1 /* 2 * Copyright (C) 2014 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 * See the License for the specific language governing permissions and 14 * limitations under the License. 15 */ 16 17 #include "minikin/GraphemeBreak.h" 18 19 #include <algorithm> 20 #include <cstdint> 21 22 #include <unicode/uchar.h> 23 #include <unicode/utf16.h> 24 25 #include "minikin/Emoji.h" 26 27 namespace minikin { 28 29 int32_t tailoredGraphemeClusterBreak(uint32_t c) { 30 // Characters defined as Control that we want to treat them as Extend. 31 // These are curated manually. 32 if (c == 0x00AD // SHY 33 || c == 0x061C // ALM 34 || c == 0x180E // MONGOLIAN VOWEL SEPARATOR 35 || c == 0x200B // ZWSP 36 || c == 0x200E // LRM 37 || c == 0x200F // RLM 38 || (0x202A <= c && c <= 0x202E) // LRE, RLE, PDF, LRO, RLO 39 || ((c | 0xF) == 0x206F) // WJ, invisible math operators, LRI, RLI, FSI, PDI, 40 // and the deprecated invisible format controls 41 || c == 0xFEFF // BOM 42 || ((c | 0x7F) == 0xE007F)) // recently undeprecated tag characters in Plane 14 43 return U_GCB_EXTEND; 44 // THAI CHARACTER SARA AM is treated as a normal letter by most other implementations: they 45 // allow a grapheme break before it. 46 else if (c == 0x0E33) 47 return U_GCB_OTHER; 48 else 49 return u_getIntPropertyValue(c, UCHAR_GRAPHEME_CLUSTER_BREAK); 50 } 51 52 // Returns true for all characters whose IndicSyllabicCategory is Pure_Killer. 53 // From http://www.unicode.org/Public/9.0.0/ucd/IndicSyllabicCategory.txt 54 bool isPureKiller(uint32_t c) { 55 return (c == 0x0E3A || c == 0x0E4E || c == 0x0F84 || c == 0x103A || c == 0x1714 || 56 c == 0x1734 || c == 0x17D1 || c == 0x1BAA || c == 0x1BF2 || c == 0x1BF3 || 57 c == 0xA806 || c == 0xA953 || c == 0xABED || c == 0x11134 || c == 0x112EA || 58 c == 0x1172B); 59 } 60 61 bool GraphemeBreak::isGraphemeBreak(const float* advances, const uint16_t* buf, size_t start, 62 size_t count, const size_t offset) { 63 // This implementation closely follows Unicode Standard Annex #29 on 64 // Unicode Text Segmentation (http://www.unicode.org/reports/tr29/), 65 // implementing a tailored version of extended grapheme clusters. 66 // The GB rules refer to section 3.1.1, Grapheme Cluster Boundary Rules. 67 68 // Rule GB1, sot ; Rule GB2, eot 69 if (offset <= start || offset >= start + count) { 70 return true; 71 } 72 if (U16_IS_TRAIL(buf[offset])) { 73 // Don't break a surrogate pair, but a lonely trailing surrogate pair is a break 74 return !U16_IS_LEAD(buf[offset - 1]); 75 } 76 uint32_t c1 = 0; 77 uint32_t c2 = 0; 78 size_t offset_back = offset; 79 size_t offset_forward = offset; 80 U16_PREV(buf, start, offset_back, c1); 81 U16_NEXT(buf, offset_forward, start + count, c2); 82 int32_t p1 = tailoredGraphemeClusterBreak(c1); 83 int32_t p2 = tailoredGraphemeClusterBreak(c2); 84 // Rule GB3, CR x LF 85 if (p1 == U_GCB_CR && p2 == U_GCB_LF) { 86 return false; 87 } 88 // Rule GB4, (Control | CR | LF) 89 if (p1 == U_GCB_CONTROL || p1 == U_GCB_CR || p1 == U_GCB_LF) { 90 return true; 91 } 92 // Rule GB5, (Control | CR | LF) 93 if (p2 == U_GCB_CONTROL || p2 == U_GCB_CR || p2 == U_GCB_LF) { 94 return true; 95 } 96 // Rule GB6, L x ( L | V | LV | LVT ) 97 if (p1 == U_GCB_L && (p2 == U_GCB_L || p2 == U_GCB_V || p2 == U_GCB_LV || p2 == U_GCB_LVT)) { 98 return false; 99 } 100 // Rule GB7, ( LV | V ) x ( V | T ) 101 if ((p1 == U_GCB_LV || p1 == U_GCB_V) && (p2 == U_GCB_V || p2 == U_GCB_T)) { 102 return false; 103 } 104 // Rule GB8, ( LVT | T ) x T 105 if ((p1 == U_GCB_LVT || p1 == U_GCB_T) && p2 == U_GCB_T) { 106 return false; 107 } 108 // Rule GB9, x (Extend | ZWJ); Rule GB9a, x SpacingMark; Rule GB9b, Prepend x 109 if (p2 == U_GCB_EXTEND || p2 == U_GCB_ZWJ || p2 == U_GCB_SPACING_MARK || p1 == U_GCB_PREPEND) { 110 return false; 111 } 112 113 // This is used to decide font-dependent grapheme clusters. If we don't have the advance 114 // information, we become conservative in grapheme breaking and assume that it has no advance. 115 const bool c2_has_advance = (advances != nullptr && advances[offset - start] != 0.0); 116 117 // All the following rules are font-dependent, in the way that if we know c2 has an advance, 118 // we definitely know that it cannot form a grapheme with the character(s) before it. So we 119 // make the decision in favor a grapheme break early. 120 if (c2_has_advance) { 121 return true; 122 } 123 124 // Note: For Rule GB10 and GB11 below, we do not use the Unicode line breaking properties for 125 // determining emoji-ness and carry our own data, because our data could be more fresh than what 126 // ICU provides. 127 // 128 // Tailored version of Rule GB10, (E_Base | EBG) Extend* E_Modifier. 129 // The rule itself says do not break between emoji base and emoji modifiers, skipping all Extend 130 // characters. Variation selectors are considered Extend, so they are handled fine. 131 // 132 // We tailor this by requiring that an actual ligature is formed. If the font doesn't form a 133 // ligature, we allow a break before the modifier. 134 if (isEmojiModifier(c2)) { 135 uint32_t c0 = c1; 136 size_t offset_backback = offset_back; 137 int32_t p0 = p1; 138 if (p0 == U_GCB_EXTEND && offset_backback > start) { 139 // skip over emoji variation selector 140 U16_PREV(buf, start, offset_backback, c0); 141 p0 = tailoredGraphemeClusterBreak(c0); 142 } 143 if (isEmojiBase(c0)) { 144 return false; 145 } 146 } 147 // Tailored version of Rule GB11, ZWJ (Glue_After_Zwj | EBG) 148 // We try to make emoji sequences with ZWJ a single grapheme cluster, but only if they actually 149 // merge to one cluster. So we are more relaxed than the UAX #29 rules in accepting any emoji 150 // character after the ZWJ, but are tighter in that we only treat it as one cluster if a 151 // ligature is actually formed and we also require the character before the ZWJ to also be an 152 // emoji. 153 if (p1 == U_GCB_ZWJ && isEmoji(c2) && offset_back > start) { 154 // look at character before ZWJ to see that both can participate in an emoji zwj sequence 155 uint32_t c0 = 0; 156 size_t offset_backback = offset_back; 157 U16_PREV(buf, start, offset_backback, c0); 158 if (c0 == 0xFE0F && offset_backback > start) { 159 // skip over emoji variation selector 160 U16_PREV(buf, start, offset_backback, c0); 161 } 162 if (isEmoji(c0)) { 163 return false; 164 } 165 } 166 // Tailored version of Rule GB12 and Rule GB13 that look at even-odd cases. 167 // sot (RI RI)* RI x RI 168 // [^RI] (RI RI)* RI x RI 169 // 170 // If we have font information, we have already broken the cluster if and only if the second 171 // character had no advance, which means a ligature was formed. If we don't, we look back like 172 // UAX #29 recommends, but only up to 1000 code units. 173 if (p1 == U_GCB_REGIONAL_INDICATOR && p2 == U_GCB_REGIONAL_INDICATOR) { 174 if (advances != nullptr) { 175 // We have advances information. But if we are here, we already know c2 has no advance. 176 // So we should definitely disallow a break. 177 return false; 178 } else { 179 // Look at up to 1000 code units. 180 const size_t lookback_barrier = std::max((ssize_t)start, (ssize_t)offset_back - 1000); 181 size_t offset_backback = offset_back; 182 while (offset_backback > lookback_barrier) { 183 uint32_t c0 = 0; 184 U16_PREV(buf, lookback_barrier, offset_backback, c0); 185 if (tailoredGraphemeClusterBreak(c0) != U_GCB_REGIONAL_INDICATOR) { 186 offset_backback += U16_LENGTH(c0); 187 break; 188 } 189 } 190 // The number 4 comes from the number of code units in a whole flag. 191 return (offset - offset_backback) % 4 == 0; 192 } 193 } 194 // Cluster Indic syllables together (tailoring of UAX #29). 195 // Immediately after each virama (that is not just a pure killer) followed by a letter, we 196 // disallow grapheme breaks (if we are here, we don't know about advances, or we already know 197 // that c2 has no advance). 198 if (u_getIntPropertyValue(c1, UCHAR_CANONICAL_COMBINING_CLASS) == 9 // virama 199 && !isPureKiller(c1) && 200 u_getIntPropertyValue(c2, UCHAR_GENERAL_CATEGORY) == U_OTHER_LETTER) { 201 return false; 202 } 203 // Rule GB999, Any Any 204 return true; 205 } 206 207 size_t GraphemeBreak::getTextRunCursor(const float* advances, const uint16_t* buf, size_t start, 208 size_t count, size_t offset, MoveOpt opt) { 209 switch (opt) { 210 case AFTER: 211 if (offset < start + count) { 212 offset++; 213 } 214 // fall through 215 case AT_OR_AFTER: 216 while (!isGraphemeBreak(advances, buf, start, count, offset)) { 217 offset++; 218 } 219 break; 220 case BEFORE: 221 if (offset > start) { 222 offset--; 223 } 224 // fall through 225 case AT_OR_BEFORE: 226 while (!isGraphemeBreak(advances, buf, start, count, offset)) { 227 offset--; 228 } 229 break; 230 case AT: 231 if (!isGraphemeBreak(advances, buf, start, count, offset)) { 232 offset = (size_t)-1; 233 } 234 break; 235 } 236 return offset; 237 } 238 239 } // namespace minikin 240
