1 /*
2 * Copyright (C) 2003, 2006, 2008, 2009, 2010, 2011 Apple Inc. All rights reserved.
3 * Copyright (C) 2008 Holger Hans Peter Freyther
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 #include "config.h"
23 #include "WidthIterator.h"
24
25 #include "Font.h"
26 #include "GlyphBuffer.h"
27 #include "SimpleFontData.h"
28 #include "TextRun.h"
29 #include <wtf/MathExtras.h>
30
31 #if USE(ICU_UNICODE)
32 #include <unicode/unorm.h>
33 #endif
34
35 using namespace WTF;
36 using namespace Unicode;
37 using namespace std;
38
39 namespace WebCore {
40
41 // According to http://www.unicode.org/Public/UNIDATA/UCD.html#Canonical_Combining_Class_Values
42 static const uint8_t hiraganaKatakanaVoicingMarksCombiningClass = 8;
43
WidthIterator(const Font * font,const TextRun & run,HashSet<const SimpleFontData * > * fallbackFonts,bool accountForGlyphBounds,bool forTextEmphasis)44 WidthIterator::WidthIterator(const Font* font, const TextRun& run, HashSet<const SimpleFontData*>* fallbackFonts, bool accountForGlyphBounds, bool forTextEmphasis)
45 : m_font(font)
46 , m_run(run)
47 , m_end(run.length())
48 , m_currentCharacter(0)
49 , m_runWidthSoFar(0)
50 , m_isAfterExpansion(!run.allowsLeadingExpansion())
51 , m_fallbackFonts(fallbackFonts)
52 , m_accountForGlyphBounds(accountForGlyphBounds)
53 , m_maxGlyphBoundingBoxY(numeric_limits<float>::min())
54 , m_minGlyphBoundingBoxY(numeric_limits<float>::max())
55 , m_firstGlyphOverflow(0)
56 , m_lastGlyphOverflow(0)
57 , m_forTextEmphasis(forTextEmphasis)
58 {
59 // If the padding is non-zero, count the number of spaces in the run
60 // and divide that by the padding for per space addition.
61 m_expansion = m_run.expansion();
62 if (!m_expansion)
63 m_expansionPerOpportunity = 0;
64 else {
65 bool isAfterExpansion = m_isAfterExpansion;
66 unsigned expansionOpportunityCount = Font::expansionOpportunityCount(m_run.characters(), m_end, m_run.ltr() ? LTR : RTL, isAfterExpansion);
67 if (isAfterExpansion && !m_run.allowsTrailingExpansion())
68 expansionOpportunityCount--;
69
70 if (!expansionOpportunityCount)
71 m_expansionPerOpportunity = 0;
72 else
73 m_expansionPerOpportunity = m_expansion / expansionOpportunityCount;
74 }
75 }
76
advance(int offset,GlyphBuffer * glyphBuffer)77 void WidthIterator::advance(int offset, GlyphBuffer* glyphBuffer)
78 {
79 if (offset > m_end)
80 offset = m_end;
81
82 int currentCharacter = m_currentCharacter;
83 const UChar* cp = m_run.data(currentCharacter);
84
85 bool rtl = m_run.rtl();
86 bool hasExtraSpacing = (m_font->letterSpacing() || m_font->wordSpacing() || m_expansion) && !m_run.spacingDisabled();
87
88 FloatRect bounds;
89
90 const SimpleFontData* primaryFont = m_font->primaryFont();
91 const SimpleFontData* lastFontData = primaryFont;
92
93 while (currentCharacter < offset) {
94 UChar32 c = *cp;
95 unsigned clusterLength = 1;
96 if (c >= 0x3041) {
97 if (c <= 0x30FE) {
98 // Deal with Hiragana and Katakana voiced and semi-voiced syllables.
99 // Normalize into composed form, and then look for glyph with base + combined mark.
100 // Check above for character range to minimize performance impact.
101 UChar32 normalized = normalizeVoicingMarks(currentCharacter);
102 if (normalized) {
103 c = normalized;
104 clusterLength = 2;
105 }
106 } else if (U16_IS_SURROGATE(c)) {
107 if (!U16_IS_SURROGATE_LEAD(c))
108 break;
109
110 // Do we have a surrogate pair? If so, determine the full Unicode (32 bit)
111 // code point before glyph lookup.
112 // Make sure we have another character and it's a low surrogate.
113 if (currentCharacter + 1 >= m_run.length())
114 break;
115 UChar low = cp[1];
116 if (!U16_IS_TRAIL(low))
117 break;
118 c = U16_GET_SUPPLEMENTARY(c, low);
119 clusterLength = 2;
120 }
121 }
122
123 const GlyphData& glyphData = m_font->glyphDataForCharacter(c, rtl);
124 Glyph glyph = glyphData.glyph;
125 const SimpleFontData* fontData = glyphData.fontData;
126
127 ASSERT(fontData);
128
129 // Now that we have a glyph and font data, get its width.
130 float width;
131 if (c == '\t' && m_run.allowTabs()) {
132 float tabWidth = m_font->tabWidth(*fontData);
133 width = tabWidth - fmodf(m_run.xPos() + m_runWidthSoFar, tabWidth);
134 } else {
135 width = fontData->widthForGlyph(glyph);
136
137 #if ENABLE(SVG)
138 // SVG uses horizontalGlyphStretch(), when textLength is used to stretch/squeeze text.
139 width *= m_run.horizontalGlyphStretch();
140 #endif
141 }
142
143 if (fontData != lastFontData && width) {
144 lastFontData = fontData;
145 if (m_fallbackFonts && fontData != primaryFont) {
146 // FIXME: This does a little extra work that could be avoided if
147 // glyphDataForCharacter() returned whether it chose to use a small caps font.
148 if (!m_font->isSmallCaps() || c == toUpper(c))
149 m_fallbackFonts->add(fontData);
150 else {
151 const GlyphData& uppercaseGlyphData = m_font->glyphDataForCharacter(toUpper(c), rtl);
152 if (uppercaseGlyphData.fontData != primaryFont)
153 m_fallbackFonts->add(uppercaseGlyphData.fontData);
154 }
155 }
156 }
157
158 if (hasExtraSpacing) {
159 // Account for letter-spacing.
160 if (width && m_font->letterSpacing())
161 width += m_font->letterSpacing();
162
163 static bool expandAroundIdeographs = Font::canExpandAroundIdeographsInComplexText();
164 bool treatAsSpace = Font::treatAsSpace(c);
165 if (treatAsSpace || (expandAroundIdeographs && Font::isCJKIdeographOrSymbol(c))) {
166 // Distribute the run's total expansion evenly over all expansion opportunities in the run.
167 if (m_expansion) {
168 if (!treatAsSpace && !m_isAfterExpansion) {
169 // Take the expansion opportunity before this ideograph.
170 m_expansion -= m_expansionPerOpportunity;
171 m_runWidthSoFar += m_expansionPerOpportunity;
172 if (glyphBuffer) {
173 if (glyphBuffer->isEmpty())
174 glyphBuffer->add(fontData->spaceGlyph(), fontData, m_expansionPerOpportunity);
175 else
176 glyphBuffer->expandLastAdvance(m_expansionPerOpportunity);
177 }
178 }
179 if (m_run.allowsTrailingExpansion() || (m_run.ltr() && currentCharacter + clusterLength < static_cast<size_t>(m_run.length()))
180 || (m_run.rtl() && currentCharacter)) {
181 m_expansion -= m_expansionPerOpportunity;
182 width += m_expansionPerOpportunity;
183 m_isAfterExpansion = true;
184 }
185 } else
186 m_isAfterExpansion = false;
187
188 // Account for word spacing.
189 // We apply additional space between "words" by adding width to the space character.
190 if (treatAsSpace && currentCharacter && !Font::treatAsSpace(cp[-1]) && m_font->wordSpacing())
191 width += m_font->wordSpacing();
192 } else
193 m_isAfterExpansion = false;
194 }
195
196 if (m_accountForGlyphBounds) {
197 bounds = fontData->boundsForGlyph(glyph);
198 if (!currentCharacter)
199 m_firstGlyphOverflow = max<float>(0, -bounds.x());
200 }
201
202 if (m_forTextEmphasis && !Font::canReceiveTextEmphasis(c))
203 glyph = 0;
204
205 // Advance past the character we just dealt with.
206 cp += clusterLength;
207 currentCharacter += clusterLength;
208
209 m_runWidthSoFar += width;
210
211 if (glyphBuffer)
212 glyphBuffer->add(glyph, fontData, width);
213
214 if (m_accountForGlyphBounds) {
215 m_maxGlyphBoundingBoxY = max(m_maxGlyphBoundingBoxY, bounds.maxY());
216 m_minGlyphBoundingBoxY = min(m_minGlyphBoundingBoxY, bounds.y());
217 m_lastGlyphOverflow = max<float>(0, bounds.maxX() - width);
218 }
219 }
220
221 m_currentCharacter = currentCharacter;
222 }
223
advanceOneCharacter(float & width,GlyphBuffer * glyphBuffer)224 bool WidthIterator::advanceOneCharacter(float& width, GlyphBuffer* glyphBuffer)
225 {
226 int oldSize = glyphBuffer->size();
227 advance(m_currentCharacter + 1, glyphBuffer);
228 float w = 0;
229 for (int i = oldSize; i < glyphBuffer->size(); ++i)
230 w += glyphBuffer->advanceAt(i);
231 width = w;
232 return glyphBuffer->size() > oldSize;
233 }
234
normalizeVoicingMarks(int currentCharacter)235 UChar32 WidthIterator::normalizeVoicingMarks(int currentCharacter)
236 {
237 if (currentCharacter + 1 < m_end) {
238 if (combiningClass(m_run[currentCharacter + 1]) == hiraganaKatakanaVoicingMarksCombiningClass) {
239 #if USE(ICU_UNICODE)
240 // Normalize into composed form using 3.2 rules.
241 UChar normalizedCharacters[2] = { 0, 0 };
242 UErrorCode uStatus = U_ZERO_ERROR;
243 int32_t resultLength = unorm_normalize(m_run.data(currentCharacter), 2,
244 UNORM_NFC, UNORM_UNICODE_3_2, &normalizedCharacters[0], 2, &uStatus);
245 if (resultLength == 1 && uStatus == 0)
246 return normalizedCharacters[0];
247 #elif USE(QT4_UNICODE)
248 QString tmp(reinterpret_cast<const QChar*>(m_run.data(currentCharacter)), 2);
249 QString res = tmp.normalized(QString::NormalizationForm_C, QChar::Unicode_3_2);
250 if (res.length() == 1)
251 return res.at(0).unicode();
252 #endif
253 }
254 }
255 return 0;
256 }
257
258 }
259