bump product version to 6.4.0.3
[LibreOffice.git] / vcl / source / gdi / scrptrun.cxx
blobb18816cac26c0b3811e4afe76d40b09281d82204
1 /*
2 *******************************************************************************
4 * Copyright (c) 1995-2013 International Business Machines Corporation and others
6 * All rights reserved.
8 * Permission is hereby granted, free of charge, to any person obtaining a copy of
9 * this software and associated documentation files (the "Software"), to deal in
10 * the Software without restriction, including without limitation the rights to
11 * use, copy, modify, merge, publish, distribute, and/or sell copies of the
12 * Software, and to permit persons to whom the Software is furnished to do so,
13 * provided that the above copyright notice(s) and this permission notice appear
14 * in all copies of the Software and that both the above copyright notice(s) and
15 * this permission notice appear in supporting documentation.
17 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
18 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
19 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OF THIRD PARTY RIGHTS. IN
20 * NO EVENT SHALL THE COPYRIGHT HOLDER OR HOLDERS INCLUDED IN THIS NOTICE BE
21 * LIABLE FOR ANY CLAIM, OR ANY SPECIAL INDIRECT OR CONSEQUENTIAL DAMAGES, OR ANY
22 * DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
23 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
24 * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
26 * Except as contained in this notice, the name of a copyright holder shall not be
27 * used in advertising or otherwise to promote the sale, use or other dealings in
28 * this Software without prior written authorization of the copyright holder.
30 *******************************************************************************
31 * file name: scrptrun.cpp
33 * created on: 10/17/2001
34 * created by: Eric R. Mader
36 /**
37 * This file is largely copied from the ICU project,
38 * under folder source/extra/scrptrun/scrptrun.cpp
40 #include <unicode/utypes.h>
41 #include <unicode/uscript.h>
43 #include <scrptrun.h>
44 #include <algorithm>
46 namespace {
48 struct PairIndices
50 int8_t ma00[0xff];
51 int8_t ma20[0x7f];
52 int8_t ma30[0x7f];
54 PairIndices()
56 std::fill_n(ma00, 0xff, -1);
57 std::fill_n(ma20, 0x7f, -1);
58 std::fill_n(ma30, 0x7f, -1);
60 // characters in the range 0x0000 - 0x007e (inclusive)
61 // ascii paired punctuation
62 ma00[0x28] = 0;
63 ma00[0x29] = 1;
64 ma00[0x3c] = 2;
65 ma00[0x3e] = 3;
66 ma00[0x5b] = 4;
67 ma00[0x5d] = 5;
68 ma00[0x7b] = 6;
69 ma00[0x7d] = 7;
70 // guillemets
71 ma00[0xab] = 8;
72 ma00[0xbb] = 9;
74 // characters in the range 0x2000 - 0x207e (inclusive)
75 // general punctuation
76 ma20[0x18] = 10;
77 ma20[0x19] = 11;
78 ma20[0x1c] = 12;
79 ma20[0x1d] = 13;
80 ma20[0x39] = 14;
81 ma20[0x3a] = 15;
83 // characters in the range 0x3000 - 0x307e (inclusive)
84 // chinese paired punctuation
85 ma30[0x08] = 16;
86 ma30[0x09] = 17;
87 ma30[0x0a] = 18;
88 ma30[0x0b] = 19;
89 ma30[0x0c] = 20;
90 ma30[0x0d] = 21;
91 ma30[0x0e] = 22;
92 ma30[0x0f] = 23;
93 ma30[0x10] = 24;
94 ma30[0x11] = 25;
95 ma30[0x14] = 26;
96 ma30[0x15] = 27;
97 ma30[0x16] = 28;
98 ma30[0x17] = 29;
99 ma30[0x18] = 30;
100 ma30[0x19] = 31;
101 ma30[0x1a] = 32;
102 ma30[0x1b] = 33;
105 int32_t getPairIndex(UChar32 ch) const
107 if (ch < 0xff)
108 return ma00[ch];
109 if (ch >= 0x2000 && ch < 0x207f)
110 return ma20[ch - 0x2000];
111 if (ch >= 0x3000 && ch < 0x307f)
112 return ma30[ch - 0x3000];
113 return -1;
118 // There are three Unicode script codes for Japanese text, but only one
119 // OpenType script tag, so we want to keep them in one run as splitting is
120 // pointless for the purpose of OpenType shaping.
121 UScriptCode getScript(UChar32 ch, UErrorCode* status)
123 UScriptCode script = uscript_getScript(ch, status);
124 if (U_FAILURE(*status))
125 return script;
126 if (script == USCRIPT_KATAKANA || script == USCRIPT_KATAKANA_OR_HIRAGANA)
127 return USCRIPT_HIRAGANA;
128 return script;
133 static const PairIndices gPairIndices;
136 namespace vcl {
138 const char ScriptRun::fgClassID=0;
140 static bool sameScript(int32_t scriptOne, int32_t scriptTwo)
142 return scriptOne <= USCRIPT_INHERITED || scriptTwo <= USCRIPT_INHERITED || scriptOne == scriptTwo;
145 UBool ScriptRun::next()
147 int32_t startSP = parenSP; // used to find the first new open character
148 UErrorCode error = U_ZERO_ERROR;
150 // if we've fallen off the end of the text, we're done
151 if (scriptEnd >= charLimit) {
152 return false;
155 scriptCode = USCRIPT_COMMON;
157 for (scriptStart = scriptEnd; scriptEnd < charLimit; scriptEnd += 1) {
158 UChar high = charArray[scriptEnd];
159 UChar32 ch = high;
161 // if the character is a high surrogate and it's not the last one
162 // in the text, see if it's followed by a low surrogate
163 if (high >= 0xD800 && high <= 0xDBFF && scriptEnd < charLimit - 1)
165 UChar low = charArray[scriptEnd + 1];
167 // if it is followed by a low surrogate,
168 // consume it and form the full character
169 if (low >= 0xDC00 && low <= 0xDFFF) {
170 ch = (high - 0xD800) * 0x0400 + low - 0xDC00 + 0x10000;
171 scriptEnd += 1;
175 UScriptCode sc = getScript(ch, &error);
176 int32_t pairIndex = gPairIndices.getPairIndex(ch);
178 // Paired character handling:
180 // if it's an open character, push it onto the stack.
181 // if it's a close character, find the matching open on the
182 // stack, and use that script code. Any non-matching open
183 // characters above it on the stack will be popped.
184 if (pairIndex >= 0) {
185 if ((pairIndex & 1) == 0) {
186 ++parenSP;
187 int32_t nVecSize = parenStack.size();
188 if (parenSP == nVecSize)
189 parenStack.resize(nVecSize + 128);
190 parenStack[parenSP].pairIndex = pairIndex;
191 parenStack[parenSP].scriptCode = scriptCode;
192 } else if (parenSP >= 0) {
193 int32_t pi = pairIndex & ~1;
195 while (parenSP >= 0 && parenStack[parenSP].pairIndex != pi) {
196 parenSP -= 1;
199 if (parenSP < startSP) {
200 startSP = parenSP;
203 if (parenSP >= 0) {
204 sc = parenStack[parenSP].scriptCode;
209 if (sameScript(scriptCode, sc)) {
210 if (scriptCode <= USCRIPT_INHERITED && sc > USCRIPT_INHERITED) {
211 scriptCode = sc;
213 // now that we have a final script code, fix any open
214 // characters we pushed before we knew the script code.
215 while (startSP < parenSP) {
216 parenStack[++startSP].scriptCode = scriptCode;
220 // if this character is a close paired character,
221 // pop it from the stack
222 if (pairIndex >= 0 && (pairIndex & 1) != 0 && parenSP >= 0) {
223 parenSP -= 1;
224 /* decrement startSP only if it is >= 0,
225 decrementing it unnecessarily will lead to memory corruption
226 while processing the above while block.
227 e.g. startSP = -4 , parenSP = -1
229 if (startSP >= 0) {
230 startSP -= 1;
233 } else {
234 // if the run broke on a surrogate pair,
235 // end it before the high surrogate
236 if (ch >= 0x10000) {
237 scriptEnd -= 1;
240 break;
244 return true;