[release-branch.go1.24] go1.24.0

Change-Id: I98457f219e75fb99233804d15c8b9577ee3d4a24
Reviewed-on: https://go-review.googlesource.com/c/go/+/648555
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Bypass: Dmitri Shuralyov <dmitshur@golang.org>
Auto-Submit: Gopher Robot <gobot@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>

VERSION

@@ -0,0 +1,2 @@
+go1.24.0
+time 2025-02-10T23:33:55Z

codereview.cfg

@@ -1 +1,2 @@
-branch: master
+branch: release-branch.go1.24
+parent-branch: master

src/bytes/iter.go

@@ -67,26 +67,26 @@ func splitSeq(s, sep []byte, sepSave int) iter.Seq[[]byte] {
 	}
 }
 
-// SplitSeq returns an iterator over all substrings of s separated by sep.
+// SplitSeq returns an iterator over all subslices of s separated by sep.
-// The iterator yields the same strings that would be returned by [Split](s, sep),
+// The iterator yields the same subslices that would be returned by [Split](s, sep),
-// but without constructing the slice.
+// but without constructing a new slice containing the subslices.
 // It returns a single-use iterator.
 func SplitSeq(s, sep []byte) iter.Seq[[]byte] {
 	return splitSeq(s, sep, 0)
 }
 
-// SplitAfterSeq returns an iterator over substrings of s split after each instance of sep.
+// SplitAfterSeq returns an iterator over subslices of s split after each instance of sep.
-// The iterator yields the same strings that would be returned by [SplitAfter](s, sep),
+// The iterator yields the same subslices that would be returned by [SplitAfter](s, sep),
-// but without constructing the slice.
+// but without constructing a new slice containing the subslices.
 // It returns a single-use iterator.
 func SplitAfterSeq(s, sep []byte) iter.Seq[[]byte] {
 	return splitSeq(s, sep, len(sep))
 }
 
-// FieldsSeq returns an iterator over substrings of s split around runs of
+// FieldsSeq returns an iterator over subslices of s split around runs of
 // whitespace characters, as defined by [unicode.IsSpace].
-// The iterator yields the same strings that would be returned by [Fields](s),
+// The iterator yields the same subslices that would be returned by [Fields](s),
-// but without constructing the slice.
+// but without constructing a new slice containing the subslices.
 func FieldsSeq(s []byte) iter.Seq[[]byte] {
 	return func(yield func([]byte) bool) {
 		start := -1
@@ -116,10 +116,10 @@ func FieldsSeq(s []byte) iter.Seq[[]byte] {
 	}
 }
 
-// FieldsFuncSeq returns an iterator over substrings of s split around runs of
+// FieldsFuncSeq returns an iterator over subslices of s split around runs of
 // Unicode code points satisfying f(c).
-// The iterator yields the same strings that would be returned by [FieldsFunc](s),
+// The iterator yields the same subslices that would be returned by [FieldsFunc](s),
-// but without constructing the slice.
+// but without constructing a new slice containing the subslices.
 func FieldsFuncSeq(s []byte, f func(rune) bool) iter.Seq[[]byte] {
 	return func(yield func([]byte) bool) {
 		start := -1

src/cmd/go/internal/work/security.go

@@ -227,21 +227,6 @@ var validLinkerFlags = []*lazyregexp.Regexp{
 	re(`\./.*\.(a|o|obj|dll|dylib|so|tbd)`),
 }
 
-var validLinkerFlagsOnDarwin = []*lazyregexp.Regexp{
-	// The GNU linker interprets `@file` as "read command-line options from
-	// file". Thus, we forbid values starting with `@` on linker flags.
-	// However, this causes a problem when targeting Darwin.
-	// `@executable_path`, `@loader_path`, and `@rpath` are special values
-	// used in Mach-O to change the library search path and can be used in
-	// conjunction with the `-install_name` and `-rpath` linker flags.
-	// Since the GNU linker does not support Mach-O, targeting Darwin
-	// implies not using the GNU linker. Therefore, we allow @ in the linker
-	// flags if and only if cfg.Goos == "darwin" || cfg.Goos == "ios".
-	re(`-Wl,-dylib_install_name,@rpath(/[^,]*)?`),
-	re(`-Wl,-install_name,@rpath(/[^,]*)?`),
-	re(`-Wl,-rpath,@(executable_path|loader_path)(/[^,]*)?`),
-}
-
 var validLinkerFlagsWithNextArg = []string{
 	"-arch",
 	"-F",
@@ -264,13 +249,8 @@ func checkCompilerFlags(name, source string, list []string) error {
 }
 
 func checkLinkerFlags(name, source string, list []string) error {
-	validLinkerFlagsForPlatform := validLinkerFlags
-	if cfg.Goos == "darwin" || cfg.Goos == "ios" {
-		validLinkerFlagsForPlatform = append(validLinkerFlags, validLinkerFlagsOnDarwin...)
-	}
-
 	checkOverrides := true
-	return checkFlags(name, source, list, invalidLinkerFlags, validLinkerFlagsForPlatform, validLinkerFlagsWithNextArg, checkOverrides)
+	return checkFlags(name, source, list, invalidLinkerFlags, validLinkerFlags, validLinkerFlagsWithNextArg, checkOverrides)
 }
 
 // checkCompilerFlagsForInternalLink returns an error if 'list'

src/cmd/go/internal/work/security_test.go

@@ -8,8 +8,6 @@ import (
 	"os"
 	"strings"
 	"testing"
-
-	"cmd/go/internal/cfg"
 )
 
 var goodCompilerFlags = [][]string{
@@ -247,8 +245,6 @@ var badLinkerFlags = [][]string{
 	{"-Wl,--hash-style=foo"},
 	{"-x", "--c"},
 	{"-x", "@obj"},
-	{"-Wl,-dylib_install_name,@foo"},
-	{"-Wl,-install_name,@foo"},
 	{"-Wl,-rpath,@foo"},
 	{"-Wl,-R,foo,bar"},
 	{"-Wl,-R,@foo"},
@@ -265,21 +261,6 @@ var badLinkerFlags = [][]string{
 	{"./-Wl,--push-state,-R.c"},
 }
 
-var goodLinkerFlagsOnDarwin = [][]string{
-	{"-Wl,-dylib_install_name,@rpath"},
-	{"-Wl,-dylib_install_name,@rpath/"},
-	{"-Wl,-dylib_install_name,@rpath/foo"},
-	{"-Wl,-install_name,@rpath"},
-	{"-Wl,-install_name,@rpath/"},
-	{"-Wl,-install_name,@rpath/foo"},
-	{"-Wl,-rpath,@executable_path"},
-	{"-Wl,-rpath,@executable_path/"},
-	{"-Wl,-rpath,@executable_path/foo"},
-	{"-Wl,-rpath,@loader_path"},
-	{"-Wl,-rpath,@loader_path/"},
-	{"-Wl,-rpath,@loader_path/foo"},
-}
-
 func TestCheckLinkerFlags(t *testing.T) {
 	for _, f := range goodLinkerFlags {
 		if err := checkLinkerFlags("test", "test", f); err != nil {
@@ -291,31 +272,6 @@ func TestCheckLinkerFlags(t *testing.T) {
 			t.Errorf("missing error for %q", f)
 		}
 	}
-
-	goos := cfg.Goos
-
-	cfg.Goos = "darwin"
-	for _, f := range goodLinkerFlagsOnDarwin {
-		if err := checkLinkerFlags("test", "test", f); err != nil {
-			t.Errorf("unexpected error for %q: %v", f, err)
-		}
-	}
-
-	cfg.Goos = "ios"
-	for _, f := range goodLinkerFlagsOnDarwin {
-		if err := checkLinkerFlags("test", "test", f); err != nil {
-			t.Errorf("unexpected error for %q: %v", f, err)
-		}
-	}
-
-	cfg.Goos = "linux"
-	for _, f := range goodLinkerFlagsOnDarwin {
-		if err := checkLinkerFlags("test", "test", f); err == nil {
-			t.Errorf("missing error for %q", f)
-		}
-	}
-
-	cfg.Goos = goos
 }
 
 func TestCheckFlagAllowDisallow(t *testing.T) {

src/os/root.go

@@ -60,7 +60,7 @@ func OpenInRoot(dir, name string) (*File, error) {
 //   - When GOOS=plan9 or GOOS=js, Root does not track directories across renames.
 //     On these platforms, a Root references a directory name, not a file descriptor.
 type Root struct {
-	root root
+	root *root
 }
 
 const (

src/os/root_noopenat.go

@@ -49,7 +49,7 @@ func newRoot(name string) (*Root, error) {
 	if !fi.IsDir() {
 		return nil, errors.New("not a directory")
 	}
-	return &Root{root{name: name}}, nil
+	return &Root{&root{name: name}}, nil
 }
 
 func (r *root) Close() error {

src/os/root_unix.go

@@ -48,11 +48,11 @@ func newRoot(fd int, name string) (*Root, error) {
 		syscall.CloseOnExec(fd)
 	}
 
-	r := &Root{root{
+	r := &Root{&root{
 		fd:   fd,
 		name: name,
 	}}
-	runtime.SetFinalizer(&r.root, (*root).Close)
+	runtime.SetFinalizer(r.root, (*root).Close)
 	return r, nil
 }
 

src/os/root_windows.go

@@ -105,11 +105,11 @@ func newRoot(fd syscall.Handle, name string) (*Root, error) {
 		return nil, &PathError{Op: "open", Path: name, Err: errors.New("not a directory")}
 	}
 
-	r := &Root{root{
+	r := &Root{&root{
 		fd:   fd,
 		name: name,
 	}}
-	runtime.SetFinalizer(&r.root, (*root).Close)
+	runtime.SetFinalizer(r.root, (*root).Close)
 	return r, nil
 }
 

src/weak/pointer.go

@@ -56,6 +56,9 @@ import (
 // referenced object. Typically, this batching only happens for tiny
 // (on the order of 16 bytes or less) and pointer-free objects.
 type Pointer[T any] struct {
+	// Mention T in the type definition to prevent conversions
+	// between Pointer types, like we do for sync/atomic.Pointer.
+	_ [0]*T
 	u unsafe.Pointer
 }
 
@@ -69,7 +72,7 @@ func Make[T any](ptr *T) Pointer[T] {
 		u = runtime_registerWeakPointer(unsafe.Pointer(ptr))
 	}
 	runtime.KeepAlive(ptr)
-	return Pointer[T]{u}
+	return Pointer[T]{u: u}
 }
 
 // Value returns the original pointer used to create the weak pointer.

src/weak/pointer_test.go

@@ -6,10 +6,12 @@ package weak_test
 
 import (
 	"context"
+	"internal/goarch"
 	"runtime"
 	"sync"
 	"testing"
 	"time"
+	"unsafe"
 	"weak"
 )
 
@@ -155,6 +157,14 @@ func TestPointerFinalizer(t *testing.T) {
 	}
 }
 
+func TestPointerSize(t *testing.T) {
+	var p weak.Pointer[T]
+	size := unsafe.Sizeof(p)
+	if size != goarch.PtrSize {
+		t.Errorf("weak.Pointer[T] size = %d, want %d", size, goarch.PtrSize)
+	}
+}
+
 // Regression test for issue 69210.
 //
 // Weak-to-strong conversions must shade the new strong pointer, otherwise

test/weak.go

@@ -0,0 +1,24 @@
+// errorcheck
+
+// Copyright 2025 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.
+
+// Test weak pointers.
+
+package p
+
+import (
+	"runtime"
+	"weak"
+)
+
+// Adapted from example in https://github.com/golang/go/issues/67552#issuecomment-2639661220
+func conversion() {
+	p := "hello"
+	a := weak.Make(&p)
+	b := (weak.Pointer[*byte])(a) // ERROR "cannot convert a \(variable of struct type weak\.Pointer\[string\]\) to type weak.Pointer\[\*byte\]"
+	c := b.Value()
+	println(**c)
+	runtime.KeepAlive(p)
+}