cmd/go: encode backslash and newline in response files

Fixes #42295

Change-Id: Ie324bc99a74c1d864c6c2da2e7b929b338c2e033
Reviewed-on: https://go-review.googlesource.com/c/go/+/272870
Trust: Jeremy Faller <jeremy@golang.org>
Run-TryBot: Jeremy Faller <jeremy@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Bryan C. Mills <bcmills@google.com>

5 files changed, 206 insertions, 6 deletions

diff --git a/src/cmd/go/go_test.go b/src/cmd/go/go_test.go
index 19764bfc60..c472620db2 100644
--- a/src/cmd/go/go_test.go
+++ b/src/cmd/go/go_test.go
@@ -31,6 +31,7 @@ import (
 	"cmd/go/internal/cache"
 	"cmd/go/internal/cfg"
 	"cmd/go/internal/robustio"
+	"cmd/go/internal/work"
 	"cmd/internal/sys"
 )
 
@@ -1365,6 +1366,30 @@ func TestLdflagsArgumentsWithSpacesIssue3941(t *testing.T) {
 	tg.grepStderr("^hello world", `ldflags -X "main.extern=hello world"' failed`)
 }
 
+func TestLdFlagsLongArgumentsIssue42295(t *testing.T) {
+	// Test the extremely long command line arguments that contain '\n' characters
+	// get encoded and passed correctly.
+	skipIfGccgo(t, "gccgo does not support -ldflags -X")
+	tooSlow(t)
+	tg := testgo(t)
+	defer tg.cleanup()
+	tg.parallel()
+	tg.tempFile("main.go", `package main
+		var extern string
+		func main() {
+			print(extern)
+		}`)
+	testStr := "test test test test test \n\\ "
+	var buf bytes.Buffer
+	for buf.Len() < work.ArgLengthForResponseFile+1 {
+		buf.WriteString(testStr)
+	}
+	tg.run("run", "-ldflags", fmt.Sprintf(`-X "main.extern=%s"`, buf.String()), tg.path("main.go"))
+	if tg.stderr.String() != buf.String() {
+		t.Errorf("strings differ")
+	}
+}
+
 func TestGoTestDashCDashOControlsBinaryLocation(t *testing.T) {
 	skipIfGccgo(t, "gccgo has no standard packages")
 	tooSlow(t)
diff --git a/src/cmd/go/internal/work/exec.go b/src/cmd/go/internal/work/exec.go
index 336751df27..feb2299d40 100644
--- a/src/cmd/go/internal/work/exec.go
+++ b/src/cmd/go/internal/work/exec.go
@@ -3236,7 +3236,7 @@ func passLongArgsInResponseFiles(cmd *exec.Cmd) (cleanup func()) {
 	cleanup = func() { os.Remove(tf.Name()) }
 	var buf bytes.Buffer
 	for _, arg := range cmd.Args[1:] {
-		fmt.Fprintf(&buf, "%s\n", arg)
+		fmt.Fprintf(&buf, "%s\n", encodeArg(arg))
 	}
 	if _, err := tf.Write(buf.Bytes()); err != nil {
 		tf.Close()
@@ -3251,6 +3251,12 @@ func passLongArgsInResponseFiles(cmd *exec.Cmd) (cleanup func()) {
 	return cleanup
 }
 
+// Windows has a limit of 32 KB arguments. To be conservative and not worry
+// about whether that includes spaces or not, just use 30 KB. Darwin's limit is
+// less clear. The OS claims 256KB, but we've seen failures with arglen as
+// small as 50KB.
+const ArgLengthForResponseFile = (30 << 10)
+
 func useResponseFile(path string, argLen int) bool {
 	// Unless the program uses objabi.Flagparse, which understands
 	// response files, don't use response files.
@@ -3262,11 +3268,7 @@ func useResponseFile(path string, argLen int) bool {
 		return false
 	}
 
-	// Windows has a limit of 32 KB arguments. To be conservative and not
-	// worry about whether that includes spaces or not, just use 30 KB.
-	// Darwin's limit is less clear. The OS claims 256KB, but we've seen
-	// failures with arglen as small as 50KB.
-	if argLen > (30 << 10) {
+	if argLen > ArgLengthForResponseFile {
 		return true
 	}
 
@@ -3279,3 +3281,25 @@ func useResponseFile(path string, argLen int) bool {
 
 	return false
 }
+
+// encodeArg encodes an argument for response file writing.
+func encodeArg(arg string) string {
+	// If there aren't any characters we need to reencode, fastpath out.
+	if !strings.ContainsAny(arg, "\\\n") {
+		return arg
+	}
+	var b strings.Builder
+	for _, r := range arg {
+		switch r {
+		case '\\':
+			b.WriteByte('\\')
+			b.WriteByte('\\')
+		case '\n':
+			b.WriteByte('\\')
+			b.WriteByte('n')
+		default:
+			b.WriteRune(r)
+		}
+	}
+	return b.String()
+}
diff --git a/src/cmd/go/internal/work/exec_test.go b/src/cmd/go/internal/work/exec_test.go
new file mode 100644
index 0000000000..4eb762cb28
--- /dev/null
+++ b/src/cmd/go/internal/work/exec_test.go
@@ -0,0 +1,86 @@
+// Copyright 2011 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 work
+
+import (
+	"bytes"
+	"cmd/internal/objabi"
+	"fmt"
+	"math/rand"
+	"testing"
+	"time"
+	"unicode/utf8"
+)
+
+func TestEncodeArgs(t *testing.T) {
+	t.Parallel()
+	tests := []struct {
+		arg, want string
+	}{
+		{"", ""},
+		{"hello", "hello"},
+		{"hello\n", "hello\\n"},
+		{"hello\\", "hello\\\\"},
+		{"hello\nthere", "hello\\nthere"},
+		{"\\\n", "\\\\\\n"},
+	}
+	for _, test := range tests {
+		if got := encodeArg(test.arg); got != test.want {
+			t.Errorf("encodeArg(%q) = %q, want %q", test.arg, got, test.want)
+		}
+	}
+}
+
+func TestEncodeDecode(t *testing.T) {
+	t.Parallel()
+	tests := []string{
+		"",
+		"hello",
+		"hello\\there",
+		"hello\nthere",
+		"hello 中国",
+		"hello \n中\\国",
+	}
+	for _, arg := range tests {
+		if got := objabi.DecodeArg(encodeArg(arg)); got != arg {
+			t.Errorf("objabi.DecodeArg(encodeArg(%q)) = %q", arg, got)
+		}
+	}
+}
+
+func TestEncodeDecodeFuzz(t *testing.T) {
+	if testing.Short() {
+		t.Skip("fuzz test is slow")
+	}
+	t.Parallel()
+
+	nRunes := ArgLengthForResponseFile + 100
+	rBuffer := make([]rune, nRunes)
+	buf := bytes.NewBuffer([]byte(string(rBuffer)))
+
+	seed := time.Now().UnixNano()
+	t.Logf("rand seed: %v", seed)
+	rng := rand.New(rand.NewSource(seed))
+
+	for i := 0; i < 50; i++ {
+		// Generate a random string of runes.
+		buf.Reset()
+		for buf.Len() < ArgLengthForResponseFile+1 {
+			var r rune
+			for {
+				r = rune(rng.Intn(utf8.MaxRune + 1))
+				if utf8.ValidRune(r) {
+					break
+				}
+			}
+			fmt.Fprintf(buf, "%c", r)
+		}
+		arg := buf.String()
+
+		if got := objabi.DecodeArg(encodeArg(arg)); got != arg {
+			t.Errorf("[%d] objabi.DecodeArg(encodeArg(%q)) = %q [seed: %v]", i, arg, got, seed)
+		}
+	}
+}
diff --git a/src/cmd/internal/objabi/flag.go b/src/cmd/internal/objabi/flag.go
index 79ad2ccf74..3fd73f3c57 100644
--- a/src/cmd/internal/objabi/flag.go
+++ b/src/cmd/internal/objabi/flag.go
@@ -5,6 +5,7 @@
 package objabi
 
 import (
+	"bytes"
 	"flag"
 	"fmt"
 	"io"
@@ -59,6 +60,9 @@ func expandArgs(in []string) (out []string) {
 				log.Fatal(err)
 			}
 			args := strings.Split(strings.TrimSpace(strings.Replace(string(slurp), "\r", "", -1)), "\n")
+			for i, arg := range args {
+				args[i] = DecodeArg(arg)
+			}
 			out = append(out, expandArgs(args)...)
 		} else if out != nil {
 			out = append(out, s)
@@ -160,3 +164,38 @@ func (f fn1) Set(s string) error {
 }
 
 func (f fn1) String() string { return "" }
+
+// DecodeArg decodes an argument.
+//
+// This function is public for testing with the parallel encoder.
+func DecodeArg(arg string) string {
+	// If no encoding, fastpath out.
+	if !strings.ContainsAny(arg, "\\\n") {
+		return arg
+	}
+
+	// We can't use strings.Builder as this must work at bootstrap.
+	var b bytes.Buffer
+	var wasBS bool
+	for _, r := range arg {
+		if wasBS {
+			switch r {
+			case '\\':
+				b.WriteByte('\\')
+			case 'n':
+				b.WriteByte('\n')
+			default:
+				// This shouldn't happen. The only backslashes that reach here
+				// should encode '\n' and '\\' exclusively.
+				panic("badly formatted input")
+			}
+		} else if r == '\\' {
+			wasBS = true
+			continue
+		} else {
+			b.WriteRune(r)
+		}
+		wasBS = false
+	}
+	return b.String()
+}
diff --git a/src/cmd/internal/objabi/flag_test.go b/src/cmd/internal/objabi/flag_test.go
new file mode 100644
index 0000000000..935b9c2193
--- /dev/null
+++ b/src/cmd/internal/objabi/flag_test.go
@@ -0,0 +1,26 @@
+// Copyright 2020 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 objabi
+
+import "testing"
+
+func TestDecodeArg(t *testing.T) {
+	t.Parallel()
+	tests := []struct {
+		arg, want string
+	}{
+		{"", ""},
+		{"hello", "hello"},
+		{"hello\\n", "hello\n"},
+		{"hello\\nthere", "hello\nthere"},
+		{"hello\\\\there", "hello\\there"},
+		{"\\\\\\n", "\\\n"},
+	}
+	for _, test := range tests {
+		if got := DecodeArg(test.arg); got != test.want {
+			t.Errorf("decodoeArg(%q) = %q, want %q", test.arg, got, test.want)
+		}
+	}
+}