Source file src/pkg/runtime/utf8.go

     1	// Copyright 2016 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 runtime
     6	
     7	// Numbers fundamental to the encoding.
     8	const (
     9		runeError = '\uFFFD'     // the "error" Rune or "Unicode replacement character"
    10		runeSelf  = 0x80         // characters below Runeself are represented as themselves in a single byte.
    11		maxRune   = '\U0010FFFF' // Maximum valid Unicode code point.
    12	)
    13	
    14	// Code points in the surrogate range are not valid for UTF-8.
    15	const (
    16		surrogateMin = 0xD800
    17		surrogateMax = 0xDFFF
    18	)
    19	
    20	const (
    21		t1 = 0x00 // 0000 0000
    22		tx = 0x80 // 1000 0000
    23		t2 = 0xC0 // 1100 0000
    24		t3 = 0xE0 // 1110 0000
    25		t4 = 0xF0 // 1111 0000
    26		t5 = 0xF8 // 1111 1000
    27	
    28		maskx = 0x3F // 0011 1111
    29		mask2 = 0x1F // 0001 1111
    30		mask3 = 0x0F // 0000 1111
    31		mask4 = 0x07 // 0000 0111
    32	
    33		rune1Max = 1<<7 - 1
    34		rune2Max = 1<<11 - 1
    35		rune3Max = 1<<16 - 1
    36	
    37		// The default lowest and highest continuation byte.
    38		locb = 0x80 // 1000 0000
    39		hicb = 0xBF // 1011 1111
    40	)
    41	
    42	// countrunes returns the number of runes in s.
    43	func countrunes(s string) int {
    44		n := 0
    45		for range s {
    46			n++
    47		}
    48		return n
    49	}
    50	
    51	// decoderune returns the non-ASCII rune at the start of
    52	// s[k:] and the index after the rune in s.
    53	//
    54	// decoderune assumes that caller has checked that
    55	// the to be decoded rune is a non-ASCII rune.
    56	//
    57	// If the string appears to be incomplete or decoding problems
    58	// are encountered (runeerror, k + 1) is returned to ensure
    59	// progress when decoderune is used to iterate over a string.
    60	func decoderune(s string, k int) (r rune, pos int) {
    61		pos = k
    62	
    63		if k >= len(s) {
    64			return runeError, k + 1
    65		}
    66	
    67		s = s[k:]
    68	
    69		switch {
    70		case t2 <= s[0] && s[0] < t3:
    71			// 0080-07FF two byte sequence
    72			if len(s) > 1 && (locb <= s[1] && s[1] <= hicb) {
    73				r = rune(s[0]&mask2)<<6 | rune(s[1]&maskx)
    74				pos += 2
    75				if rune1Max < r {
    76					return
    77				}
    78			}
    79		case t3 <= s[0] && s[0] < t4:
    80			// 0800-FFFF three byte sequence
    81			if len(s) > 2 && (locb <= s[1] && s[1] <= hicb) && (locb <= s[2] && s[2] <= hicb) {
    82				r = rune(s[0]&mask3)<<12 | rune(s[1]&maskx)<<6 | rune(s[2]&maskx)
    83				pos += 3
    84				if rune2Max < r && !(surrogateMin <= r && r <= surrogateMax) {
    85					return
    86				}
    87			}
    88		case t4 <= s[0] && s[0] < t5:
    89			// 10000-1FFFFF four byte sequence
    90			if len(s) > 3 && (locb <= s[1] && s[1] <= hicb) && (locb <= s[2] && s[2] <= hicb) && (locb <= s[3] && s[3] <= hicb) {
    91				r = rune(s[0]&mask4)<<18 | rune(s[1]&maskx)<<12 | rune(s[2]&maskx)<<6 | rune(s[3]&maskx)
    92				pos += 4
    93				if rune3Max < r && r <= maxRune {
    94					return
    95				}
    96			}
    97		}
    98	
    99		return runeError, k + 1
   100	}
   101	
   102	// encoderune writes into p (which must be large enough) the UTF-8 encoding of the rune.
   103	// It returns the number of bytes written.
   104	func encoderune(p []byte, r rune) int {
   105		// Negative values are erroneous. Making it unsigned addresses the problem.
   106		switch i := uint32(r); {
   107		case i <= rune1Max:
   108			p[0] = byte(r)
   109			return 1
   110		case i <= rune2Max:
   111			_ = p[1] // eliminate bounds checks
   112			p[0] = t2 | byte(r>>6)
   113			p[1] = tx | byte(r)&maskx
   114			return 2
   115		case i > maxRune, surrogateMin <= i && i <= surrogateMax:
   116			r = runeError
   117			fallthrough
   118		case i <= rune3Max:
   119			_ = p[2] // eliminate bounds checks
   120			p[0] = t3 | byte(r>>12)
   121			p[1] = tx | byte(r>>6)&maskx
   122			p[2] = tx | byte(r)&maskx
   123			return 3
   124		default:
   125			_ = p[3] // eliminate bounds checks
   126			p[0] = t4 | byte(r>>18)
   127			p[1] = tx | byte(r>>12)&maskx
   128			p[2] = tx | byte(r>>6)&maskx
   129			p[3] = tx | byte(r)&maskx
   130			return 4
   131		}
   132	}
   133	

View as plain text