1 | /* |
2 | * Copyright (C) 2005, 2007, 2010, 2013, 2016 Apple Inc. All rights reserved. |
3 | * Copyright (C) 2011 Google Inc. All rights reserved. |
4 | * |
5 | * This library is free software; you can redistribute it and/or |
6 | * modify it under the terms of the GNU Library General Public |
7 | * License as published by the Free Software Foundation; either |
8 | * version 2 of the License, or (at your option) any later version. |
9 | * |
10 | * This library is distributed in the hope that it will be useful, |
11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
13 | * Library General Public License for more details. |
14 | * |
15 | * You should have received a copy of the GNU Library General Public License |
16 | * along with this library; see the file COPYING.LIB. If not, write to |
17 | * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, |
18 | * Boston, MA 02110-1301, USA. |
19 | * |
20 | */ |
21 | |
22 | #pragma once |
23 | |
24 | #include <unicode/ubrk.h> |
25 | #include <wtf/ASCIICType.h> |
26 | #include <wtf/StdLibExtras.h> |
27 | #include <wtf/text/TextBreakIterator.h> |
28 | #include <wtf/unicode/CharacterNames.h> |
29 | |
30 | namespace WebCore { |
31 | |
32 | static const UChar lineBreakTableFirstCharacter = '!'; |
33 | static const UChar lineBreakTableLastCharacter = 127; |
34 | static const unsigned lineBreakTableColumnCount = (lineBreakTableLastCharacter - lineBreakTableFirstCharacter) / 8 + 1; |
35 | |
36 | WEBCORE_EXPORT extern const unsigned char lineBreakTable[][lineBreakTableColumnCount]; |
37 | |
38 | enum class NonBreakingSpaceBehavior { |
39 | IgnoreNonBreakingSpace, |
40 | TreatNonBreakingSpaceAsBreak, |
41 | }; |
42 | |
43 | enum class CanUseShortcut { |
44 | Yes, |
45 | No |
46 | }; |
47 | |
48 | template<NonBreakingSpaceBehavior nonBreakingSpaceBehavior> |
49 | static inline bool isBreakableSpace(UChar character) |
50 | { |
51 | switch (character) { |
52 | case ' ': |
53 | case '\n': |
54 | case '\t': |
55 | return true; |
56 | case noBreakSpace: |
57 | return nonBreakingSpaceBehavior == NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak; |
58 | default: |
59 | return false; |
60 | } |
61 | } |
62 | |
63 | inline bool shouldBreakAfter(UChar lastCharacter, UChar character, UChar nextCharacter) |
64 | { |
65 | // Don't allow line breaking between '-' and a digit if the '-' may mean a minus sign in the context, |
66 | // while allow breaking in 'ABCD-1234' and '1234-5678' which may be in long URLs. |
67 | if (character == '-' && isASCIIDigit(nextCharacter)) |
68 | return isASCIIAlphanumeric(lastCharacter); |
69 | |
70 | // If both ch and nextCh are ASCII characters, use a lookup table for enhanced speed and for compatibility |
71 | // with other browsers (see comments for asciiLineBreakTable for details). |
72 | if (character >= lineBreakTableFirstCharacter && character <= lineBreakTableLastCharacter && nextCharacter >= lineBreakTableFirstCharacter && nextCharacter <= lineBreakTableLastCharacter) { |
73 | const unsigned char* tableRow = lineBreakTable[character - lineBreakTableFirstCharacter]; |
74 | unsigned nextCharacterIndex = nextCharacter - lineBreakTableFirstCharacter; |
75 | return tableRow[nextCharacterIndex / 8] & (1 << (nextCharacterIndex % 8)); |
76 | } |
77 | // Otherwise defer to the Unicode algorithm by returning false. |
78 | return false; |
79 | } |
80 | |
81 | template<NonBreakingSpaceBehavior nonBreakingSpaceBehavior> |
82 | inline bool needsLineBreakIterator(UChar character) |
83 | { |
84 | if (nonBreakingSpaceBehavior == NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak) |
85 | return character > lineBreakTableLastCharacter; |
86 | return character > lineBreakTableLastCharacter && character != noBreakSpace; |
87 | } |
88 | |
89 | // When in non-loose mode, we can use the ASCII shortcut table. |
90 | template<typename CharacterType, NonBreakingSpaceBehavior nonBreakingSpaceBehavior, CanUseShortcut canUseShortcut> |
91 | inline unsigned nextBreakablePosition(LazyLineBreakIterator& lazyBreakIterator, const CharacterType* string, unsigned length, unsigned startPosition) |
92 | { |
93 | Optional<unsigned> nextBreak; |
94 | |
95 | CharacterType lastLastCharacter = startPosition > 1 ? string[startPosition - 2] : static_cast<CharacterType>(lazyBreakIterator.secondToLastCharacter()); |
96 | CharacterType lastCharacter = startPosition > 0 ? string[startPosition - 1] : static_cast<CharacterType>(lazyBreakIterator.lastCharacter()); |
97 | unsigned priorContextLength = lazyBreakIterator.priorContextLength(); |
98 | for (unsigned i = startPosition; i < length; i++) { |
99 | CharacterType character = string[i]; |
100 | |
101 | if (isBreakableSpace<nonBreakingSpaceBehavior>(character) || (canUseShortcut == CanUseShortcut::Yes && shouldBreakAfter(lastLastCharacter, lastCharacter, character))) |
102 | return i; |
103 | |
104 | if (canUseShortcut == CanUseShortcut::No || needsLineBreakIterator<nonBreakingSpaceBehavior>(character) || needsLineBreakIterator<nonBreakingSpaceBehavior>(lastCharacter)) { |
105 | if (!nextBreak || nextBreak.value() < i) { |
106 | // Don't break if positioned at start of primary context and there is no prior context. |
107 | if (i || priorContextLength) { |
108 | UBreakIterator* breakIterator = lazyBreakIterator.get(priorContextLength); |
109 | if (breakIterator) { |
110 | int candidate = ubrk_following(breakIterator, i - 1 + priorContextLength); |
111 | if (candidate == UBRK_DONE) |
112 | nextBreak = WTF::nullopt; |
113 | else { |
114 | unsigned result = candidate; |
115 | ASSERT(result >= priorContextLength); |
116 | nextBreak = result - priorContextLength; |
117 | } |
118 | } |
119 | } |
120 | } |
121 | if (i == nextBreak && !isBreakableSpace<nonBreakingSpaceBehavior>(lastCharacter)) |
122 | return i; |
123 | } |
124 | |
125 | lastLastCharacter = lastCharacter; |
126 | lastCharacter = character; |
127 | } |
128 | |
129 | return length; |
130 | } |
131 | |
132 | template<typename CharacterType, NonBreakingSpaceBehavior nonBreakingSpaceBehavior> |
133 | inline unsigned nextBreakablePositionKeepingAllWords(const CharacterType* string, unsigned length, unsigned startPosition) |
134 | { |
135 | for (unsigned i = startPosition; i < length; i++) { |
136 | if (isBreakableSpace<nonBreakingSpaceBehavior>(string[i])) |
137 | return i; |
138 | } |
139 | return length; |
140 | } |
141 | |
142 | inline unsigned nextBreakablePositionKeepingAllWords(LazyLineBreakIterator& lazyBreakIterator, unsigned startPosition) |
143 | { |
144 | auto stringView = lazyBreakIterator.stringView(); |
145 | if (stringView.is8Bit()) |
146 | return nextBreakablePositionKeepingAllWords<LChar, NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak>(stringView.characters8(), stringView.length(), startPosition); |
147 | return nextBreakablePositionKeepingAllWords<UChar, NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak>(stringView.characters16(), stringView.length(), startPosition); |
148 | } |
149 | |
150 | inline unsigned nextBreakablePositionKeepingAllWordsIgnoringNBSP(LazyLineBreakIterator& iterator, unsigned startPosition) |
151 | { |
152 | auto stringView = iterator.stringView(); |
153 | if (stringView.is8Bit()) |
154 | return nextBreakablePositionKeepingAllWords<LChar, NonBreakingSpaceBehavior::IgnoreNonBreakingSpace>(stringView.characters8(), stringView.length(), startPosition); |
155 | return nextBreakablePositionKeepingAllWords<UChar, NonBreakingSpaceBehavior::IgnoreNonBreakingSpace>(stringView.characters16(), stringView.length(), startPosition); |
156 | } |
157 | |
158 | inline unsigned nextBreakablePosition(LazyLineBreakIterator& iterator, unsigned startPosition) |
159 | { |
160 | auto stringView = iterator.stringView(); |
161 | if (stringView.is8Bit()) |
162 | return nextBreakablePosition<LChar, NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak, CanUseShortcut::Yes>(iterator, stringView.characters8(), stringView.length(), startPosition); |
163 | return nextBreakablePosition<UChar, NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak, CanUseShortcut::Yes>(iterator, stringView.characters16(), stringView.length(), startPosition); |
164 | } |
165 | |
166 | inline unsigned nextBreakablePositionIgnoringNBSP(LazyLineBreakIterator& lazyBreakIterator, unsigned startPosition) |
167 | { |
168 | auto stringView = lazyBreakIterator.stringView(); |
169 | if (stringView.is8Bit()) |
170 | return nextBreakablePosition<LChar, NonBreakingSpaceBehavior::IgnoreNonBreakingSpace, CanUseShortcut::Yes>(lazyBreakIterator, stringView.characters8(), stringView.length(), startPosition); |
171 | return nextBreakablePosition<UChar, NonBreakingSpaceBehavior::IgnoreNonBreakingSpace, CanUseShortcut::Yes>(lazyBreakIterator, stringView.characters16(), stringView.length(), startPosition); |
172 | } |
173 | |
174 | inline unsigned nextBreakablePositionWithoutShortcut(LazyLineBreakIterator& lazyBreakIterator, unsigned startPosition) |
175 | { |
176 | auto stringView = lazyBreakIterator.stringView(); |
177 | if (stringView.is8Bit()) |
178 | return nextBreakablePosition<LChar, NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak, CanUseShortcut::No>(lazyBreakIterator, stringView.characters8(), stringView.length(), startPosition); |
179 | return nextBreakablePosition<UChar, NonBreakingSpaceBehavior::TreatNonBreakingSpaceAsBreak, CanUseShortcut::No>(lazyBreakIterator, stringView.characters16(), stringView.length(), startPosition); |
180 | } |
181 | |
182 | inline unsigned nextBreakablePositionIgnoringNBSPWithoutShortcut(LazyLineBreakIterator& lazyBreakIterator, unsigned startPosition) |
183 | { |
184 | auto stringView = lazyBreakIterator.stringView(); |
185 | if (stringView.is8Bit()) |
186 | return nextBreakablePosition<LChar, NonBreakingSpaceBehavior::IgnoreNonBreakingSpace, CanUseShortcut::No>(lazyBreakIterator, stringView.characters8(), stringView.length(), startPosition); |
187 | return nextBreakablePosition<UChar, NonBreakingSpaceBehavior::IgnoreNonBreakingSpace, CanUseShortcut::No>(lazyBreakIterator, stringView.characters16(), stringView.length(), startPosition); |
188 | } |
189 | |
190 | inline unsigned nextBreakablePositionBreakCharacter(LazyLineBreakIterator& lazyBreakIterator, unsigned startPosition) |
191 | { |
192 | auto stringView = lazyBreakIterator.stringView(); |
193 | ASSERT(startPosition <= stringView.length()); |
194 | // FIXME: Can/Should we implement this using a Shared Iterator (performance issue) |
195 | // https://bugs.webkit.org/show_bug.cgi?id=197876 |
196 | NonSharedCharacterBreakIterator iterator(stringView); |
197 | Optional<unsigned> next = ubrk_following(iterator, startPosition); |
198 | return next.valueOr(stringView.length()); |
199 | } |
200 | |
201 | inline bool isBreakable(LazyLineBreakIterator& lazyBreakIterator, unsigned startPosition, Optional<unsigned>& nextBreakable, bool breakNBSP, bool canUseShortcut, bool keepAllWords, bool breakAnywhere) |
202 | { |
203 | if (nextBreakable && nextBreakable.value() >= startPosition) |
204 | return startPosition == nextBreakable; |
205 | |
206 | if (breakAnywhere) |
207 | nextBreakable = nextBreakablePositionBreakCharacter(lazyBreakIterator, startPosition); |
208 | else if (keepAllWords) { |
209 | if (breakNBSP) |
210 | nextBreakable = nextBreakablePositionKeepingAllWords(lazyBreakIterator, startPosition); |
211 | else |
212 | nextBreakable = nextBreakablePositionKeepingAllWordsIgnoringNBSP(lazyBreakIterator, startPosition); |
213 | } else if (!canUseShortcut) { |
214 | if (breakNBSP) |
215 | nextBreakable = nextBreakablePositionWithoutShortcut(lazyBreakIterator, startPosition); |
216 | else |
217 | nextBreakable = nextBreakablePositionIgnoringNBSPWithoutShortcut(lazyBreakIterator, startPosition); |
218 | } else { |
219 | if (breakNBSP) |
220 | nextBreakable = nextBreakablePosition(lazyBreakIterator, startPosition); |
221 | else |
222 | nextBreakable = nextBreakablePositionIgnoringNBSP(lazyBreakIterator, startPosition); |
223 | } |
224 | return startPosition == nextBreakable; |
225 | } |
226 | |
227 | } // namespace WebCore |
228 | |