OLD | NEW |
1 // Copyright (c) 2012 The Chromium Authors. All rights reserved. | 1 // Copyright (c) 2012 The Chromium Authors. All rights reserved. |
2 // Use of this source code is governed by a BSD-style license that can be | 2 // Use of this source code is governed by a BSD-style license that can be |
3 // found in the LICENSE file. | 3 // found in the LICENSE file. |
4 | 4 |
5 #include "ui/gfx/utf16_indexing.h" | 5 #include "ui/gfx/utf16_indexing.h" |
6 | 6 |
7 #include "base/logging.h" | 7 #include "base/logging.h" |
8 #include "base/third_party/icu/icu_utf.h" | 8 #include "base/third_party/icu/icu_utf.h" |
9 | 9 |
10 namespace gfx { | 10 namespace gfx { |
11 | 11 |
12 bool IsValidCodePointIndex(const string16& s, size_t index) { | 12 bool IsValidCodePointIndex(const base::string16& s, size_t index) { |
13 return index == 0 || index == s.length() || | 13 return index == 0 || index == s.length() || |
14 !(CBU16_IS_TRAIL(s[index]) && CBU16_IS_LEAD(s[index - 1])); | 14 !(CBU16_IS_TRAIL(s[index]) && CBU16_IS_LEAD(s[index - 1])); |
15 } | 15 } |
16 | 16 |
17 ptrdiff_t UTF16IndexToOffset(const string16& s, size_t base, size_t pos) { | 17 ptrdiff_t UTF16IndexToOffset(const base::string16& s, size_t base, size_t pos) { |
18 // The indices point between UTF-16 words (range 0 to s.length() inclusive). | 18 // The indices point between UTF-16 words (range 0 to s.length() inclusive). |
19 // In order to consistently handle indices that point to the middle of a | 19 // In order to consistently handle indices that point to the middle of a |
20 // surrogate pair, we count the first word in that surrogate pair and not | 20 // surrogate pair, we count the first word in that surrogate pair and not |
21 // the second. The test "s[i] is not the second half of a surrogate pair" is | 21 // the second. The test "s[i] is not the second half of a surrogate pair" is |
22 // "IsValidCodePointIndex(s, i)". | 22 // "IsValidCodePointIndex(s, i)". |
23 DCHECK_LE(base, s.length()); | 23 DCHECK_LE(base, s.length()); |
24 DCHECK_LE(pos, s.length()); | 24 DCHECK_LE(pos, s.length()); |
25 ptrdiff_t delta = 0; | 25 ptrdiff_t delta = 0; |
26 while (base < pos) | 26 while (base < pos) |
27 delta += IsValidCodePointIndex(s, base++) ? 1 : 0; | 27 delta += IsValidCodePointIndex(s, base++) ? 1 : 0; |
28 while (pos < base) | 28 while (pos < base) |
29 delta -= IsValidCodePointIndex(s, pos++) ? 1 : 0; | 29 delta -= IsValidCodePointIndex(s, pos++) ? 1 : 0; |
30 return delta; | 30 return delta; |
31 } | 31 } |
32 | 32 |
33 size_t UTF16OffsetToIndex(const string16& s, size_t base, ptrdiff_t offset) { | 33 size_t UTF16OffsetToIndex(const base::string16& s, |
| 34 size_t base, |
| 35 ptrdiff_t offset) { |
34 DCHECK_LE(base, s.length()); | 36 DCHECK_LE(base, s.length()); |
35 // As in UTF16IndexToOffset, we count the first half of a surrogate pair, not | 37 // As in UTF16IndexToOffset, we count the first half of a surrogate pair, not |
36 // the second. When stepping from pos to pos+1 we check s[pos:pos+1] == s[pos] | 38 // the second. When stepping from pos to pos+1 we check s[pos:pos+1] == s[pos] |
37 // (Python syntax), hence pos++. When stepping from pos to pos-1 we check | 39 // (Python syntax), hence pos++. When stepping from pos to pos-1 we check |
38 // s[pos-1], hence --pos. | 40 // s[pos-1], hence --pos. |
39 size_t pos = base; | 41 size_t pos = base; |
40 while (offset > 0 && pos < s.length()) | 42 while (offset > 0 && pos < s.length()) |
41 offset -= IsValidCodePointIndex(s, pos++) ? 1 : 0; | 43 offset -= IsValidCodePointIndex(s, pos++) ? 1 : 0; |
42 while (offset < 0 && pos > 0) | 44 while (offset < 0 && pos > 0) |
43 offset += IsValidCodePointIndex(s, --pos) ? 1 : 0; | 45 offset += IsValidCodePointIndex(s, --pos) ? 1 : 0; |
44 // If offset != 0 then we ran off the edge of the string, which is a contract | 46 // If offset != 0 then we ran off the edge of the string, which is a contract |
45 // violation but is handled anyway (by clamping) in release for safety. | 47 // violation but is handled anyway (by clamping) in release for safety. |
46 DCHECK_EQ(offset, 0); | 48 DCHECK_EQ(offset, 0); |
47 // Since the second half of a surrogate pair has "length" zero, there is an | 49 // Since the second half of a surrogate pair has "length" zero, there is an |
48 // ambiguity in the returned position. Resolve it by always returning a valid | 50 // ambiguity in the returned position. Resolve it by always returning a valid |
49 // index. | 51 // index. |
50 if (!IsValidCodePointIndex(s, pos)) | 52 if (!IsValidCodePointIndex(s, pos)) |
51 ++pos; | 53 ++pos; |
52 return pos; | 54 return pos; |
53 } | 55 } |
54 | 56 |
55 } // namespace gfx | 57 } // namespace gfx |
OLD | NEW |