William Kurkian | ea86948 | 2019-04-09 15:16:11 -0400 | [diff] [blame] | 1 | // Copyright 2011 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 norm |
| 6 | |
| 7 | import "unicode/utf8" |
| 8 | |
| 9 | type input struct { |
| 10 | str string |
| 11 | bytes []byte |
| 12 | } |
| 13 | |
| 14 | func inputBytes(str []byte) input { |
| 15 | return input{bytes: str} |
| 16 | } |
| 17 | |
| 18 | func inputString(str string) input { |
| 19 | return input{str: str} |
| 20 | } |
| 21 | |
| 22 | func (in *input) setBytes(str []byte) { |
| 23 | in.str = "" |
| 24 | in.bytes = str |
| 25 | } |
| 26 | |
| 27 | func (in *input) setString(str string) { |
| 28 | in.str = str |
| 29 | in.bytes = nil |
| 30 | } |
| 31 | |
| 32 | func (in *input) _byte(p int) byte { |
| 33 | if in.bytes == nil { |
| 34 | return in.str[p] |
| 35 | } |
| 36 | return in.bytes[p] |
| 37 | } |
| 38 | |
| 39 | func (in *input) skipASCII(p, max int) int { |
| 40 | if in.bytes == nil { |
| 41 | for ; p < max && in.str[p] < utf8.RuneSelf; p++ { |
| 42 | } |
| 43 | } else { |
| 44 | for ; p < max && in.bytes[p] < utf8.RuneSelf; p++ { |
| 45 | } |
| 46 | } |
| 47 | return p |
| 48 | } |
| 49 | |
| 50 | func (in *input) skipContinuationBytes(p int) int { |
| 51 | if in.bytes == nil { |
| 52 | for ; p < len(in.str) && !utf8.RuneStart(in.str[p]); p++ { |
| 53 | } |
| 54 | } else { |
| 55 | for ; p < len(in.bytes) && !utf8.RuneStart(in.bytes[p]); p++ { |
| 56 | } |
| 57 | } |
| 58 | return p |
| 59 | } |
| 60 | |
| 61 | func (in *input) appendSlice(buf []byte, b, e int) []byte { |
| 62 | if in.bytes != nil { |
| 63 | return append(buf, in.bytes[b:e]...) |
| 64 | } |
| 65 | for i := b; i < e; i++ { |
| 66 | buf = append(buf, in.str[i]) |
| 67 | } |
| 68 | return buf |
| 69 | } |
| 70 | |
| 71 | func (in *input) copySlice(buf []byte, b, e int) int { |
| 72 | if in.bytes == nil { |
| 73 | return copy(buf, in.str[b:e]) |
| 74 | } |
| 75 | return copy(buf, in.bytes[b:e]) |
| 76 | } |
| 77 | |
| 78 | func (in *input) charinfoNFC(p int) (uint16, int) { |
| 79 | if in.bytes == nil { |
| 80 | return nfcData.lookupString(in.str[p:]) |
| 81 | } |
| 82 | return nfcData.lookup(in.bytes[p:]) |
| 83 | } |
| 84 | |
| 85 | func (in *input) charinfoNFKC(p int) (uint16, int) { |
| 86 | if in.bytes == nil { |
| 87 | return nfkcData.lookupString(in.str[p:]) |
| 88 | } |
| 89 | return nfkcData.lookup(in.bytes[p:]) |
| 90 | } |
| 91 | |
| 92 | func (in *input) hangul(p int) (r rune) { |
| 93 | var size int |
| 94 | if in.bytes == nil { |
| 95 | if !isHangulString(in.str[p:]) { |
| 96 | return 0 |
| 97 | } |
| 98 | r, size = utf8.DecodeRuneInString(in.str[p:]) |
| 99 | } else { |
| 100 | if !isHangul(in.bytes[p:]) { |
| 101 | return 0 |
| 102 | } |
| 103 | r, size = utf8.DecodeRune(in.bytes[p:]) |
| 104 | } |
| 105 | if size != hangulUTF8Size { |
| 106 | return 0 |
| 107 | } |
| 108 | return r |
| 109 | } |