David K. Bainbridge | 215e024 | 2017-09-05 23:18:24 -0700 | [diff] [blame] | 1 | // Copyright 2012 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_test |
| 6 | |
| 7 | import ( |
| 8 | "bytes" |
| 9 | "fmt" |
| 10 | "unicode/utf8" |
| 11 | |
| 12 | "golang.org/x/text/unicode/norm" |
| 13 | ) |
| 14 | |
| 15 | // EqualSimple uses a norm.Iter to compare two non-normalized |
| 16 | // strings for equivalence. |
| 17 | func EqualSimple(a, b string) bool { |
| 18 | var ia, ib norm.Iter |
| 19 | ia.InitString(norm.NFKD, a) |
| 20 | ib.InitString(norm.NFKD, b) |
| 21 | for !ia.Done() && !ib.Done() { |
| 22 | if !bytes.Equal(ia.Next(), ib.Next()) { |
| 23 | return false |
| 24 | } |
| 25 | } |
| 26 | return ia.Done() && ib.Done() |
| 27 | } |
| 28 | |
| 29 | // FindPrefix finds the longest common prefix of ASCII characters |
| 30 | // of a and b. |
| 31 | func FindPrefix(a, b string) int { |
| 32 | i := 0 |
| 33 | for ; i < len(a) && i < len(b) && a[i] < utf8.RuneSelf && a[i] == b[i]; i++ { |
| 34 | } |
| 35 | return i |
| 36 | } |
| 37 | |
| 38 | // EqualOpt is like EqualSimple, but optimizes the special |
| 39 | // case for ASCII characters. |
| 40 | func EqualOpt(a, b string) bool { |
| 41 | n := FindPrefix(a, b) |
| 42 | a, b = a[n:], b[n:] |
| 43 | var ia, ib norm.Iter |
| 44 | ia.InitString(norm.NFKD, a) |
| 45 | ib.InitString(norm.NFKD, b) |
| 46 | for !ia.Done() && !ib.Done() { |
| 47 | if !bytes.Equal(ia.Next(), ib.Next()) { |
| 48 | return false |
| 49 | } |
| 50 | if n := int64(FindPrefix(a[ia.Pos():], b[ib.Pos():])); n != 0 { |
| 51 | ia.Seek(n, 1) |
| 52 | ib.Seek(n, 1) |
| 53 | } |
| 54 | } |
| 55 | return ia.Done() && ib.Done() |
| 56 | } |
| 57 | |
| 58 | var compareTests = []struct{ a, b string }{ |
| 59 | {"aaa", "aaa"}, |
| 60 | {"aaa", "aab"}, |
| 61 | {"a\u0300a", "\u00E0a"}, |
| 62 | {"a\u0300\u0320b", "a\u0320\u0300b"}, |
| 63 | {"\u1E0A\u0323", "\x44\u0323\u0307"}, |
| 64 | // A character that decomposes into multiple segments |
| 65 | // spans several iterations. |
| 66 | {"\u3304", "\u30A4\u30CB\u30F3\u30AF\u3099"}, |
| 67 | } |
| 68 | |
| 69 | func ExampleIter() { |
| 70 | for i, t := range compareTests { |
| 71 | r0 := EqualSimple(t.a, t.b) |
| 72 | r1 := EqualOpt(t.a, t.b) |
| 73 | fmt.Printf("%d: %v %v\n", i, r0, r1) |
| 74 | } |
| 75 | // Output: |
| 76 | // 0: true true |
| 77 | // 1: false false |
| 78 | // 2: true true |
| 79 | // 3: true true |
| 80 | // 4: true true |
| 81 | // 5: true true |
| 82 | } |