// 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 (
	
	
	
	

	
	
	
	
	
	
)

// 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 {
	return "UTF-8-BOM"
}

func (utf8bomEncoding) () (identifier.MIB, string) {
	return identifier.Unofficial, "x-utf8bom"
}

func (utf8bomEncoding) () *encoding.Encoder {
	return &encoding.Encoder{
		Transformer: &utf8bomEncoder{t: runes.ReplaceIllFormed()},
	}
}

func (utf8bomEncoding) () *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 ( *utf8bomDecoder) () {
	.checked = false
}

func ( *utf8bomDecoder) (,  []byte,  bool) (,  int,  error) {
	if !.checked {
		if ! && len() < len(utf8BOM) {
			if len() == 0 {
				return 0, 0, nil
			}
			return 0, 0, transform.ErrShortSrc
		}
		if bytes.HasPrefix(, []byte(utf8BOM)) {
			 += len(utf8BOM)
			 = [len(utf8BOM):]
		}
		.checked = true
	}
	, ,  := utf8Decoder.Transform(utf8Decoder{}, [:], , )
	 += 
	return , , 
}

type utf8bomEncoder struct {
	written bool
	t       transform.Transformer
}

func ( *utf8bomEncoder) () {
	.written = false
	.t.Reset()
}

func ( *utf8bomEncoder) (,  []byte,  bool) (,  int,  error) {
	if !.written {
		if len() < len(utf8BOM) {
			return , 0, transform.ErrShortDst
		}
		 = copy(, utf8BOM)
		.written = true
	}
	, ,  := utf8Decoder.Transform(utf8Decoder{}, [:], , )
	 += 
	return , , 
}

type utf8Decoder struct{ transform.NopResetter }

func (utf8Decoder) (,  []byte,  bool) (,  int,  error) {
	var  int // point from which to start copy in src
	var  utf8internal.AcceptRange

	// The decoder can only make the input larger, not smaller.
	 := len()
	if len() <  {
		 = transform.ErrShortDst
		 = len()
		 = false
	}
	for  <  {
		 := []
		if  < utf8.RuneSelf {
			++
			continue
		}
		 := utf8internal.First[]
		 := int( & utf8internal.SizeMask)
		if  == utf8internal.FirstInvalid {
			goto  // invalid starter byte
		}
		 = utf8internal.AcceptRanges[>>utf8internal.AcceptShift]
		if + >  {
			if ! {
				// 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  == nil {
					 = transform.ErrShortSrc
				}
				break
			}
			// Determine the maximal subpart of an ill-formed subsequence.
			switch {
			case +1 >=  || [+1] < .Lo || .Hi < [+1]:
				 = 1
			case +2 >=  || [+2] < utf8internal.LoCB || utf8internal.HiCB < [+2]:
				 = 2
			default:
				 = 3 // As we are short, the maximum is 3.
			}
			goto 
		}
		if  = [+1];  < .Lo || .Hi <  {
			 = 1
			goto  // invalid continuation byte
		} else if  == 2 {
		} else if  = [+2];  < utf8internal.LoCB || utf8internal.HiCB <  {
			 = 2
			goto  // invalid continuation byte
		} else if  == 3 {
		} else if  = [+3];  < utf8internal.LoCB || utf8internal.HiCB <  {
			 = 3
			goto  // invalid continuation byte
		}
		 += 
		continue

	:
		// Copy the scanned input so far.
		 += copy([:], [:])

		// Append RuneError to the destination.
		const  = "\ufffd"
		if +len() > len() {
			return , , transform.ErrShortDst
		}
		 += copy([:], )

		// 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.
		 += 
		 = 

		// Recompute the maximum source length.
		if  := len() - ;  < len()- {
			 = transform.ErrShortDst
			 =  + 
			 = false
		}
	}
	return  + copy([:], [:]), , 
}

// 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 ( Endianness,  BOMPolicy) encoding.Encoding {
	return utf16Encoding{config{, }, mibValue[][&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 ( utf16Encoding) () *encoding.Decoder {
	return &encoding.Decoder{Transformer: &utf16Decoder{
		initial: .config,
		current: .config,
	}}
}

func ( utf16Encoding) () *encoding.Encoder {
	return &encoding.Encoder{Transformer: &utf16Encoder{
		endianness:       .endianness,
		initialBOMPolicy: .bomPolicy,
		currentBOMPolicy: .bomPolicy,
	}}
}

func ( utf16Encoding) () ( identifier.MIB,  string) {
	return .mib, ""
}

func ( utf16Encoding) () string {
	,  := "B", ""
	if .endianness == LittleEndian {
		 = "L"
	}
	switch .bomPolicy {
	case ExpectBOM:
		 = "Expect"
	case UseBOM:
		 = "Use"
	case IgnoreBOM:
		 = "Ignore"
	}
	return "UTF-16" +  + "E (" +  + " BOM)"
}

type utf16Decoder struct {
	initial config
	current config
}

func ( *utf16Decoder) () {
	.current = .initial
}

func ( *utf16Decoder) (,  []byte,  bool) (,  int,  error) {
	if len() < 2 &&  && .current.bomPolicy&requireBOM != 0 {
		return 0, 0, ErrMissingBOM
	}
	if len() == 0 {
		return 0, 0, nil
	}
	if len() >= 2 && .current.bomPolicy&acceptBOM != 0 {
		switch {
		case [0] == 0xfe && [1] == 0xff:
			.current.endianness = BigEndian
			 = 2
		case [0] == 0xff && [1] == 0xfe:
			.current.endianness = LittleEndian
			 = 2
		default:
			if .current.bomPolicy&requireBOM != 0 {
				return 0, 0, ErrMissingBOM
			}
		}
		.current.bomPolicy = IgnoreBOM
	}

	var  rune
	var ,  int
	for  < len() {
		if +1 < len() {
			 := uint16([+0])<<8 | uint16([+1])
			if .current.endianness == LittleEndian {
				 = >>8 | <<8
			}
			,  = rune(), 2
			if utf16.IsSurrogate() {
				if +3 < len() {
					 = uint16([+2])<<8 | uint16([+3])
					if .current.endianness == LittleEndian {
						 = >>8 | <<8
					}
					// Save for next iteration if it is not a high surrogate.
					if isHighSurrogate(rune()) {
						,  = utf16.DecodeRune(, rune()), 4
					}
				} else if ! {
					 = transform.ErrShortSrc
					break
				}
			}
			if  = utf8.RuneLen();  < 0 {
				,  = utf8.RuneError, 3
			}
		} else if  {
			// Single trailing byte.
			, ,  = utf8.RuneError, 3, 1
		} else {
			 = transform.ErrShortSrc
			break
		}
		if + > len() {
			 = transform.ErrShortDst
			break
		}
		 += utf8.EncodeRune([:], )
		 += 
	}
	return , , 
}

func isHighSurrogate( rune) bool {
	return 0xDC00 <=  &&  <= 0xDFFF
}

type utf16Encoder struct {
	endianness       Endianness
	initialBOMPolicy BOMPolicy
	currentBOMPolicy BOMPolicy
}

func ( *utf16Encoder) () {
	.currentBOMPolicy = .initialBOMPolicy
}

func ( *utf16Encoder) (,  []byte,  bool) (,  int,  error) {
	if .currentBOMPolicy&writeBOM != 0 {
		if len() < 2 {
			return 0, 0, transform.ErrShortDst
		}
		[0], [1] = 0xfe, 0xff
		.currentBOMPolicy = IgnoreBOM
		 = 2
	}

	,  := rune(0), 0
	for  < len() {
		 = rune([])

		// Decode a 1-byte rune.
		if  < utf8.RuneSelf {
			 = 1

		} else {
			// Decode a multi-byte rune.
			,  = utf8.DecodeRune([:])
			if  == 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 ! && !utf8.FullRune([:]) {
					 = transform.ErrShortSrc
					break
				}
			}
		}

		if  <= 0xffff {
			if +2 > len() {
				 = transform.ErrShortDst
				break
			}
			[+0] = uint8( >> 8)
			[+1] = uint8()
			 += 2
		} else {
			if +4 > len() {
				 = transform.ErrShortDst
				break
			}
			,  := utf16.EncodeRune()
			[+0] = uint8( >> 8)
			[+1] = uint8()
			[+2] = uint8( >> 8)
			[+3] = uint8()
			 += 4
		}
		 += 
	}

	if .endianness == LittleEndian {
		for  := 0;  < ;  += 2 {
			[], [+1] = [+1], []
		}
	}
	return , , 
}