)
+		ir.DumpList("", fn.Inl.Body)
+	}
 }
 
 func (r *importReader) readNames(fn *ir.Func) []*ir.Name {
@@ -1349,7 +1353,7 @@ func (r *importReader) node() ir.Node {
 		n := ir.NewIndexExpr(r.pos(), r.expr(), r.expr())
 		if go117ExportTypes {
 			n.SetOp(op)
-			n.SetType(r.typ())
+			n.SetType(r.exoticType())
 			if op == ir.OINDEXMAP {
 				n.Assigned = r.bool()
 			}

test/typeparam/mapsimp.dir/a.go

@@ -0,0 +1,108 @@
+// Copyright 2021 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package a
+
+// SliceEqual reports whether two slices are equal: the same length and all
+// elements equal. All floating point NaNs are considered equal.
+func SliceEqual[Elem comparable](s1, s2 []Elem) bool {
+	if len(s1) != len(s2) {
+		return false
+	}
+	for i, v1 := range s1 {
+		v2 := s2[i]
+		if v1 != v2 {
+			isNaN := func(f Elem) bool { return f != f }
+			if !isNaN(v1) || !isNaN(v2) {
+				return false
+			}
+		}
+	}
+	return true
+}
+
+// Keys returns the keys of the map m.
+// The keys will be an indeterminate order.
+func Keys[K comparable, V any](m map[K]V) []K {
+	r := make([]K, 0, len(m))
+	for k := range m {
+		r = append(r, k)
+	}
+	return r
+}
+
+// Values returns the values of the map m.
+// The values will be in an indeterminate order.
+func Values[K comparable, V any](m map[K]V) []V {
+	r := make([]V, 0, len(m))
+	for _, v := range m {
+		r = append(r, v)
+	}
+	return r
+}
+
+// Equal reports whether two maps contain the same key/value pairs.
+// Values are compared using ==.
+func Equal[K, V comparable](m1, m2 map[K]V) bool {
+	if len(m1) != len(m2) {
+		return false
+	}
+	for k, v1 := range m1 {
+		if v2, ok := m2[k]; !ok || v1 != v2 {
+			return false
+		}
+	}
+	return true
+}
+
+// Copy returns a copy of m.
+func Copy[K comparable, V any](m map[K]V) map[K]V {
+	r := make(map[K]V, len(m))
+	for k, v := range m {
+		r[k] = v
+	}
+	return r
+}
+
+// Add adds all key/value pairs in m2 to m1. Keys in m2 that are already
+// present in m1 will be overwritten with the value in m2.
+func Add[K comparable, V any](m1, m2 map[K]V) {
+	for k, v := range m2 {
+		m1[k] = v
+	}
+}
+
+// Sub removes all keys in m2 from m1. Keys in m2 that are not present
+// in m1 are ignored. The values in m2 are ignored.
+func Sub[K comparable, V any](m1, m2 map[K]V) {
+	for k := range m2 {
+		delete(m1, k)
+	}
+}
+
+// Intersect removes all keys from m1 that are not present in m2.
+// Keys in m2 that are not in m1 are ignored. The values in m2 are ignored.
+func Intersect[K comparable, V any](m1, m2 map[K]V) {
+	for k := range m1 {
+		if _, ok := m2[k]; !ok {
+			delete(m1, k)
+		}
+	}
+}
+
+// Filter deletes any key/value pairs from m for which f returns false.
+func Filter[K comparable, V any](m map[K]V, f func(K, V) bool) {
+	for k, v := range m {
+		if !f(k, v) {
+			delete(m, k)
+		}
+	}
+}
+
+// TransformValues applies f to each value in m. The keys remain unchanged.
+func TransformValues[K comparable, V any](m map[K]V, f func(V) V) {
+	for k, v := range m {
+		m[k] = f(v)
+	}
+}

test/typeparam/mapsimp.dir/main.go

@@ -0,0 +1,156 @@
+// Copyright 2021 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package main
+
+import (
+	"a"
+	"fmt"
+	"math"
+	"sort"
+)
+
+var m1 = map[int]int{1: 2, 2: 4, 4: 8, 8: 16}
+var m2 = map[int]string{1: "2", 2: "4", 4: "8", 8: "16"}
+
+func TestKeys() {
+	want := []int{1, 2, 4, 8}
+
+	got1 := a.Keys(m1)
+	sort.Ints(got1)
+	if !a.SliceEqual(got1, want) {
+		panic(fmt.Sprintf("a.Keys(%v) = %v, want %v", m1, got1, want))
+	}
+
+	got2 := a.Keys(m2)
+	sort.Ints(got2)
+	if !a.SliceEqual(got2, want) {
+		panic(fmt.Sprintf("a.Keys(%v) = %v, want %v", m2, got2, want))
+	}
+}
+
+func TestValues() {
+	got1 := a.Values(m1)
+	want1 := []int{2, 4, 8, 16}
+	sort.Ints(got1)
+	if !a.SliceEqual(got1, want1) {
+		panic(fmt.Sprintf("a.Values(%v) = %v, want %v", m1, got1, want1))
+	}
+
+	got2 := a.Values(m2)
+	want2 := []string{"16", "2", "4", "8"}
+	sort.Strings(got2)
+	if !a.SliceEqual(got2, want2) {
+		panic(fmt.Sprintf("a.Values(%v) = %v, want %v", m2, got2, want2))
+	}
+}
+
+func TestEqual() {
+	if !a.Equal(m1, m1) {
+		panic(fmt.Sprintf("a.Equal(%v, %v) = false, want true", m1, m1))
+	}
+	if a.Equal(m1, nil) {
+		panic(fmt.Sprintf("a.Equal(%v, nil) = true, want false", m1))
+	}
+	if a.Equal(nil, m1) {
+		panic(fmt.Sprintf("a.Equal(nil, %v) = true, want false", m1))
+	}
+	if !a.Equal[int, int](nil, nil) {
+		panic("a.Equal(nil, nil) = false, want true")
+	}
+	if ms := map[int]int{1: 2}; a.Equal(m1, ms) {
+		panic(fmt.Sprintf("a.Equal(%v, %v) = true, want false", m1, ms))
+	}
+
+	// Comparing NaN for equality is expected to fail.
+	mf := map[int]float64{1: 0, 2: math.NaN()}
+	if a.Equal(mf, mf) {
+		panic(fmt.Sprintf("a.Equal(%v, %v) = true, want false", mf, mf))
+	}
+}
+
+func TestCopy() {
+	m2 := a.Copy(m1)
+	if !a.Equal(m1, m2) {
+		panic(fmt.Sprintf("a.Copy(%v) = %v, want %v", m1, m2, m1))
+	}
+	m2[16] = 32
+	if a.Equal(m1, m2) {
+		panic(fmt.Sprintf("a.Equal(%v, %v) = true, want false", m1, m2))
+	}
+}
+
+func TestAdd() {
+	mc := a.Copy(m1)
+	a.Add(mc, mc)
+	if !a.Equal(mc, m1) {
+		panic(fmt.Sprintf("a.Add(%v, %v) = %v, want %v", m1, m1, mc, m1))
+	}
+	a.Add(mc, map[int]int{16: 32})
+	want := map[int]int{1: 2, 2: 4, 4: 8, 8: 16, 16: 32}
+	if !a.Equal(mc, want) {
+		panic(fmt.Sprintf("a.Add result = %v, want %v", mc, want))
+	}
+}
+
+func TestSub() {
+	mc := a.Copy(m1)
+	a.Sub(mc, mc)
+	if len(mc) > 0 {
+		panic(fmt.Sprintf("a.Sub(%v, %v) = %v, want empty map", m1, m1, mc))
+	}
+	mc = a.Copy(m1)
+	a.Sub(mc, map[int]int{1: 0})
+	want := map[int]int{2: 4, 4: 8, 8: 16}
+	if !a.Equal(mc, want) {
+		panic(fmt.Sprintf("a.Sub result = %v, want %v", mc, want))
+	}
+}
+
+func TestIntersect() {
+	mc := a.Copy(m1)
+	a.Intersect(mc, mc)
+	if !a.Equal(mc, m1) {
+		panic(fmt.Sprintf("a.Intersect(%v, %v) = %v, want %v", m1, m1, mc, m1))
+	}
+	a.Intersect(mc, map[int]int{1: 0, 2: 0})
+	want := map[int]int{1: 2, 2: 4}
+	if !a.Equal(mc, want) {
+		panic(fmt.Sprintf("a.Intersect result = %v, want %v", mc, want))
+	}
+}
+
+func TestFilter() {
+	mc := a.Copy(m1)
+	a.Filter(mc, func(int, int) bool { return true })
+	if !a.Equal(mc, m1) {
+		panic(fmt.Sprintf("a.Filter(%v, true) = %v, want %v", m1, mc, m1))
+	}
+	a.Filter(mc, func(k, v int) bool { return k < 3 })
+	want := map[int]int{1: 2, 2: 4}
+	if !a.Equal(mc, want) {
+		panic(fmt.Sprintf("a.Filter result = %v, want %v", mc, want))
+	}
+}
+
+func TestTransformValues() {
+	mc := a.Copy(m1)
+	a.TransformValues(mc, func(i int) int { return i / 2 })
+	want := map[int]int{1: 1, 2: 2, 4: 4, 8: 8}
+	if !a.Equal(mc, want) {
+		panic(fmt.Sprintf("a.TransformValues result = %v, want %v", mc, want))
+	}
+}
+
+func main() {
+	TestKeys()
+	TestValues()
+	TestEqual()
+	TestCopy()
+	TestAdd()
+	TestSub()
+	TestIntersect()
+	TestFilter()
+	TestTransformValues()
+}

test/typeparam/mapsimp.go

@@ -0,0 +1,7 @@
+// rundir -G=3
+
+// Copyright 2021 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package ignored