Home | History | Annotate | Download | only in utf16 
      1 // Copyright 2010 The Go Authors.  All rights reserved.
      2 // Use of this source code is governed by a BSD-style
      3 // license that can be found in the LICENSE file.
      4 
      5 // Package utf16 implements encoding and decoding of UTF-16 sequences.
      6 package utf16
      7 
      8 // The conditions replacementChar==unicode.ReplacementChar and
      9 // maxRune==unicode.MaxRune are verified in the tests.
     10 // Defining them locally avoids this package depending on package unicode.
     11 
     12 const (
     13 	replacementChar = '\uFFFD'     // Unicode replacement character
     14 	maxRune         = '\U0010FFFF' // Maximum valid Unicode code point.
     15 )
     16 
     17 const (
     18 	// 0xd800-0xdc00 encodes the high 10 bits of a pair.
     19 	// 0xdc00-0xe000 encodes the low 10 bits of a pair.
     20 	// the value is those 20 bits plus 0x10000.
     21 	surr1 = 0xd800
     22 	surr2 = 0xdc00
     23 	surr3 = 0xe000
     24 
     25 	surrSelf = 0x10000
     26 )
     27 
     28 // IsSurrogate reports whether the specified Unicode code point
     29 // can appear in a surrogate pair.
     30 func IsSurrogate(r rune) bool {
     31 	return surr1 <= r && r < surr3
     32 }
     33 
     34 // DecodeRune returns the UTF-16 decoding of a surrogate pair.
     35 // If the pair is not a valid UTF-16 surrogate pair, DecodeRune returns
     36 // the Unicode replacement code point U+FFFD.
     37 func DecodeRune(r1, r2 rune) rune {
     38 	if surr1 <= r1 && r1 < surr2 && surr2 <= r2 && r2 < surr3 {
     39 		return (r1-surr1)<<10 | (r2 - surr2) + 0x10000
     40 	}
     41 	return replacementChar
     42 }
     43 
     44 // EncodeRune returns the UTF-16 surrogate pair r1, r2 for the given rune.
     45 // If the rune is not a valid Unicode code point or does not need encoding,
     46 // EncodeRune returns U+FFFD, U+FFFD.
     47 func EncodeRune(r rune) (r1, r2 rune) {
     48 	if r < surrSelf || r > maxRune || IsSurrogate(r) {
     49 		return replacementChar, replacementChar
     50 	}
     51 	r -= surrSelf
     52 	return surr1 + (r>>10)&0x3ff, surr2 + r&0x3ff
     53 }
     54 
     55 // Encode returns the UTF-16 encoding of the Unicode code point sequence s.
     56 func Encode(s []rune) []uint16 {
     57 	n := len(s)
     58 	for _, v := range s {
     59 		if v >= surrSelf {
     60 			n++
     61 		}
     62 	}
     63 
     64 	a := make([]uint16, n)
     65 	n = 0
     66 	for _, v := range s {
     67 		switch {
     68 		case v < 0, surr1 <= v && v < surr3, v > maxRune:
     69 			v = replacementChar
     70 			fallthrough
     71 		case v < surrSelf:
     72 			a[n] = uint16(v)
     73 			n++
     74 		default:
     75 			r1, r2 := EncodeRune(v)
     76 			a[n] = uint16(r1)
     77 			a[n+1] = uint16(r2)
     78 			n += 2
     79 		}
     80 	}
     81 	return a[0:n]
     82 }
     83 
     84 // Decode returns the Unicode code point sequence represented
     85 // by the UTF-16 encoding s.
     86 func Decode(s []uint16) []rune {
     87 	a := make([]rune, len(s))
     88 	n := 0
     89 	for i := 0; i < len(s); i++ {
     90 		switch r := s[i]; {
     91 		case surr1 <= r && r < surr2 && i+1 < len(s) &&
     92 			surr2 <= s[i+1] && s[i+1] < surr3:
     93 			// valid surrogate sequence
     94 			a[n] = DecodeRune(rune(r), rune(s[i+1]))
     95 			i++
     96 			n++
     97 		case surr1 <= r && r < surr3:
     98 			// invalid surrogate sequence
     99 			a[n] = replacementChar
    100 			n++
    101 		default:
    102 			// normal rune
    103 			a[n] = rune(r)
    104 			n++
    105 		}
    106 	}
    107 	return a[0:n]
    108 }
    109