strconv: remove dependence on unicode and strings

We need a compact, reasonably efficient IsPrint. That adds about 2K of data,
plus a modest amount of code, but now strconv is a near-leaf package.

R=r, bradfitz, adg, rsc, minux.ma
CC=golang-dev
https://golang.org/cl/5756050

src/pkg/go/build/deps_test.go

@@ -52,7 +52,7 @@ var pkgDeps = map[string][]string{
 	"math/rand":     {"L0", "math"},
 	"path":          {"L0", "unicode/utf8", "strings"},
 	"sort":          {"math"},
-	"strconv":       {"L0", "unicode", "unicode/utf8", "math", "strings"},
+	"strconv":       {"L0", "unicode/utf8", "math"},
 	"strings":       {"L0", "unicode", "unicode/utf8"},
 	"unicode":       {},
 	"unicode/utf16": {},

src/pkg/strconv/isprint.go

@@ -3,7 +3,7 @@
 
 package strconv
 
-// (474+134)*2 + (180+42)*4 = 2104 bytes
+// (474+134+42)*2 + (180)*4 = 2020 bytes
 
 var isPrint16 = []uint16{
 	0x0020, 0x007e,
@@ -383,139 +383,139 @@ var isNotPrint16 = []uint16{
 }
 
 var isPrint32 = []uint32{
-	0x000020, 0x00007e,
+	0x010000, 0x01004d,
-	0x0000a1, 0x000377,
+	0x010050, 0x01005d,
-	0x00037a, 0x00037e,
+	0x010080, 0x0100fa,
-	0x000384, 0x000527,
+	0x010100, 0x010102,
-	0x000531, 0x000556,
+	0x010107, 0x010133,
-	0x000559, 0x00058a,
+	0x010137, 0x01018a,
-	0x000591, 0x0005c7,
+	0x010190, 0x01019b,
-	0x0005d0, 0x0005ea,
+	0x0101d0, 0x0101fd,
-	0x0005f0, 0x0005f4,
+	0x010280, 0x01029c,
-	0x000606, 0x00061b,
+	0x0102a0, 0x0102d0,
-	0x00061e, 0x00070d,
+	0x010300, 0x010323,
-	0x000710, 0x00074a,
+	0x010330, 0x01034a,
-	0x00074d, 0x0007b1,
+	0x010380, 0x0103c3,
-	0x0007c0, 0x0007fa,
+	0x0103c8, 0x0103d5,
-	0x000800, 0x00082d,
+	0x010400, 0x01049d,
-	0x000830, 0x00085b,
+	0x0104a0, 0x0104a9,
-	0x00085e, 0x00085e,
+	0x010800, 0x010805,
-	0x000900, 0x00098c,
+	0x010808, 0x010838,
-	0x00098f, 0x000990,
+	0x01083c, 0x01083c,
-	0x000993, 0x0009b2,
+	0x01083f, 0x01085f,
-	0x0009b6, 0x0009b9,
+	0x010900, 0x01091b,
-	0x0009bc, 0x0009c4,
+	0x01091f, 0x010939,
-	0x0009c7, 0x0009c8,
+	0x01093f, 0x01093f,
-	0x0009cb, 0x0009ce,
+	0x010a00, 0x010a06,
-	0x0009d7, 0x0009d7,
+	0x010a0c, 0x010a33,
-	0x0009dc, 0x0009e3,
+	0x010a38, 0x010a3a,
-	0x0009e6, 0x0009fb,
+	0x010a3f, 0x010a47,
-	0x000a01, 0x000a0a,
+	0x010a50, 0x010a58,
-	0x000a0f, 0x000a10,
+	0x010a60, 0x010a7f,
-	0x000a13, 0x000a39,
+	0x010b00, 0x010b35,
-	0x000a3c, 0x000a42,
+	0x010b39, 0x010b55,
-	0x000a47, 0x000a48,
+	0x010b58, 0x010b72,
-	0x000a4b, 0x000a4d,
+	0x010b78, 0x010b7f,
-	0x000a51, 0x000a51,
+	0x010c00, 0x010c48,
-	0x000a59, 0x000a5e,
+	0x010e60, 0x010e7e,
-	0x000a66, 0x000a75,
+	0x011000, 0x01104d,
-	0x000a81, 0x000ab9,
+	0x011052, 0x01106f,
-	0x000abc, 0x000acd,
+	0x011080, 0x0110c1,
-	0x000ad0, 0x000ad0,
+	0x012000, 0x01236e,
-	0x000ae0, 0x000ae3,
+	0x012400, 0x012462,
-	0x000ae6, 0x000af1,
+	0x012470, 0x012473,
-	0x000b01, 0x000b0c,
+	0x013000, 0x01342e,
-	0x000b0f, 0x000b10,
+	0x016800, 0x016a38,
-	0x000b13, 0x000b39,
+	0x01b000, 0x01b001,
-	0x000b3c, 0x000b44,
+	0x01d000, 0x01d0f5,
-	0x000b47, 0x000b48,
+	0x01d100, 0x01d126,
-	0x000b4b, 0x000b4d,
+	0x01d129, 0x01d172,
-	0x000b56, 0x000b57,
+	0x01d17b, 0x01d1dd,
-	0x000b5c, 0x000b63,
+	0x01d200, 0x01d245,
-	0x000b66, 0x000b77,
+	0x01d300, 0x01d356,
-	0x000b82, 0x000b8a,
+	0x01d360, 0x01d371,
-	0x000b8e, 0x000b95,
+	0x01d400, 0x01d49f,
-	0x000b99, 0x000b9f,
+	0x01d4a2, 0x01d4a2,
-	0x000ba3, 0x000ba4,
+	0x01d4a5, 0x01d4a6,
-	0x000ba8, 0x000baa,
+	0x01d4a9, 0x01d50a,
-	0x000bae, 0x000bb9,
+	0x01d50d, 0x01d546,
-	0x000bbe, 0x000bc2,
+	0x01d54a, 0x01d6a5,
-	0x000bc6, 0x000bcd,
+	0x01d6a8, 0x01d7cb,
-	0x000bd0, 0x000bd0,
+	0x01d7ce, 0x01d7ff,
-	0x000bd7, 0x000bd7,
+	0x01f000, 0x01f02b,
-	0x000be6, 0x000bfa,
+	0x01f030, 0x01f093,
-	0x000c01, 0x000c39,
+	0x01f0a0, 0x01f0ae,
-	0x000c3d, 0x000c4d,
+	0x01f0b1, 0x01f0be,
-	0x000c55, 0x000c59,
+	0x01f0c1, 0x01f0df,
-	0x000c60, 0x000c63,
+	0x01f100, 0x01f10a,
-	0x000c66, 0x000c6f,
+	0x01f110, 0x01f169,
-	0x000c78, 0x000c7f,
+	0x01f170, 0x01f19a,
-	0x000c82, 0x000cb9,
+	0x01f1e6, 0x01f202,
-	0x000cbc, 0x000ccd,
+	0x01f210, 0x01f23a,
-	0x000cd5, 0x000cd6,
+	0x01f240, 0x01f248,
-	0x000cde, 0x000ce3,
+	0x01f250, 0x01f251,
-	0x000ce6, 0x000cf2,
+	0x01f300, 0x01f320,
-	0x000d02, 0x000d3a,
+	0x01f330, 0x01f37c,
-	0x000d3d, 0x000d4e,
+	0x01f380, 0x01f393,
-	0x000d57, 0x000d57,
+	0x01f3a0, 0x01f3ca,
-	0x000d60, 0x000d63,
+	0x01f3e0, 0x01f3f0,
-	0x000d66, 0x000d75,
+	0x01f400, 0x01f4fc,
-	0x000d79, 0x000d7f,
+	0x01f500, 0x01f53d,
-	0x000d82, 0x000d96,
+	0x01f550, 0x01f567,
-	0x000d9a, 0x000dbd,
+	0x01f5fb, 0x01f625,
-	0x000dc0, 0x000dc6,
+	0x01f628, 0x01f62d,
-	0x000dca, 0x000dca,
+	0x01f630, 0x01f640,
-	0x000dcf, 0x000ddf,
+	0x01f645, 0x01f64f,
-	0x000df2, 0x000df4,
+	0x01f680, 0x01f6c5,
-	0x000e01, 0x000e3a,
+	0x01f700, 0x01f773,
-	0x000e3f, 0x000e5b,
+	0x020000, 0x02a6d6,
-	0x000e81, 0x000e84,
+	0x02a700, 0x02b734,
-	0x000e87, 0x000e8a,
+	0x02b740, 0x02b81d,
-	0x000e8d, 0x000e8d,
+	0x02f800, 0x02fa1d,
-	0x000e94, 0x000ea7,
+	0x0e0100, 0x0e01ef,
 }
 
-var isNotPrint32 = []uint32{
+var isNotPrint32 = []uint16{ // add 0x10000 to each entry
-	0x1000c,
+	0x000c,
-	0x10027,
+	0x0027,
-	0x1003b,
+	0x003b,
-	0x1003e,
+	0x003e,
-	0x1031f,
+	0x031f,
-	0x1039e,
+	0x039e,
-	0x10809,
+	0x0809,
-	0x10836,
+	0x0836,
-	0x10856,
+	0x0856,
-	0x10a04,
+	0x0a04,
-	0x10a14,
+	0x0a14,
-	0x10a18,
+	0x0a18,
-	0x110bd,
+	0x10bd,
-	0x1d455,
+	0xd455,
-	0x1d49d,
+	0xd49d,
-	0x1d4ad,
+	0xd4ad,
-	0x1d4ba,
+	0xd4ba,
-	0x1d4bc,
+	0xd4bc,
-	0x1d4c4,
+	0xd4c4,
-	0x1d506,
+	0xd506,
-	0x1d515,
+	0xd515,
-	0x1d51d,
+	0xd51d,
-	0x1d53a,
+	0xd53a,
-	0x1d53f,
+	0xd53f,
-	0x1d545,
+	0xd545,
-	0x1d551,
+	0xd551,
-	0x1f0d0,
+	0xf0d0,
-	0x1f12f,
+	0xf12f,
-	0x1f336,
+	0xf336,
-	0x1f3c5,
+	0xf3c5,
-	0x1f43f,
+	0xf43f,
-	0x1f441,
+	0xf441,
-	0x1f4f8,
+	0xf4f8,
-	0x1f600,
+	0xf600,
-	0x1f611,
+	0xf611,
-	0x1f615,
+	0xf615,
-	0x1f617,
+	0xf617,
-	0x1f619,
+	0xf619,
-	0x1f61b,
+	0xf61b,
-	0x1f61f,
+	0xf61f,
-	0x1f62c,
+	0xf62c,
-	0x1f634,
+	0xf634,
 }

src/pkg/strconv/makeisprint.go

@@ -9,6 +9,7 @@ package main
 
 import (
 	"fmt"
+	"os"
 	"unicode"
 )
 
@@ -116,8 +117,8 @@ func main() {
 
 	for i := rune(0); i <= unicode.MaxRune; i++ {
 		if isPrint(i) != unicode.IsPrint(i) {
-			fmt.Printf("%U: isPrint=%v, want %v\n", i, isPrint(i), unicode.IsPrint(i))
+			fmt.Fprintf(os.Stderr, "%U: isPrint=%v, want %v\n", i, isPrint(i), unicode.IsPrint(i))
-			break
+			return
 		}
 	}
 
@@ -125,11 +126,11 @@ func main() {
 	fmt.Printf("//     go run makeisprint.go >x && mv x isprint.go\n\n")
 	fmt.Printf("package strconv\n\n")
 
-	fmt.Printf("// (%d+%d)*2 + (%d+%d)*4 = %d bytes\n\n",
+	fmt.Printf("// (%d+%d+%d)*2 + (%d)*4 = %d bytes\n\n",
-		len(range16), len(except16),
+		len(range16), len(except16), len(except32),
-		len(range32), len(except32),
+		len(range32),
-		(len(range16)+len(except16))*2+
+		(len(range16)+len(except16)+len(except32))*2+
-			(len(range32)+len(except32))*4)
+			(len(range32))*4)
 
 	fmt.Printf("var isPrint16 = []uint16{\n")
 	for i := 0; i < len(range16); i += 2 {
@@ -145,13 +146,17 @@ func main() {
 
 	fmt.Printf("var isPrint32 = []uint32{\n")
 	for i := 0; i < len(range32); i += 2 {
-		fmt.Printf("\t%#06x, %#06x,\n", range16[i], range16[i+1])
+		fmt.Printf("\t%#06x, %#06x,\n", range32[i], range32[i+1])
 	}
 	fmt.Printf("}\n\n")
 
-	fmt.Printf("var isNotPrint32 = []uint32{\n")
+	fmt.Printf("var isNotPrint32 = []uint16{ // add 0x10000 to each entry\n")
 	for _, r := range except32 {
-		fmt.Printf("\t%#04x,\n", r)
+		if r >= 0x20000 {
+			fmt.Fprintf(os.Stderr, "%U too big for isNotPrint32\n", r)
+			return
+		}
+		fmt.Printf("\t%#04x,\n", r-0x10000)
 	}
 	fmt.Printf("}\n")
 }

src/pkg/strconv/quote.go

@@ -5,8 +5,6 @@
 package strconv
 
 import (
-	"strings"
-	"unicode"
 	"unicode/utf8"
 )
 
@@ -34,11 +32,11 @@ func quoteWith(s string, quote byte, ASCIIonly bool) string {
 			continue
 		}
 		if ASCIIonly {
-			if r <= unicode.MaxASCII && unicode.IsPrint(r) {
+			if r < utf8.RuneSelf && IsPrint(r) {
 				buf = append(buf, byte(r))
 				continue
 			}
-		} else if unicode.IsPrint(r) {
+		} else if IsPrint(r) {
 			n := utf8.EncodeRune(runeTmp[:], r)
 			buf = append(buf, runeTmp[:n]...)
 			continue
@@ -64,7 +62,7 @@ func quoteWith(s string, quote byte, ASCIIonly bool) string {
 				buf = append(buf, `\x`...)
 				buf = append(buf, lowerhex[s[0]>>4])
 				buf = append(buf, lowerhex[s[0]&0xF])
-			case r > unicode.MaxRune:
+			case r > utf8.MaxRune:
 				r = 0xFFFD
 				fallthrough
 			case r < 0x10000:
@@ -88,7 +86,7 @@ func quoteWith(s string, quote byte, ASCIIonly bool) string {
 // Quote returns a double-quoted Go string literal representing s.  The
 // returned string uses Go escape sequences (\t, \n, \xFF, \u0100) for
 // control characters and non-printable characters as defined by
-// unicode.IsPrint.
+// IsPrint.
 func Quote(s string) string {
 	return quoteWith(s, '"', false)
 }
@@ -101,8 +99,7 @@ func AppendQuote(dst []byte, s string) []byte {
 
 // QuoteToASCII returns a double-quoted Go string literal representing s.
 // The returned string uses Go escape sequences (\t, \n, \xFF, \u0100) for
-// non-ASCII characters and non-printable characters as defined by
+// non-ASCII characters and non-printable characters as defined by IsPrint.
-// unicode.IsPrint.
 func QuoteToASCII(s string) string {
 	return quoteWith(s, '"', true)
 }
@@ -115,8 +112,7 @@ func AppendQuoteToASCII(dst []byte, s string) []byte {
 
 // QuoteRune returns a single-quoted Go character literal representing the
 // rune.  The returned string uses Go escape sequences (\t, \n, \xFF, \u0100)
-// for control characters and non-printable characters as defined by
+// for control characters and non-printable characters as defined by IsPrint.
-// unicode.IsPrint.
 func QuoteRune(r rune) string {
 	// TODO: avoid the allocation here.
 	return quoteWith(string(r), '\'', false)
@@ -131,7 +127,7 @@ func AppendQuoteRune(dst []byte, r rune) []byte {
 // QuoteRuneToASCII returns a single-quoted Go character literal representing
 // the rune.  The returned string uses Go escape sequences (\t, \n, \xFF,
 // \u0100) for non-ASCII characters and non-printable characters as defined
-// by unicode.IsPrint.
+// by IsPrint.
 func QuoteRuneToASCII(r rune) string {
 	// TODO: avoid the allocation here.
 	return quoteWith(string(r), '\'', true)
@@ -246,7 +242,7 @@ func UnquoteChar(s string, quote byte) (value rune, multibyte bool, tail string,
 			value = v
 			break
 		}
-		if v > unicode.MaxRune {
+		if v > utf8.MaxRune {
 			err = ErrSyntax
 			return
 		}
@@ -305,7 +301,7 @@ func Unquote(s string) (t string, err error) {
 	s = s[1 : n-1]
 
 	if quote == '`' {
-		if strings.Contains(s, "`") {
+		if contains(s, '`') {
 			return "", ErrSyntax
 		}
 		return s, nil
@@ -313,12 +309,12 @@ func Unquote(s string) (t string, err error) {
 	if quote != '"' && quote != '\'' {
 		return "", ErrSyntax
 	}
-	if strings.Index(s, "\n") >= 0 {
+	if contains(s, '\n') {
 		return "", ErrSyntax
 	}
 
 	// Is it trivial?  Avoid allocation.
-	if strings.Index(s, `\`) < 0 && strings.IndexRune(s, rune(quote)) < 0 {
+	if !contains(s, '\\') && !contains(s, quote) {
 		switch quote {
 		case '"':
 			return s, nil
@@ -352,6 +348,16 @@ func Unquote(s string) (t string, err error) {
 	return string(buf), nil
 }
 
+// contains reports whether the string contains the byte c.
+func contains(s string, c byte) bool {
+	for i := 0; i < len(s); i++ {
+		if s[i] == c {
+			return true
+		}
+	}
+	return false
+}
+
 // bsearch16 returns the smallest i such that a[i] >= x.
 // If there is no such i, bsearch16 returns len(a).
 func bsearch16(a []uint16, x uint16) int {
@@ -382,7 +388,29 @@ func bsearch32(a []uint32, x uint32) int {
 	return i
 }
 
-func isPrint(r rune) bool {
+// TODO: IsPrint is a local implementation of unicode.IsPrint, verified by the tests
+// to give the same answer. It allows this package not to depend on unicode,
+// and therefore not pull in all the Unicode tables. If the linker were better
+// at tossing unused tables, we could get rid of this implementation.
+// That would be nice.
+
+// IsPrint reports whether the rune is defined as printable by Go, with
+// the same definition as unicode.IsPrint: letters, numbers, punctuation,
+// symbols and ASCII space.
+func IsPrint(r rune) bool {
+	// Fast check for Latin-1
+	if r <= 0xFF {
+		if 0x20 <= r && r <= 0x7E {
+			// All the ASCII is printable from space through DEL-1.
+			return true
+		}
+		if 0xA1 <= r && r <= 0xFF {
+			// Similarly for ¡ through ÿ...
+			return r != 0xAD // ...except for the bizarre soft hyphen.
+		}
+		return false
+	}
+
 	// Same algorithm, either on uint16 or uint32 value.
 	// First, find first i such that isPrint[i] >= x.
 	// This is the index of either the start or end of a pair that might span x.
@@ -404,6 +432,10 @@ func isPrint(r rune) bool {
 	if i >= len(isPrint) || rr < isPrint[i&^1] || isPrint[i|1] < rr {
 		return false
 	}
-	j := bsearch32(isNotPrint, rr)
+	if r >= 0x20000 {
-	return j >= len(isNotPrint) || isNotPrint[j] != rr
+		return true
+	}
+	r -= 0x10000
+	j := bsearch16(isNotPrint, uint16(r))
+	return j >= len(isNotPrint) || isNotPrint[j] != uint16(r)
 }

src/pkg/strconv/quote_test.go

@@ -7,8 +7,23 @@ package strconv_test
 import (
 	. "strconv"
 	"testing"
+	"unicode"
 )
 
+// Verify that our isPrint agrees with unicode.IsPrint
+func TestIsPrint(t *testing.T) {
+	n := 0
+	for r := rune(0); r <= unicode.MaxRune; r++ {
+		if IsPrint(r) != unicode.IsPrint(r) {
+			t.Errorf("IsPrint(%U)=%t incorrect", r, IsPrint(r))
+			n++
+			if n > 10 {
+				return
+			}
+		}
+	}
+}
+
 type quoteTest struct {
 	in    string
 	out   string