Refactor utf-8 conversion (irc) (#1767)

3 files changed, 681 insertions, 0 deletions

diff --git a/vendor/golang.org/x/text/encoding/unicode/override.go b/vendor/golang.org/x/text/encoding/unicode/override.go
new file mode 100644
index 00000000..35d62fcc
--- /dev/null
+++ b/vendor/golang.org/x/text/encoding/unicode/override.go
@@ -0,0 +1,82 @@
+// Copyright 2015 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 unicode
+
+import (
+	"golang.org/x/text/transform"
+)
+
+// BOMOverride returns a new decoder transformer that is identical to fallback,
+// except that the presence of a Byte Order Mark at the start of the input
+// causes it to switch to the corresponding Unicode decoding. It will only
+// consider BOMs for UTF-8, UTF-16BE, and UTF-16LE.
+//
+// This differs from using ExpectBOM by allowing a BOM to switch to UTF-8, not
+// just UTF-16 variants, and allowing falling back to any encoding scheme.
+//
+// This technique is recommended by the W3C for use in HTML 5: "For
+// compatibility with deployed content, the byte order mark (also known as BOM)
+// is considered more authoritative than anything else."
+// http://www.w3.org/TR/encoding/#specification-hooks
+//
+// Using BOMOverride is mostly intended for use cases where the first characters
+// of a fallback encoding are known to not be a BOM, for example, for valid HTML
+// and most encodings.
+func BOMOverride(fallback transform.Transformer) transform.Transformer {
+	// TODO: possibly allow a variadic argument of unicode encodings to allow
+	// specifying details of which fallbacks are supported as well as
+	// specifying the details of the implementations. This would also allow for
+	// support for UTF-32, which should not be supported by default.
+	return &bomOverride{fallback: fallback}
+}
+
+type bomOverride struct {
+	fallback transform.Transformer
+	current  transform.Transformer
+}
+
+func (d *bomOverride) Reset() {
+	d.current = nil
+	d.fallback.Reset()
+}
+
+var (
+	// TODO: we could use decode functions here, instead of allocating a new
+	// decoder on every NewDecoder as IgnoreBOM decoders can be stateless.
+	utf16le = UTF16(LittleEndian, IgnoreBOM)
+	utf16be = UTF16(BigEndian, IgnoreBOM)
+)
+
+const utf8BOM = "\ufeff"
+
+func (d *bomOverride) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	if d.current != nil {
+		return d.current.Transform(dst, src, atEOF)
+	}
+	if len(src) < 3 && !atEOF {
+		return 0, 0, transform.ErrShortSrc
+	}
+	d.current = d.fallback
+	bomSize := 0
+	if len(src) >= 2 {
+		if src[0] == 0xFF && src[1] == 0xFE {
+			d.current = utf16le.NewDecoder()
+			bomSize = 2
+		} else if src[0] == 0xFE && src[1] == 0xFF {
+			d.current = utf16be.NewDecoder()
+			bomSize = 2
+		} else if len(src) >= 3 &&
+			src[0] == utf8BOM[0] &&
+			src[1] == utf8BOM[1] &&
+			src[2] == utf8BOM[2] {
+			d.current = transform.Nop
+			bomSize = 3
+		}
+	}
+	if bomSize < len(src) {
+		nDst, nSrc, err = d.current.Transform(dst, src[bomSize:], atEOF)
+	}
+	return nDst, nSrc + bomSize, err
+}
diff --git a/vendor/golang.org/x/text/encoding/unicode/unicode.go b/vendor/golang.org/x/text/encoding/unicode/unicode.go
new file mode 100644
index 00000000..dd99ad14
--- /dev/null
+++ b/vendor/golang.org/x/text/encoding/unicode/unicode.go
@@ -0,0 +1,512 @@
+// Copyright 2013 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 unicode provides Unicode encodings such as UTF-16.
+package unicode // import "golang.org/x/text/encoding/unicode"
+
+import (
+	"bytes"
+	"errors"
+	"unicode/utf16"
+	"unicode/utf8"
+
+	"golang.org/x/text/encoding"
+	"golang.org/x/text/encoding/internal"
+	"golang.org/x/text/encoding/internal/identifier"
+	"golang.org/x/text/internal/utf8internal"
+	"golang.org/x/text/runes"
+	"golang.org/x/text/transform"
+)
+
+// TODO: I think the Transformers really should return errors on unmatched
+// surrogate pairs and odd numbers of bytes. This is not required by RFC 2781,
+// which leaves it open, but is suggested by WhatWG. It will allow for all error
+// modes as defined by WhatWG: fatal, HTML and Replacement. This would require
+// the introduction of some kind of error type for conveying the erroneous code
+// point.
+
+// UTF8 is the UTF-8 encoding. It neither removes nor adds byte order marks.
+var UTF8 encoding.Encoding = utf8enc
+
+// UTF8BOM is an UTF-8 encoding where the decoder strips a leading byte order
+// mark while the encoder adds one.
+//
+// Some editors add a byte order mark as a signature to UTF-8 files. Although
+// the byte order mark is not useful for detecting byte order in UTF-8, it is
+// sometimes used as a convention to mark UTF-8-encoded files. This relies on
+// the observation that the UTF-8 byte order mark is either an illegal or at
+// least very unlikely sequence in any other character encoding.
+var UTF8BOM encoding.Encoding = utf8bomEncoding{}
+
+type utf8bomEncoding struct{}
+
+func (utf8bomEncoding) String() string {
+	return "UTF-8-BOM"
+}
+
+func (utf8bomEncoding) ID() (identifier.MIB, string) {
+	return identifier.Unofficial, "x-utf8bom"
+}
+
+func (utf8bomEncoding) NewEncoder() *encoding.Encoder {
+	return &encoding.Encoder{
+		Transformer: &utf8bomEncoder{t: runes.ReplaceIllFormed()},
+	}
+}
+
+func (utf8bomEncoding) NewDecoder() *encoding.Decoder {
+	return &encoding.Decoder{Transformer: &utf8bomDecoder{}}
+}
+
+var utf8enc = &internal.Encoding{
+	&internal.SimpleEncoding{utf8Decoder{}, runes.ReplaceIllFormed()},
+	"UTF-8",
+	identifier.UTF8,
+}
+
+type utf8bomDecoder struct {
+	checked bool
+}
+
+func (t *utf8bomDecoder) Reset() {
+	t.checked = false
+}
+
+func (t *utf8bomDecoder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	if !t.checked {
+		if !atEOF && len(src) < len(utf8BOM) {
+			if len(src) == 0 {
+				return 0, 0, nil
+			}
+			return 0, 0, transform.ErrShortSrc
+		}
+		if bytes.HasPrefix(src, []byte(utf8BOM)) {
+			nSrc += len(utf8BOM)
+			src = src[len(utf8BOM):]
+		}
+		t.checked = true
+	}
+	nDst, n, err := utf8Decoder.Transform(utf8Decoder{}, dst[nDst:], src, atEOF)
+	nSrc += n
+	return nDst, nSrc, err
+}
+
+type utf8bomEncoder struct {
+	written bool
+	t       transform.Transformer
+}
+
+func (t *utf8bomEncoder) Reset() {
+	t.written = false
+	t.t.Reset()
+}
+
+func (t *utf8bomEncoder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	if !t.written {
+		if len(dst) < len(utf8BOM) {
+			return nDst, 0, transform.ErrShortDst
+		}
+		nDst = copy(dst, utf8BOM)
+		t.written = true
+	}
+	n, nSrc, err := utf8Decoder.Transform(utf8Decoder{}, dst[nDst:], src, atEOF)
+	nDst += n
+	return nDst, nSrc, err
+}
+
+type utf8Decoder struct{ transform.NopResetter }
+
+func (utf8Decoder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	var pSrc int // point from which to start copy in src
+	var accept utf8internal.AcceptRange
+
+	// The decoder can only make the input larger, not smaller.
+	n := len(src)
+	if len(dst) < n {
+		err = transform.ErrShortDst
+		n = len(dst)
+		atEOF = false
+	}
+	for nSrc < n {
+		c := src[nSrc]
+		if c < utf8.RuneSelf {
+			nSrc++
+			continue
+		}
+		first := utf8internal.First[c]
+		size := int(first & utf8internal.SizeMask)
+		if first == utf8internal.FirstInvalid {
+			goto handleInvalid // invalid starter byte
+		}
+		accept = utf8internal.AcceptRanges[first>>utf8internal.AcceptShift]
+		if nSrc+size > n {
+			if !atEOF {
+				// We may stop earlier than necessary here if the short sequence
+				// has invalid bytes. Not checking for this simplifies the code
+				// and may avoid duplicate computations in certain conditions.
+				if err == nil {
+					err = transform.ErrShortSrc
+				}
+				break
+			}
+			// Determine the maximal subpart of an ill-formed subsequence.
+			switch {
+			case nSrc+1 >= n || src[nSrc+1] < accept.Lo || accept.Hi < src[nSrc+1]:
+				size = 1
+			case nSrc+2 >= n || src[nSrc+2] < utf8internal.LoCB || utf8internal.HiCB < src[nSrc+2]:
+				size = 2
+			default:
+				size = 3 // As we are short, the maximum is 3.
+			}
+			goto handleInvalid
+		}
+		if c = src[nSrc+1]; c < accept.Lo || accept.Hi < c {
+			size = 1
+			goto handleInvalid // invalid continuation byte
+		} else if size == 2 {
+		} else if c = src[nSrc+2]; c < utf8internal.LoCB || utf8internal.HiCB < c {
+			size = 2
+			goto handleInvalid // invalid continuation byte
+		} else if size == 3 {
+		} else if c = src[nSrc+3]; c < utf8internal.LoCB || utf8internal.HiCB < c {
+			size = 3
+			goto handleInvalid // invalid continuation byte
+		}
+		nSrc += size
+		continue
+
+	handleInvalid:
+		// Copy the scanned input so far.
+		nDst += copy(dst[nDst:], src[pSrc:nSrc])
+
+		// Append RuneError to the destination.
+		const runeError = "\ufffd"
+		if nDst+len(runeError) > len(dst) {
+			return nDst, nSrc, transform.ErrShortDst
+		}
+		nDst += copy(dst[nDst:], runeError)
+
+		// Skip the maximal subpart of an ill-formed subsequence according to
+		// the W3C standard way instead of the Go way. This Transform is
+		// probably the only place in the text repo where it is warranted.
+		nSrc += size
+		pSrc = nSrc
+
+		// Recompute the maximum source length.
+		if sz := len(dst) - nDst; sz < len(src)-nSrc {
+			err = transform.ErrShortDst
+			n = nSrc + sz
+			atEOF = false
+		}
+	}
+	return nDst + copy(dst[nDst:], src[pSrc:nSrc]), nSrc, err
+}
+
+// UTF16 returns a UTF-16 Encoding for the given default endianness and byte
+// order mark (BOM) policy.
+//
+// When decoding from UTF-16 to UTF-8, if the BOMPolicy is IgnoreBOM then
+// neither BOMs U+FEFF nor noncharacters U+FFFE in the input stream will affect
+// the endianness used for decoding, and will instead be output as their
+// standard UTF-8 encodings: "\xef\xbb\xbf" and "\xef\xbf\xbe". If the BOMPolicy
+// is UseBOM or ExpectBOM a staring BOM is not written to the UTF-8 output.
+// Instead, it overrides the default endianness e for the remainder of the
+// transformation. Any subsequent BOMs U+FEFF or noncharacters U+FFFE will not
+// affect the endianness used, and will instead be output as their standard
+// UTF-8 encodings. For UseBOM, if there is no starting BOM, it will proceed
+// with the default Endianness. For ExpectBOM, in that case, the transformation
+// will return early with an ErrMissingBOM error.
+//
+// When encoding from UTF-8 to UTF-16, a BOM will be inserted at the start of
+// the output if the BOMPolicy is UseBOM or ExpectBOM. Otherwise, a BOM will not
+// be inserted. The UTF-8 input does not need to contain a BOM.
+//
+// There is no concept of a 'native' endianness. If the UTF-16 data is produced
+// and consumed in a greater context that implies a certain endianness, use
+// IgnoreBOM. Otherwise, use ExpectBOM and always produce and consume a BOM.
+//
+// In the language of https://www.unicode.org/faq/utf_bom.html#bom10, IgnoreBOM
+// corresponds to "Where the precise type of the data stream is known... the
+// BOM should not be used" and ExpectBOM corresponds to "A particular
+// protocol... may require use of the BOM".
+func UTF16(e Endianness, b BOMPolicy) encoding.Encoding {
+	return utf16Encoding{config{e, b}, mibValue[e][b&bomMask]}
+}
+
+// mibValue maps Endianness and BOMPolicy settings to MIB constants. Note that
+// some configurations map to the same MIB identifier. RFC 2781 has requirements
+// and recommendations. Some of the "configurations" are merely recommendations,
+// so multiple configurations could match.
+var mibValue = map[Endianness][numBOMValues]identifier.MIB{
+	BigEndian: [numBOMValues]identifier.MIB{
+		IgnoreBOM: identifier.UTF16BE,
+		UseBOM:    identifier.UTF16, // BigEnding default is preferred by RFC 2781.
+		// TODO: acceptBOM | strictBOM would map to UTF16BE as well.
+	},
+	LittleEndian: [numBOMValues]identifier.MIB{
+		IgnoreBOM: identifier.UTF16LE,
+		UseBOM:    identifier.UTF16, // LittleEndian default is allowed and preferred on Windows.
+		// TODO: acceptBOM | strictBOM would map to UTF16LE as well.
+	},
+	// ExpectBOM is not widely used and has no valid MIB identifier.
+}
+
+// All lists a configuration for each IANA-defined UTF-16 variant.
+var All = []encoding.Encoding{
+	UTF8,
+	UTF16(BigEndian, UseBOM),
+	UTF16(BigEndian, IgnoreBOM),
+	UTF16(LittleEndian, IgnoreBOM),
+}
+
+// BOMPolicy is a UTF-16 encoding's byte order mark policy.
+type BOMPolicy uint8
+
+const (
+	writeBOM   BOMPolicy = 0x01
+	acceptBOM  BOMPolicy = 0x02
+	requireBOM BOMPolicy = 0x04
+	bomMask    BOMPolicy = 0x07
+
+	// HACK: numBOMValues == 8 triggers a bug in the 1.4 compiler (cannot have a
+	// map of an array of length 8 of a type that is also used as a key or value
+	// in another map). See golang.org/issue/11354.
+	// TODO: consider changing this value back to 8 if the use of 1.4.* has
+	// been minimized.
+	numBOMValues = 8 + 1
+
+	// IgnoreBOM means to ignore any byte order marks.
+	IgnoreBOM BOMPolicy = 0
+	// Common and RFC 2781-compliant interpretation for UTF-16BE/LE.
+
+	// UseBOM means that the UTF-16 form may start with a byte order mark, which
+	// will be used to override the default encoding.
+	UseBOM BOMPolicy = writeBOM | acceptBOM
+	// Common and RFC 2781-compliant interpretation for UTF-16.
+
+	// ExpectBOM means that the UTF-16 form must start with a byte order mark,
+	// which will be used to override the default encoding.
+	ExpectBOM BOMPolicy = writeBOM | acceptBOM | requireBOM
+	// Used in Java as Unicode (not to be confused with Java's UTF-16) and
+	// ICU's UTF-16,version=1. Not compliant with RFC 2781.
+
+	// TODO (maybe): strictBOM: BOM must match Endianness. This would allow:
+	// - UTF-16(B|L)E,version=1: writeBOM | acceptBOM | requireBOM | strictBOM
+	//    (UnicodeBig and UnicodeLittle in Java)
+	// - RFC 2781-compliant, but less common interpretation for UTF-16(B|L)E:
+	//    acceptBOM | strictBOM (e.g. assigned to CheckBOM).
+	// This addition would be consistent with supporting ExpectBOM.
+)
+
+// Endianness is a UTF-16 encoding's default endianness.
+type Endianness bool
+
+const (
+	// BigEndian is UTF-16BE.
+	BigEndian Endianness = false
+	// LittleEndian is UTF-16LE.
+	LittleEndian Endianness = true
+)
+
+// ErrMissingBOM means that decoding UTF-16 input with ExpectBOM did not find a
+// starting byte order mark.
+var ErrMissingBOM = errors.New("encoding: missing byte order mark")
+
+type utf16Encoding struct {
+	config
+	mib identifier.MIB
+}
+
+type config struct {
+	endianness Endianness
+	bomPolicy  BOMPolicy
+}
+
+func (u utf16Encoding) NewDecoder() *encoding.Decoder {
+	return &encoding.Decoder{Transformer: &utf16Decoder{
+		initial: u.config,
+		current: u.config,
+	}}
+}
+
+func (u utf16Encoding) NewEncoder() *encoding.Encoder {
+	return &encoding.Encoder{Transformer: &utf16Encoder{
+		endianness:       u.endianness,
+		initialBOMPolicy: u.bomPolicy,
+		currentBOMPolicy: u.bomPolicy,
+	}}
+}
+
+func (u utf16Encoding) ID() (mib identifier.MIB, other string) {
+	return u.mib, ""
+}
+
+func (u utf16Encoding) String() string {
+	e, b := "B", ""
+	if u.endianness == LittleEndian {
+		e = "L"
+	}
+	switch u.bomPolicy {
+	case ExpectBOM:
+		b = "Expect"
+	case UseBOM:
+		b = "Use"
+	case IgnoreBOM:
+		b = "Ignore"
+	}
+	return "UTF-16" + e + "E (" + b + " BOM)"
+}
+
+type utf16Decoder struct {
+	initial config
+	current config
+}
+
+func (u *utf16Decoder) Reset() {
+	u.current = u.initial
+}
+
+func (u *utf16Decoder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	if len(src) < 2 && atEOF && u.current.bomPolicy&requireBOM != 0 {
+		return 0, 0, ErrMissingBOM
+	}
+	if len(src) == 0 {
+		return 0, 0, nil
+	}
+	if len(src) >= 2 && u.current.bomPolicy&acceptBOM != 0 {
+		switch {
+		case src[0] == 0xfe && src[1] == 0xff:
+			u.current.endianness = BigEndian
+			nSrc = 2
+		case src[0] == 0xff && src[1] == 0xfe:
+			u.current.endianness = LittleEndian
+			nSrc = 2
+		default:
+			if u.current.bomPolicy&requireBOM != 0 {
+				return 0, 0, ErrMissingBOM
+			}
+		}
+		u.current.bomPolicy = IgnoreBOM
+	}
+
+	var r rune
+	var dSize, sSize int
+	for nSrc < len(src) {
+		if nSrc+1 < len(src) {
+			x := uint16(src[nSrc+0])<<8 | uint16(src[nSrc+1])
+			if u.current.endianness == LittleEndian {
+				x = x>>8 | x<<8
+			}
+			r, sSize = rune(x), 2
+			if utf16.IsSurrogate(r) {
+				if nSrc+3 < len(src) {
+					x = uint16(src[nSrc+2])<<8 | uint16(src[nSrc+3])
+					if u.current.endianness == LittleEndian {
+						x = x>>8 | x<<8
+					}
+					// Save for next iteration if it is not a high surrogate.
+					if isHighSurrogate(rune(x)) {
+						r, sSize = utf16.DecodeRune(r, rune(x)), 4
+					}
+				} else if !atEOF {
+					err = transform.ErrShortSrc
+					break
+				}
+			}
+			if dSize = utf8.RuneLen(r); dSize < 0 {
+				r, dSize = utf8.RuneError, 3
+			}
+		} else if atEOF {
+			// Single trailing byte.
+			r, dSize, sSize = utf8.RuneError, 3, 1
+		} else {
+			err = transform.ErrShortSrc
+			break
+		}
+		if nDst+dSize > len(dst) {
+			err = transform.ErrShortDst
+			break
+		}
+		nDst += utf8.EncodeRune(dst[nDst:], r)
+		nSrc += sSize
+	}
+	return nDst, nSrc, err
+}
+
+func isHighSurrogate(r rune) bool {
+	return 0xDC00 <= r && r <= 0xDFFF
+}
+
+type utf16Encoder struct {
+	endianness       Endianness
+	initialBOMPolicy BOMPolicy
+	currentBOMPolicy BOMPolicy
+}
+
+func (u *utf16Encoder) Reset() {
+	u.currentBOMPolicy = u.initialBOMPolicy
+}
+
+func (u *utf16Encoder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	if u.currentBOMPolicy&writeBOM != 0 {
+		if len(dst) < 2 {
+			return 0, 0, transform.ErrShortDst
+		}
+		dst[0], dst[1] = 0xfe, 0xff
+		u.currentBOMPolicy = IgnoreBOM
+		nDst = 2
+	}
+
+	r, size := rune(0), 0
+	for nSrc < len(src) {
+		r = rune(src[nSrc])
+
+		// Decode a 1-byte rune.
+		if r < utf8.RuneSelf {
+			size = 1
+
+		} else {
+			// Decode a multi-byte rune.
+			r, size = utf8.DecodeRune(src[nSrc:])
+			if size == 1 {
+				// All valid runes of size 1 (those below utf8.RuneSelf) were
+				// handled above. We have invalid UTF-8 or we haven't seen the
+				// full character yet.
+				if !atEOF && !utf8.FullRune(src[nSrc:]) {
+					err = transform.ErrShortSrc
+					break
+				}
+			}
+		}
+
+		if r <= 0xffff {
+			if nDst+2 > len(dst) {
+				err = transform.ErrShortDst
+				break
+			}
+			dst[nDst+0] = uint8(r >> 8)
+			dst[nDst+1] = uint8(r)
+			nDst += 2
+		} else {
+			if nDst+4 > len(dst) {
+				err = transform.ErrShortDst
+				break
+			}
+			r1, r2 := utf16.EncodeRune(r)
+			dst[nDst+0] = uint8(r1 >> 8)
+			dst[nDst+1] = uint8(r1)
+			dst[nDst+2] = uint8(r2 >> 8)
+			dst[nDst+3] = uint8(r2)
+			nDst += 4
+		}
+		nSrc += size
+	}
+
+	if u.endianness == LittleEndian {
+		for i := 0; i < nDst; i += 2 {
+			dst[i], dst[i+1] = dst[i+1], dst[i]
+		}
+	}
+	return nDst, nSrc, err
+}
diff --git a/vendor/golang.org/x/text/internal/utf8internal/utf8internal.go b/vendor/golang.org/x/text/internal/utf8internal/utf8internal.go
new file mode 100644
index 00000000..575cea87
--- /dev/null
+++ b/vendor/golang.org/x/text/internal/utf8internal/utf8internal.go
@@ -0,0 +1,87 @@
+// Copyright 2015 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 utf8internal contains low-level utf8-related constants, tables, etc.
+// that are used internally by the text package.
+package utf8internal
+
+// The default lowest and highest continuation byte.
+const (
+	LoCB = 0x80 // 1000 0000
+	HiCB = 0xBF // 1011 1111
+)
+
+// Constants related to getting information of first bytes of UTF-8 sequences.
+const (
+	// ASCII identifies a UTF-8 byte as ASCII.
+	ASCII = as
+
+	// FirstInvalid indicates a byte is invalid as a first byte of a UTF-8
+	// sequence.
+	FirstInvalid = xx
+
+	// SizeMask is a mask for the size bits. Use use x&SizeMask to get the size.
+	SizeMask = 7
+
+	// AcceptShift is the right-shift count for the first byte info byte to get
+	// the index into the AcceptRanges table. See AcceptRanges.
+	AcceptShift = 4
+
+	// The names of these constants are chosen to give nice alignment in the
+	// table below. The first nibble is an index into acceptRanges or F for
+	// special one-byte cases. The second nibble is the Rune length or the
+	// Status for the special one-byte case.
+	xx = 0xF1 // invalid: size 1
+	as = 0xF0 // ASCII: size 1
+	s1 = 0x02 // accept 0, size 2
+	s2 = 0x13 // accept 1, size 3
+	s3 = 0x03 // accept 0, size 3
+	s4 = 0x23 // accept 2, size 3
+	s5 = 0x34 // accept 3, size 4
+	s6 = 0x04 // accept 0, size 4
+	s7 = 0x44 // accept 4, size 4
+)
+
+// First is information about the first byte in a UTF-8 sequence.
+var First = [256]uint8{
+	//   1   2   3   4   5   6   7   8   9   A   B   C   D   E   F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x00-0x0F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x10-0x1F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x20-0x2F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x30-0x3F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x40-0x4F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x50-0x5F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x60-0x6F
+	as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, as, // 0x70-0x7F
+	//   1   2   3   4   5   6   7   8   9   A   B   C   D   E   F
+	xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0x80-0x8F
+	xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0x90-0x9F
+	xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xA0-0xAF
+	xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xB0-0xBF
+	xx, xx, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, // 0xC0-0xCF
+	s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, s1, // 0xD0-0xDF
+	s2, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s3, s4, s3, s3, // 0xE0-0xEF
+	s5, s6, s6, s6, s7, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, xx, // 0xF0-0xFF
+}
+
+// AcceptRange gives the range of valid values for the second byte in a UTF-8
+// sequence for any value for First that is not ASCII or FirstInvalid.
+type AcceptRange struct {
+	Lo uint8 // lowest value for second byte.
+	Hi uint8 // highest value for second byte.
+}
+
+// AcceptRanges is a slice of AcceptRange values. For a given byte sequence b
+//
+//		AcceptRanges[First[b[0]]>>AcceptShift]
+//
+// will give the value of AcceptRange for the multi-byte UTF-8 sequence starting
+// at b[0].
+var AcceptRanges = [...]AcceptRange{
+	0: {LoCB, HiCB},
+	1: {0xA0, HiCB},
+	2: {LoCB, 0x9F},
+	3: {0x90, HiCB},
+	4: {LoCB, 0x8F},
+}