author | Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com> |
Mon, 04 Oct 2010 01:32:07 +0300 | |
changeset 2 | 303757a437d3 |
parent 0 | 4f2f89ce4247 |
permissions | -rw-r--r-- |
0
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
1 |
/* |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
2 |
* Copyright (C) 2007 Apple Inc. All rights reserved. |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
3 |
* Copyright (C) 2008 Tony Chang <idealisms@gmail.com> |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
4 |
* |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
5 |
* Redistribution and use in source and binary forms, with or without |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
6 |
* modification, are permitted provided that the following conditions |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
7 |
* are met: |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
8 |
* |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
9 |
* 1. Redistributions of source code must retain the above copyright |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
10 |
* notice, this list of conditions and the following disclaimer. |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
11 |
* 2. Redistributions in binary form must reproduce the above copyright |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
12 |
* notice, this list of conditions and the following disclaimer in the |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
13 |
* documentation and/or other materials provided with the distribution. |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
14 |
* 3. Neither the name of Apple Computer, Inc. ("Apple") nor the names of |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
15 |
* its contributors may be used to endorse or promote products derived |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
16 |
* from this software without specific prior written permission. |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
17 |
* |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
18 |
* THIS SOFTWARE IS PROVIDED BY APPLE AND ITS CONTRIBUTORS "AS IS" AND ANY |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
19 |
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
20 |
* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
21 |
* DISCLAIMED. IN NO EVENT SHALL APPLE OR ITS CONTRIBUTORS BE LIABLE FOR ANY |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
22 |
* DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
23 |
* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
24 |
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
25 |
* ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
26 |
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
27 |
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
28 |
*/ |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
29 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
30 |
#include "config.h" |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
31 |
#include "SmartReplace.h" |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
32 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
33 |
#if !PLATFORM(CF) && USE(ICU_UNICODE) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
34 |
#include "PlatformString.h" |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
35 |
#include <unicode/uset.h> |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
36 |
#include <wtf/Assertions.h> |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
37 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
38 |
namespace WebCore { |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
39 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
40 |
static void addAllCodePoints(USet* smartSet, const String& string) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
41 |
{ |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
42 |
const UChar* characters = string.characters(); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
43 |
for (size_t i = 0; i < string.length(); i++) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
44 |
uset_add(smartSet, characters[i]); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
45 |
} |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
46 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
47 |
// This is mostly a port of the code in WebCore/editing/SmartReplaceCF.cpp |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
48 |
// except we use icu in place of CoreFoundations character classes. |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
49 |
static USet* getSmartSet(bool isPreviousCharacter) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
50 |
{ |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
51 |
static USet* preSmartSet = NULL; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
52 |
static USet* postSmartSet = NULL; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
53 |
USet* smartSet = isPreviousCharacter ? preSmartSet : postSmartSet; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
54 |
if (!smartSet) { |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
55 |
// Whitespace and newline (kCFCharacterSetWhitespaceAndNewline) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
56 |
UErrorCode ec = U_ZERO_ERROR; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
57 |
String whitespaceAndNewline = "[[:WSpace:] [\\u000A\\u000B\\u000C\\u000D\\u0085]]"; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
58 |
smartSet = uset_openPattern(whitespaceAndNewline.characters(), whitespaceAndNewline.length(), &ec); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
59 |
ASSERT(U_SUCCESS(ec)); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
60 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
61 |
// CJK ranges |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
62 |
uset_addRange(smartSet, 0x1100, 0x1100 + 256); // Hangul Jamo (0x1100 - 0x11FF) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
63 |
uset_addRange(smartSet, 0x2E80, 0x2E80 + 352); // CJK & Kangxi Radicals (0x2E80 - 0x2FDF) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
64 |
uset_addRange(smartSet, 0x2FF0, 0x2FF0 + 464); // Ideograph Descriptions, CJK Symbols, Hiragana, Katakana, Bopomofo, Hangul Compatibility Jamo, Kanbun, & Bopomofo Ext (0x2FF0 - 0x31BF) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
65 |
uset_addRange(smartSet, 0x3200, 0x3200 + 29392); // Enclosed CJK, CJK Ideographs (Uni Han & Ext A), & Yi (0x3200 - 0xA4CF) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
66 |
uset_addRange(smartSet, 0xAC00, 0xAC00 + 11183); // Hangul Syllables (0xAC00 - 0xD7AF) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
67 |
uset_addRange(smartSet, 0xF900, 0xF900 + 352); // CJK Compatibility Ideographs (0xF900 - 0xFA5F) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
68 |
uset_addRange(smartSet, 0xFE30, 0xFE30 + 32); // CJK Compatibility From (0xFE30 - 0xFE4F) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
69 |
uset_addRange(smartSet, 0xFF00, 0xFF00 + 240); // Half/Full Width Form (0xFF00 - 0xFFEF) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
70 |
uset_addRange(smartSet, 0x20000, 0x20000 + 0xA6D7); // CJK Ideograph Exntension B |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
71 |
uset_addRange(smartSet, 0x2F800, 0x2F800 + 0x021E); // CJK Compatibility Ideographs (0x2F800 - 0x2FA1D) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
72 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
73 |
if (isPreviousCharacter) { |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
74 |
addAllCodePoints(smartSet, "([\"\'#$/-`{"); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
75 |
preSmartSet = smartSet; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
76 |
} else { |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
77 |
addAllCodePoints(smartSet, ")].,;:?\'!\"%*-/}"); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
78 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
79 |
// Punctuation (kCFCharacterSetPunctuation) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
80 |
UErrorCode ec = U_ZERO_ERROR; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
81 |
String punctuationClass = "[:P:]"; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
82 |
USet* icuPunct = uset_openPattern(punctuationClass.characters(), punctuationClass.length(), &ec); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
83 |
ASSERT(U_SUCCESS(ec)); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
84 |
uset_addAll(smartSet, icuPunct); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
85 |
uset_close(icuPunct); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
86 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
87 |
postSmartSet = smartSet; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
88 |
} |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
89 |
} |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
90 |
return smartSet; |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
91 |
} |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
92 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
93 |
bool isCharacterSmartReplaceExempt(UChar32 c, bool isPreviousCharacter) |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
94 |
{ |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
95 |
return uset_contains(getSmartSet(isPreviousCharacter), c); |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
96 |
} |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
97 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
98 |
} |
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
99 |
|
4f2f89ce4247
Revision: 201037
Dremov Kirill (Nokia-D-MSW/Tampere) <kirill.dremov@nokia.com>
parents:
diff
changeset
|
100 |
#endif // !PLATFORM(CF) && USE(ICU_UNICODE) |