author | srl |
Mon, 25 Jun 2018 11:40:46 -0700 | |
changeset 50840 | 1db5917dfe1c |
parent 50348 | 008f416a79cb |
child 51923 | 16a0f33a5052 |
permissions | -rw-r--r-- |
2 | 1 |
/* |
23010
6dadb192ad81
8029235: Update copyright year to match last edit in jdk8 jdk repository for 2013
lana
parents:
16879
diff
changeset
|
2 |
* Copyright (c) 2003, 2013, Oracle and/or its affiliates. All rights reserved. |
2 | 3 |
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. |
4 |
* |
|
5 |
* This code is free software; you can redistribute it and/or modify it |
|
6 |
* under the terms of the GNU General Public License version 2 only, as |
|
5506 | 7 |
* published by the Free Software Foundation. Oracle designates this |
2 | 8 |
* particular file as subject to the "Classpath" exception as provided |
5506 | 9 |
* by Oracle in the LICENSE file that accompanied this code. |
2 | 10 |
* |
11 |
* This code is distributed in the hope that it will be useful, but WITHOUT |
|
12 |
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
|
13 |
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
|
14 |
* version 2 for more details (a copy is included in the LICENSE file that |
|
15 |
* accompanied this code). |
|
16 |
* |
|
17 |
* You should have received a copy of the GNU General Public License version |
|
18 |
* 2 along with this work; if not, write to the Free Software Foundation, |
|
19 |
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. |
|
20 |
* |
|
5506 | 21 |
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
22 |
* or visit www.oracle.com if you need additional information or have any |
|
23 |
* questions. |
|
2 | 24 |
*/ |
25 |
||
26 |
package sun.font; |
|
27 |
||
28 |
import java.nio.ByteBuffer; |
|
29 |
import java.nio.CharBuffer; |
|
30 |
import java.nio.IntBuffer; |
|
31 |
import java.util.Locale; |
|
32 |
import java.nio.charset.*; |
|
33 |
||
34 |
/* |
|
35 |
* A tt font has a CMAP table which is in turn made up of sub-tables which |
|
36 |
* describe the char to glyph mapping in (possibly) multiple ways. |
|
37 |
* CMAP subtables are described by 3 values. |
|
38 |
* 1. Platform ID (eg 3=Microsoft, which is the id we look for in JDK) |
|
39 |
* 2. Encoding (eg 0=symbol, 1=unicode) |
|
40 |
* 3. TrueType subtable format (how the char->glyph mapping for the encoding |
|
41 |
* is stored in the subtable). See the TrueType spec. Format 4 is required |
|
42 |
* by MS in fonts for windows. Its uses segmented mapping to delta values. |
|
43 |
* Most typically we see are (3,1,4) : |
|
44 |
* CMAP Platform ID=3 is what we use. |
|
45 |
* Encodings that are used in practice by JDK on Solaris are |
|
46 |
* symbol (3,0) |
|
47 |
* unicode (3,1) |
|
48 |
* GBK (3,5) (note that solaris zh fonts report 3,4 but are really 3,5) |
|
49 |
* The format for almost all subtables is 4. However the solaris (3,5) |
|
50 |
* encodings are typically in format 2. |
|
51 |
*/ |
|
52 |
abstract class CMap { |
|
53 |
||
54 |
// static char WingDings_b2c[] = { |
|
55 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
56 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
57 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
58 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
59 |
// 0xfffd, 0xfffd, 0x2702, 0x2701, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
60 |
// 0xfffd, 0x2706, 0x2709, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
61 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
62 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x2707, 0x270d, |
|
63 |
// 0xfffd, 0x270c, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
64 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
65 |
// 0xfffd, 0x2708, 0xfffd, 0xfffd, 0x2744, 0xfffd, 0x271e, 0xfffd, |
|
66 |
// 0x2720, 0x2721, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
67 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
68 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
69 |
// 0xfffd, 0x2751, 0x2752, 0xfffd, 0xfffd, 0x2756, 0xfffd, 0xfffd, |
|
70 |
// 0xfffd, 0xfffd, 0xfffd, 0x2740, 0x273f, 0x275d, 0x275e, 0xfffd, |
|
71 |
// 0xfffd, 0x2780, 0x2781, 0x2782, 0x2783, 0x2784, 0x2785, 0x2786, |
|
72 |
// 0x2787, 0x2788, 0x2789, 0xfffd, 0x278a, 0x278b, 0x278c, 0x278d, |
|
73 |
// 0x278e, 0x278f, 0x2790, 0x2791, 0x2792, 0x2793, 0xfffd, 0xfffd, |
|
74 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
75 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x274d, 0xfffd, |
|
76 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x2736, 0x2734, 0xfffd, 0x2735, |
|
77 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x272a, 0x2730, 0xfffd, |
|
78 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
79 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x27a5, 0xfffd, 0x27a6, 0xfffd, |
|
80 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
81 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
82 |
// 0x27a2, 0xfffd, 0xfffd, 0xfffd, 0x27b3, 0xfffd, 0xfffd, 0xfffd, |
|
83 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
84 |
// 0x27a1, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
85 |
// 0x27a9, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
86 |
// 0xfffd, 0xfffd, 0xfffd, 0x2717, 0x2713, 0xfffd, 0xfffd, 0xfffd, |
|
87 |
// }; |
|
88 |
||
89 |
// static char Symbols_b2c[] = { |
|
90 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
91 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
92 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
93 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
94 |
// 0xfffd, 0xfffd, 0x2200, 0xfffd, 0x2203, 0xfffd, 0xfffd, 0x220d, |
|
95 |
// 0xfffd, 0xfffd, 0x2217, 0xfffd, 0xfffd, 0x2212, 0xfffd, 0xfffd, |
|
96 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
97 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
98 |
// 0x2245, 0x0391, 0x0392, 0x03a7, 0x0394, 0x0395, 0x03a6, 0x0393, |
|
99 |
// 0x0397, 0x0399, 0x03d1, 0x039a, 0x039b, 0x039c, 0x039d, 0x039f, |
|
100 |
// 0x03a0, 0x0398, 0x03a1, 0x03a3, 0x03a4, 0x03a5, 0x03c2, 0x03a9, |
|
101 |
// 0x039e, 0x03a8, 0x0396, 0xfffd, 0x2234, 0xfffd, 0x22a5, 0xfffd, |
|
102 |
// 0xfffd, 0x03b1, 0x03b2, 0x03c7, 0x03b4, 0x03b5, 0x03c6, 0x03b3, |
|
103 |
// 0x03b7, 0x03b9, 0x03d5, 0x03ba, 0x03bb, 0x03bc, 0x03bd, 0x03bf, |
|
104 |
// 0x03c0, 0x03b8, 0x03c1, 0x03c3, 0x03c4, 0x03c5, 0x03d6, 0x03c9, |
|
105 |
// 0x03be, 0x03c8, 0x03b6, 0xfffd, 0xfffd, 0xfffd, 0x223c, 0xfffd, |
|
106 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
107 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
108 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
109 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
110 |
// 0xfffd, 0x03d2, 0xfffd, 0x2264, 0x2215, 0x221e, 0xfffd, 0xfffd, |
|
111 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
112 |
// 0x2218, 0xfffd, 0xfffd, 0x2265, 0xfffd, 0x221d, 0xfffd, 0x2219, |
|
113 |
// 0xfffd, 0x2260, 0x2261, 0x2248, 0x22ef, 0x2223, 0xfffd, 0xfffd, |
|
114 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x2297, 0x2295, 0x2205, 0x2229, |
|
115 |
// 0x222a, 0x2283, 0x2287, 0x2284, 0x2282, 0x2286, 0x2208, 0x2209, |
|
116 |
// 0xfffd, 0x2207, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x221a, 0x22c5, |
|
117 |
// 0xfffd, 0x2227, 0x2228, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
118 |
// 0x22c4, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x2211, 0xfffd, 0xfffd, |
|
119 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
120 |
// 0xfffd, 0xfffd, 0x222b, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
121 |
// 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, |
|
122 |
// }; |
|
123 |
||
124 |
static final short ShiftJISEncoding = 2; |
|
125 |
static final short GBKEncoding = 3; |
|
126 |
static final short Big5Encoding = 4; |
|
127 |
static final short WansungEncoding = 5; |
|
128 |
static final short JohabEncoding = 6; |
|
129 |
static final short MSUnicodeSurrogateEncoding = 10; |
|
130 |
||
131 |
static final char noSuchChar = (char)0xfffd; |
|
132 |
static final int SHORTMASK = 0x0000ffff; |
|
133 |
static final int INTMASK = 0xffffffff; |
|
134 |
||
135 |
static final char[][] converterMaps = new char[7][]; |
|
136 |
||
137 |
/* |
|
138 |
* Unicode->other encoding translation array. A pre-computed look up |
|
139 |
* which can be shared across all fonts using that encoding. |
|
140 |
* Using this saves running character coverters repeatedly. |
|
141 |
*/ |
|
142 |
char[] xlat; |
|
50840 | 143 |
UVS uvs = null; |
2 | 144 |
|
145 |
static CMap initialize(TrueTypeFont font) { |
|
146 |
||
147 |
CMap cmap = null; |
|
148 |
||
149 |
int offset, platformID, encodingID=-1; |
|
150 |
||
151 |
int three0=0, three1=0, three2=0, three3=0, three4=0, three5=0, |
|
152 |
three6=0, three10=0; |
|
50840 | 153 |
int zero5 = 0; // for Unicode Variation Sequences |
2 | 154 |
boolean threeStar = false; |
155 |
||
156 |
ByteBuffer cmapBuffer = font.getTableBuffer(TrueTypeFont.cmapTag); |
|
157 |
int cmapTableOffset = font.getTableSize(TrueTypeFont.cmapTag); |
|
158 |
short numberSubTables = cmapBuffer.getShort(2); |
|
159 |
||
160 |
/* locate the offsets of all 3,* (ie Microsoft platform) encodings */ |
|
161 |
for (int i=0; i<numberSubTables; i++) { |
|
162 |
cmapBuffer.position(i * 8 + 4); |
|
163 |
platformID = cmapBuffer.getShort(); |
|
164 |
if (platformID == 3) { |
|
165 |
threeStar = true; |
|
166 |
encodingID = cmapBuffer.getShort(); |
|
167 |
offset = cmapBuffer.getInt(); |
|
168 |
switch (encodingID) { |
|
169 |
case 0: three0 = offset; break; // MS Symbol encoding |
|
170 |
case 1: three1 = offset; break; // MS Unicode cmap |
|
171 |
case 2: three2 = offset; break; // ShiftJIS cmap. |
|
172 |
case 3: three3 = offset; break; // GBK cmap |
|
173 |
case 4: three4 = offset; break; // Big 5 cmap |
|
174 |
case 5: three5 = offset; break; // Wansung |
|
175 |
case 6: three6 = offset; break; // Johab |
|
176 |
case 10: three10 = offset; break; // MS Unicode surrogates |
|
177 |
} |
|
50840 | 178 |
} else if (platformID == 0) { |
179 |
encodingID = cmapBuffer.getShort(); |
|
180 |
offset = cmapBuffer.getInt(); |
|
181 |
if (encodingID == 5) { |
|
182 |
zero5 = offset; |
|
183 |
} |
|
2 | 184 |
} |
185 |
} |
|
186 |
||
187 |
/* This defines the preference order for cmap subtables */ |
|
188 |
if (threeStar) { |
|
189 |
if (three10 != 0) { |
|
190 |
cmap = createCMap(cmapBuffer, three10, null); |
|
191 |
} |
|
192 |
else if (three0 != 0) { |
|
193 |
/* The special case treatment of these fonts leads to |
|
194 |
* anomalies where a user can view "wingdings" and "wingdings2" |
|
195 |
* and the latter shows all its code points in the unicode |
|
196 |
* private use area at 0xF000->0XF0FF and the former shows |
|
197 |
* a scattered subset of its glyphs that are known mappings to |
|
198 |
* unicode code points. |
|
199 |
* The primary purpose of these mappings was to facilitate |
|
200 |
* display of symbol chars etc in composite fonts, however |
|
201 |
* this is not needed as all these code points are covered |
|
50348
008f416a79cb
8191522: Remove Bigelow&Holmes Lucida fonts from JDK sources
prr
parents:
47216
diff
changeset
|
202 |
* by some other platform symbol font. |
2 | 203 |
* Commenting this out reduces the role of these two files |
204 |
* (assuming that they continue to be used in font.properties) |
|
205 |
* to just one of contributing to the overall composite |
|
206 |
* font metrics, and also AWT can still access the fonts. |
|
207 |
* Clients which explicitly accessed these fonts as names |
|
208 |
* "Symbol" and "Wingdings" (ie as physical fonts) and |
|
209 |
* expected to see a scattering of these characters will |
|
210 |
* see them now as missing. How much of a problem is this? |
|
211 |
* Perhaps we could still support this mapping just for |
|
212 |
* "Symbol.ttf" but I suspect some users would prefer it |
|
213 |
* to be mapped in to the Latin range as that is how |
|
214 |
* the "symbol" font is used in native apps. |
|
215 |
*/ |
|
216 |
// String name = font.platName.toLowerCase(Locale.ENGLISH); |
|
217 |
// if (name.endsWith("symbol.ttf")) { |
|
218 |
// cmap = createSymbolCMap(cmapBuffer, three0, Symbols_b2c); |
|
219 |
// } else if (name.endsWith("wingding.ttf")) { |
|
220 |
// cmap = createSymbolCMap(cmapBuffer, three0, WingDings_b2c); |
|
221 |
// } else { |
|
222 |
cmap = createCMap(cmapBuffer, three0, null); |
|
223 |
// } |
|
224 |
} |
|
225 |
else if (three1 != 0) { |
|
226 |
cmap = createCMap(cmapBuffer, three1, null); |
|
227 |
} |
|
228 |
else if (three2 != 0) { |
|
229 |
cmap = createCMap(cmapBuffer, three2, |
|
230 |
getConverterMap(ShiftJISEncoding)); |
|
231 |
} |
|
232 |
else if (three3 != 0) { |
|
233 |
cmap = createCMap(cmapBuffer, three3, |
|
234 |
getConverterMap(GBKEncoding)); |
|
235 |
} |
|
236 |
else if (three4 != 0) { |
|
237 |
/* GB2312 TrueType fonts on Solaris have wrong encoding ID for |
|
238 |
* cmap table, these fonts have EncodingID 4 which is Big5 |
|
239 |
* encoding according the TrueType spec, but actually the |
|
240 |
* fonts are using gb2312 encoding, have to use this |
|
241 |
* workaround to make Solaris zh_CN locale work. -sherman |
|
242 |
*/ |
|
3928 | 243 |
if (FontUtilities.isSolaris && font.platName != null && |
2 | 244 |
(font.platName.startsWith( |
245 |
"/usr/openwin/lib/locale/zh_CN.EUC/X11/fonts/TrueType") || |
|
246 |
font.platName.startsWith( |
|
247 |
"/usr/openwin/lib/locale/zh_CN/X11/fonts/TrueType") || |
|
248 |
font.platName.startsWith( |
|
249 |
"/usr/openwin/lib/locale/zh/X11/fonts/TrueType"))) { |
|
250 |
cmap = createCMap(cmapBuffer, three4, |
|
251 |
getConverterMap(GBKEncoding)); |
|
252 |
} |
|
253 |
else { |
|
254 |
cmap = createCMap(cmapBuffer, three4, |
|
255 |
getConverterMap(Big5Encoding)); |
|
256 |
} |
|
257 |
} |
|
258 |
else if (three5 != 0) { |
|
259 |
cmap = createCMap(cmapBuffer, three5, |
|
260 |
getConverterMap(WansungEncoding)); |
|
261 |
} |
|
262 |
else if (three6 != 0) { |
|
263 |
cmap = createCMap(cmapBuffer, three6, |
|
264 |
getConverterMap(JohabEncoding)); |
|
265 |
} |
|
266 |
} else { |
|
267 |
/* No 3,* subtable was found. Just use whatever is the first |
|
268 |
* table listed. Not very useful but maybe better than |
|
269 |
* rejecting the font entirely? |
|
270 |
*/ |
|
271 |
cmap = createCMap(cmapBuffer, cmapBuffer.getInt(8), null); |
|
272 |
} |
|
50840 | 273 |
// For Unicode Variation Sequences |
274 |
if (cmap != null && zero5 != 0) { |
|
275 |
cmap.createUVS(cmapBuffer, zero5); |
|
276 |
} |
|
2 | 277 |
return cmap; |
278 |
} |
|
279 |
||
280 |
/* speed up the converting by setting the range for double |
|
281 |
* byte characters; |
|
282 |
*/ |
|
283 |
static char[] getConverter(short encodingID) { |
|
284 |
int dBegin = 0x8000; |
|
285 |
int dEnd = 0xffff; |
|
286 |
String encoding; |
|
287 |
||
288 |
switch (encodingID) { |
|
289 |
case ShiftJISEncoding: |
|
290 |
dBegin = 0x8140; |
|
291 |
dEnd = 0xfcfc; |
|
292 |
encoding = "SJIS"; |
|
293 |
break; |
|
294 |
case GBKEncoding: |
|
295 |
dBegin = 0x8140; |
|
296 |
dEnd = 0xfea0; |
|
297 |
encoding = "GBK"; |
|
298 |
break; |
|
299 |
case Big5Encoding: |
|
300 |
dBegin = 0xa140; |
|
301 |
dEnd = 0xfefe; |
|
302 |
encoding = "Big5"; |
|
303 |
break; |
|
304 |
case WansungEncoding: |
|
305 |
dBegin = 0xa1a1; |
|
306 |
dEnd = 0xfede; |
|
307 |
encoding = "EUC_KR"; |
|
308 |
break; |
|
309 |
case JohabEncoding: |
|
310 |
dBegin = 0x8141; |
|
311 |
dEnd = 0xfdfe; |
|
312 |
encoding = "Johab"; |
|
313 |
break; |
|
314 |
default: |
|
315 |
return null; |
|
316 |
} |
|
317 |
||
318 |
try { |
|
319 |
char[] convertedChars = new char[65536]; |
|
320 |
for (int i=0; i<65536; i++) { |
|
321 |
convertedChars[i] = noSuchChar; |
|
322 |
} |
|
323 |
||
324 |
byte[] inputBytes = new byte[(dEnd-dBegin+1)*2]; |
|
325 |
char[] outputChars = new char[(dEnd-dBegin+1)]; |
|
326 |
||
327 |
int j = 0; |
|
328 |
int firstByte; |
|
329 |
if (encodingID == ShiftJISEncoding) { |
|
330 |
for (int i = dBegin; i <= dEnd; i++) { |
|
331 |
firstByte = (i >> 8 & 0xff); |
|
332 |
if (firstByte >= 0xa1 && firstByte <= 0xdf) { |
|
333 |
//sjis halfwidth katakana |
|
334 |
inputBytes[j++] = (byte)0xff; |
|
335 |
inputBytes[j++] = (byte)0xff; |
|
336 |
} else { |
|
337 |
inputBytes[j++] = (byte)firstByte; |
|
338 |
inputBytes[j++] = (byte)(i & 0xff); |
|
339 |
} |
|
340 |
} |
|
341 |
} else { |
|
342 |
for (int i = dBegin; i <= dEnd; i++) { |
|
343 |
inputBytes[j++] = (byte)(i>>8 & 0xff); |
|
344 |
inputBytes[j++] = (byte)(i & 0xff); |
|
345 |
} |
|
346 |
} |
|
347 |
||
348 |
Charset.forName(encoding).newDecoder() |
|
349 |
.onMalformedInput(CodingErrorAction.REPLACE) |
|
350 |
.onUnmappableCharacter(CodingErrorAction.REPLACE) |
|
351 |
.replaceWith("\u0000") |
|
352 |
.decode(ByteBuffer.wrap(inputBytes, 0, inputBytes.length), |
|
353 |
CharBuffer.wrap(outputChars, 0, outputChars.length), |
|
354 |
true); |
|
355 |
||
356 |
// ensure single byte ascii |
|
357 |
for (int i = 0x20; i <= 0x7e; i++) { |
|
358 |
convertedChars[i] = (char)i; |
|
359 |
} |
|
360 |
||
361 |
//sjis halfwidth katakana |
|
362 |
if (encodingID == ShiftJISEncoding) { |
|
363 |
for (int i = 0xa1; i <= 0xdf; i++) { |
|
364 |
convertedChars[i] = (char)(i - 0xa1 + 0xff61); |
|
365 |
} |
|
366 |
} |
|
367 |
||
368 |
/* It would save heap space (approx 60Kbytes for each of these |
|
369 |
* converters) if stored only valid ranges (ie returned |
|
370 |
* outputChars directly. But this is tricky since want to |
|
371 |
* include the ASCII range too. |
|
372 |
*/ |
|
373 |
// System.err.println("oc.len="+outputChars.length); |
|
374 |
// System.err.println("cc.len="+convertedChars.length); |
|
375 |
// System.err.println("dbegin="+dBegin); |
|
376 |
System.arraycopy(outputChars, 0, convertedChars, dBegin, |
|
377 |
outputChars.length); |
|
378 |
||
379 |
//return convertedChars; |
|
380 |
/* invert this map as now want it to map from Unicode |
|
381 |
* to other encoding. |
|
382 |
*/ |
|
383 |
char [] invertedChars = new char[65536]; |
|
384 |
for (int i=0;i<65536;i++) { |
|
385 |
if (convertedChars[i] != noSuchChar) { |
|
386 |
invertedChars[convertedChars[i]] = (char)i; |
|
387 |
} |
|
388 |
} |
|
389 |
return invertedChars; |
|
390 |
||
391 |
} catch (Exception e) { |
|
392 |
e.printStackTrace(); |
|
393 |
} |
|
394 |
return null; |
|
395 |
} |
|
396 |
||
397 |
/* |
|
398 |
* The returned array maps to unicode from some other 2 byte encoding |
|
399 |
* eg for a 2byte index which represents a SJIS char, the indexed |
|
400 |
* value is the corresponding unicode char. |
|
401 |
*/ |
|
402 |
static char[] getConverterMap(short encodingID) { |
|
403 |
if (converterMaps[encodingID] == null) { |
|
404 |
converterMaps[encodingID] = getConverter(encodingID); |
|
405 |
} |
|
406 |
return converterMaps[encodingID]; |
|
407 |
} |
|
408 |
||
409 |
||
410 |
static CMap createCMap(ByteBuffer buffer, int offset, char[] xlat) { |
|
411 |
/* First do a sanity check that this cmap subtable is contained |
|
412 |
* within the cmap table. |
|
413 |
*/ |
|
414 |
int subtableFormat = buffer.getChar(offset); |
|
415 |
long subtableLength; |
|
416 |
if (subtableFormat < 8) { |
|
417 |
subtableLength = buffer.getChar(offset+2); |
|
418 |
} else { |
|
419 |
subtableLength = buffer.getInt(offset+4) & INTMASK; |
|
420 |
} |
|
421 |
if (offset+subtableLength > buffer.capacity()) { |
|
3928 | 422 |
if (FontUtilities.isLogging()) { |
423 |
FontUtilities.getLogger().warning("Cmap subtable overflows buffer."); |
|
2 | 424 |
} |
425 |
} |
|
426 |
switch (subtableFormat) { |
|
427 |
case 0: return new CMapFormat0(buffer, offset); |
|
428 |
case 2: return new CMapFormat2(buffer, offset, xlat); |
|
429 |
case 4: return new CMapFormat4(buffer, offset, xlat); |
|
430 |
case 6: return new CMapFormat6(buffer, offset, xlat); |
|
431 |
case 8: return new CMapFormat8(buffer, offset, xlat); |
|
432 |
case 10: return new CMapFormat10(buffer, offset, xlat); |
|
433 |
case 12: return new CMapFormat12(buffer, offset, xlat); |
|
434 |
default: throw new RuntimeException("Cmap format unimplemented: " + |
|
435 |
(int)buffer.getChar(offset)); |
|
436 |
} |
|
437 |
} |
|
438 |
||
50840 | 439 |
private void createUVS(ByteBuffer buffer, int offset) { |
440 |
int subtableFormat = buffer.getChar(offset); |
|
441 |
if (subtableFormat == 14) { |
|
442 |
long subtableLength = buffer.getInt(offset + 2) & INTMASK; |
|
443 |
if (offset + subtableLength > buffer.capacity()) { |
|
444 |
if (FontUtilities.isLogging()) { |
|
445 |
FontUtilities.getLogger() |
|
446 |
.warning("Cmap UVS subtable overflows buffer."); |
|
447 |
} |
|
448 |
} |
|
449 |
try { |
|
450 |
this.uvs = new UVS(buffer, offset); |
|
451 |
} catch (Throwable t) { |
|
452 |
t.printStackTrace(); |
|
453 |
} |
|
454 |
} |
|
455 |
return; |
|
456 |
} |
|
457 |
||
2 | 458 |
/* |
459 |
final char charVal(byte[] cmap, int index) { |
|
460 |
return (char)(((0xff & cmap[index]) << 8)+(0xff & cmap[index+1])); |
|
461 |
} |
|
462 |
||
463 |
final short shortVal(byte[] cmap, int index) { |
|
464 |
return (short)(((0xff & cmap[index]) << 8)+(0xff & cmap[index+1])); |
|
465 |
} |
|
466 |
*/ |
|
467 |
abstract char getGlyph(int charCode); |
|
468 |
||
469 |
/* Format 4 Header is |
|
470 |
* ushort format (off=0) |
|
471 |
* ushort length (off=2) |
|
472 |
* ushort language (off=4) |
|
473 |
* ushort segCountX2 (off=6) |
|
474 |
* ushort searchRange (off=8) |
|
475 |
* ushort entrySelector (off=10) |
|
476 |
* ushort rangeShift (off=12) |
|
477 |
* ushort endCount[segCount] (off=14) |
|
478 |
* ushort reservedPad |
|
479 |
* ushort startCount[segCount] |
|
480 |
* short idDelta[segCount] |
|
481 |
* idRangeOFfset[segCount] |
|
482 |
* ushort glyphIdArray[] |
|
483 |
*/ |
|
484 |
static class CMapFormat4 extends CMap { |
|
485 |
int segCount; |
|
486 |
int entrySelector; |
|
487 |
int rangeShift; |
|
488 |
char[] endCount; |
|
489 |
char[] startCount; |
|
490 |
short[] idDelta; |
|
491 |
char[] idRangeOffset; |
|
492 |
char[] glyphIds; |
|
493 |
||
494 |
CMapFormat4(ByteBuffer bbuffer, int offset, char[] xlat) { |
|
495 |
||
496 |
this.xlat = xlat; |
|
497 |
||
498 |
bbuffer.position(offset); |
|
499 |
CharBuffer buffer = bbuffer.asCharBuffer(); |
|
500 |
buffer.get(); // skip, we already know format=4 |
|
501 |
int subtableLength = buffer.get(); |
|
502 |
/* Try to recover from some bad fonts which specify a subtable |
|
503 |
* length that would overflow the byte buffer holding the whole |
|
504 |
* cmap table. If this isn't a recoverable situation an exception |
|
505 |
* may be thrown which is caught higher up the call stack. |
|
506 |
* Whilst this may seem lenient, in practice, unless the "bad" |
|
507 |
* subtable we are using is the last one in the cmap table we |
|
508 |
* would have no way of knowing about this problem anyway. |
|
509 |
*/ |
|
510 |
if (offset+subtableLength > bbuffer.capacity()) { |
|
511 |
subtableLength = bbuffer.capacity() - offset; |
|
512 |
} |
|
513 |
buffer.get(); // skip language |
|
514 |
segCount = buffer.get()/2; |
|
515 |
int searchRange = buffer.get(); |
|
516 |
entrySelector = buffer.get(); |
|
517 |
rangeShift = buffer.get()/2; |
|
518 |
startCount = new char[segCount]; |
|
519 |
endCount = new char[segCount]; |
|
520 |
idDelta = new short[segCount]; |
|
521 |
idRangeOffset = new char[segCount]; |
|
522 |
||
523 |
for (int i=0; i<segCount; i++) { |
|
524 |
endCount[i] = buffer.get(); |
|
525 |
} |
|
526 |
buffer.get(); // 2 bytes for reserved pad |
|
527 |
for (int i=0; i<segCount; i++) { |
|
528 |
startCount[i] = buffer.get(); |
|
529 |
} |
|
530 |
||
531 |
for (int i=0; i<segCount; i++) { |
|
532 |
idDelta[i] = (short)buffer.get(); |
|
533 |
} |
|
534 |
||
535 |
for (int i=0; i<segCount; i++) { |
|
536 |
char ctmp = buffer.get(); |
|
537 |
idRangeOffset[i] = (char)((ctmp>>1)&0xffff); |
|
538 |
} |
|
539 |
/* Can calculate the number of glyph IDs by subtracting |
|
540 |
* "pos" from the length of the cmap |
|
541 |
*/ |
|
542 |
int pos = (segCount*8+16)/2; |
|
543 |
buffer.position(pos); |
|
544 |
int numGlyphIds = (subtableLength/2 - pos); |
|
545 |
glyphIds = new char[numGlyphIds]; |
|
546 |
for (int i=0;i<numGlyphIds;i++) { |
|
547 |
glyphIds[i] = buffer.get(); |
|
548 |
} |
|
549 |
/* |
|
550 |
System.err.println("segcount="+segCount); |
|
551 |
System.err.println("entrySelector="+entrySelector); |
|
552 |
System.err.println("rangeShift="+rangeShift); |
|
553 |
for (int j=0;j<segCount;j++) { |
|
554 |
System.err.println("j="+j+ " sc="+(int)(startCount[j]&0xffff)+ |
|
555 |
" ec="+(int)(endCount[j]&0xffff)+ |
|
556 |
" delta="+idDelta[j] + |
|
557 |
" ro="+(int)idRangeOffset[j]); |
|
558 |
} |
|
559 |
||
560 |
//System.err.println("numglyphs="+glyphIds.length); |
|
561 |
for (int i=0;i<numGlyphIds;i++) { |
|
562 |
System.err.println("gid["+i+"]="+(int)glyphIds[i]); |
|
563 |
} |
|
564 |
*/ |
|
565 |
} |
|
566 |
||
567 |
char getGlyph(int charCode) { |
|
568 |
||
569 |
int index = 0; |
|
570 |
char glyphCode = 0; |
|
571 |
||
572 |
int controlGlyph = getControlCodeGlyph(charCode, true); |
|
573 |
if (controlGlyph >= 0) { |
|
574 |
return (char)controlGlyph; |
|
575 |
} |
|
576 |
||
577 |
/* presence of translation array indicates that this |
|
578 |
* cmap is in some other (non-unicode encoding). |
|
579 |
* In order to look-up a char->glyph mapping we need to |
|
580 |
* translate the unicode code point to the encoding of |
|
581 |
* the cmap. |
|
582 |
* REMIND: VALID CHARCODES?? |
|
583 |
*/ |
|
584 |
if (xlat != null) { |
|
585 |
charCode = xlat[charCode]; |
|
586 |
} |
|
587 |
||
588 |
/* |
|
589 |
* Citation from the TrueType (and OpenType) spec: |
|
590 |
* The segments are sorted in order of increasing endCode |
|
591 |
* values, and the segment values are specified in four parallel |
|
592 |
* arrays. You search for the first endCode that is greater than |
|
593 |
* or equal to the character code you want to map. If the |
|
594 |
* corresponding startCode is less than or equal to the |
|
595 |
* character code, then you use the corresponding idDelta and |
|
596 |
* idRangeOffset to map the character code to a glyph index |
|
597 |
* (otherwise, the missingGlyph is returned). |
|
598 |
*/ |
|
599 |
||
600 |
/* |
|
601 |
* CMAP format4 defines several fields for optimized search of |
|
602 |
* the segment list (entrySelector, searchRange, rangeShift). |
|
603 |
* However, benefits are neglible and some fonts have incorrect |
|
604 |
* data - so we use straightforward binary search (see bug 6247425) |
|
605 |
*/ |
|
606 |
int left = 0, right = startCount.length; |
|
607 |
index = startCount.length >> 1; |
|
608 |
while (left < right) { |
|
609 |
if (endCount[index] < charCode) { |
|
610 |
left = index + 1; |
|
611 |
} else { |
|
612 |
right = index; |
|
613 |
} |
|
614 |
index = (left + right) >> 1; |
|
615 |
} |
|
616 |
||
617 |
if (charCode >= startCount[index] && charCode <= endCount[index]) { |
|
618 |
int rangeOffset = idRangeOffset[index]; |
|
619 |
||
620 |
if (rangeOffset == 0) { |
|
621 |
glyphCode = (char)(charCode + idDelta[index]); |
|
622 |
} else { |
|
623 |
/* Calculate an index into the glyphIds array */ |
|
624 |
||
625 |
/* |
|
626 |
System.err.println("rangeoffset="+rangeOffset+ |
|
627 |
" charCode=" + charCode + |
|
628 |
" scnt["+index+"]="+(int)startCount[index] + |
|
629 |
" segCnt="+segCount); |
|
630 |
*/ |
|
631 |
||
632 |
int glyphIDIndex = rangeOffset - segCount + index |
|
633 |
+ (charCode - startCount[index]); |
|
634 |
glyphCode = glyphIds[glyphIDIndex]; |
|
635 |
if (glyphCode != 0) { |
|
636 |
glyphCode = (char)(glyphCode + idDelta[index]); |
|
637 |
} |
|
638 |
} |
|
639 |
} |
|
640 |
if (glyphCode != 0) { |
|
641 |
//System.err.println("cc="+Integer.toHexString((int)charCode) + " gc="+(int)glyphCode); |
|
642 |
} |
|
643 |
return glyphCode; |
|
644 |
} |
|
645 |
} |
|
646 |
||
647 |
// Format 0: Byte Encoding table |
|
648 |
static class CMapFormat0 extends CMap { |
|
649 |
byte [] cmap; |
|
650 |
||
651 |
CMapFormat0(ByteBuffer buffer, int offset) { |
|
652 |
||
653 |
/* skip 6 bytes of format, length, and version */ |
|
654 |
int len = buffer.getChar(offset+2); |
|
655 |
cmap = new byte[len-6]; |
|
656 |
buffer.position(offset+6); |
|
657 |
buffer.get(cmap); |
|
658 |
} |
|
659 |
||
660 |
char getGlyph(int charCode) { |
|
661 |
if (charCode < 256) { |
|
662 |
if (charCode < 0x0010) { |
|
663 |
switch (charCode) { |
|
664 |
case 0x0009: |
|
665 |
case 0x000a: |
|
666 |
case 0x000d: return CharToGlyphMapper.INVISIBLE_GLYPH_ID; |
|
667 |
} |
|
668 |
} |
|
669 |
return (char)(0xff & cmap[charCode]); |
|
670 |
} else { |
|
671 |
return 0; |
|
672 |
} |
|
673 |
} |
|
674 |
} |
|
675 |
||
676 |
// static CMap createSymbolCMap(ByteBuffer buffer, int offset, char[] syms) { |
|
677 |
||
678 |
// CMap cmap = createCMap(buffer, offset, null); |
|
679 |
// if (cmap == null) { |
|
680 |
// return null; |
|
681 |
// } else { |
|
682 |
// return new CMapFormatSymbol(cmap, syms); |
|
683 |
// } |
|
684 |
// } |
|
685 |
||
686 |
// static class CMapFormatSymbol extends CMap { |
|
687 |
||
688 |
// CMap cmap; |
|
689 |
// static final int NUM_BUCKETS = 128; |
|
690 |
// Bucket[] buckets = new Bucket[NUM_BUCKETS]; |
|
691 |
||
692 |
// class Bucket { |
|
693 |
// char unicode; |
|
694 |
// char glyph; |
|
695 |
// Bucket next; |
|
696 |
||
697 |
// Bucket(char u, char g) { |
|
698 |
// unicode = u; |
|
699 |
// glyph = g; |
|
700 |
// } |
|
701 |
// } |
|
702 |
||
703 |
// CMapFormatSymbol(CMap cmap, char[] syms) { |
|
704 |
||
705 |
// this.cmap = cmap; |
|
706 |
||
707 |
// for (int i=0;i<syms.length;i++) { |
|
708 |
// char unicode = syms[i]; |
|
709 |
// if (unicode != noSuchChar) { |
|
710 |
// char glyph = cmap.getGlyph(i + 0xf000); |
|
711 |
// int hash = unicode % NUM_BUCKETS; |
|
712 |
// Bucket bucket = new Bucket(unicode, glyph); |
|
713 |
// if (buckets[hash] == null) { |
|
714 |
// buckets[hash] = bucket; |
|
715 |
// } else { |
|
716 |
// Bucket b = buckets[hash]; |
|
717 |
// while (b.next != null) { |
|
718 |
// b = b.next; |
|
719 |
// } |
|
720 |
// b.next = bucket; |
|
721 |
// } |
|
722 |
// } |
|
723 |
// } |
|
724 |
// } |
|
725 |
||
726 |
// char getGlyph(int unicode) { |
|
727 |
// if (unicode >= 0x1000) { |
|
728 |
// return 0; |
|
729 |
// } |
|
730 |
// else if (unicode >=0xf000 && unicode < 0xf100) { |
|
731 |
// return cmap.getGlyph(unicode); |
|
732 |
// } else { |
|
733 |
// Bucket b = buckets[unicode % NUM_BUCKETS]; |
|
734 |
// while (b != null) { |
|
735 |
// if (b.unicode == unicode) { |
|
736 |
// return b.glyph; |
|
737 |
// } else { |
|
738 |
// b = b.next; |
|
739 |
// } |
|
740 |
// } |
|
741 |
// return 0; |
|
742 |
// } |
|
743 |
// } |
|
744 |
// } |
|
745 |
||
746 |
// Format 2: High-byte mapping through table |
|
747 |
static class CMapFormat2 extends CMap { |
|
748 |
||
749 |
char[] subHeaderKey = new char[256]; |
|
750 |
/* Store subheaders in individual arrays |
|
751 |
* A SubHeader entry theortically looks like { |
|
752 |
* char firstCode; |
|
753 |
* char entryCount; |
|
754 |
* short idDelta; |
|
755 |
* char idRangeOffset; |
|
756 |
* } |
|
757 |
*/ |
|
758 |
char[] firstCodeArray; |
|
759 |
char[] entryCountArray; |
|
760 |
short[] idDeltaArray; |
|
761 |
char[] idRangeOffSetArray; |
|
762 |
||
763 |
char[] glyphIndexArray; |
|
764 |
||
765 |
CMapFormat2(ByteBuffer buffer, int offset, char[] xlat) { |
|
766 |
||
767 |
this.xlat = xlat; |
|
768 |
||
769 |
int tableLen = buffer.getChar(offset+2); |
|
770 |
buffer.position(offset+6); |
|
771 |
CharBuffer cBuffer = buffer.asCharBuffer(); |
|
772 |
char maxSubHeader = 0; |
|
773 |
for (int i=0;i<256;i++) { |
|
774 |
subHeaderKey[i] = cBuffer.get(); |
|
775 |
if (subHeaderKey[i] > maxSubHeader) { |
|
776 |
maxSubHeader = subHeaderKey[i]; |
|
777 |
} |
|
778 |
} |
|
779 |
/* The value of the subHeaderKey is 8 * the subHeader index, |
|
780 |
* so the number of subHeaders can be obtained by dividing |
|
781 |
* this value bv 8 and adding 1. |
|
782 |
*/ |
|
783 |
int numSubHeaders = (maxSubHeader >> 3) +1; |
|
784 |
firstCodeArray = new char[numSubHeaders]; |
|
785 |
entryCountArray = new char[numSubHeaders]; |
|
786 |
idDeltaArray = new short[numSubHeaders]; |
|
787 |
idRangeOffSetArray = new char[numSubHeaders]; |
|
788 |
for (int i=0; i<numSubHeaders; i++) { |
|
789 |
firstCodeArray[i] = cBuffer.get(); |
|
790 |
entryCountArray[i] = cBuffer.get(); |
|
791 |
idDeltaArray[i] = (short)cBuffer.get(); |
|
792 |
idRangeOffSetArray[i] = cBuffer.get(); |
|
793 |
// System.out.println("sh["+i+"]:fc="+(int)firstCodeArray[i]+ |
|
794 |
// " ec="+(int)entryCountArray[i]+ |
|
795 |
// " delta="+(int)idDeltaArray[i]+ |
|
796 |
// " offset="+(int)idRangeOffSetArray[i]); |
|
797 |
} |
|
798 |
||
799 |
int glyphIndexArrSize = (tableLen-518-numSubHeaders*8)/2; |
|
800 |
glyphIndexArray = new char[glyphIndexArrSize]; |
|
801 |
for (int i=0; i<glyphIndexArrSize;i++) { |
|
802 |
glyphIndexArray[i] = cBuffer.get(); |
|
803 |
} |
|
804 |
} |
|
805 |
||
806 |
char getGlyph(int charCode) { |
|
807 |
int controlGlyph = getControlCodeGlyph(charCode, true); |
|
808 |
if (controlGlyph >= 0) { |
|
809 |
return (char)controlGlyph; |
|
810 |
} |
|
811 |
||
812 |
if (xlat != null) { |
|
813 |
charCode = xlat[charCode]; |
|
814 |
} |
|
815 |
||
816 |
char highByte = (char)(charCode >> 8); |
|
817 |
char lowByte = (char)(charCode & 0xff); |
|
818 |
int key = subHeaderKey[highByte]>>3; // index into subHeaders |
|
819 |
char mapMe; |
|
820 |
||
821 |
if (key != 0) { |
|
822 |
mapMe = lowByte; |
|
823 |
} else { |
|
824 |
mapMe = highByte; |
|
825 |
if (mapMe == 0) { |
|
826 |
mapMe = lowByte; |
|
827 |
} |
|
828 |
} |
|
829 |
||
830 |
// System.err.println("charCode="+Integer.toHexString(charCode)+ |
|
831 |
// " key="+key+ " mapMe="+Integer.toHexString(mapMe)); |
|
832 |
char firstCode = firstCodeArray[key]; |
|
833 |
if (mapMe < firstCode) { |
|
834 |
return 0; |
|
835 |
} else { |
|
836 |
mapMe -= firstCode; |
|
837 |
} |
|
838 |
||
839 |
if (mapMe < entryCountArray[key]) { |
|
840 |
/* "address" arithmetic is needed to calculate the offset |
|
841 |
* into glyphIndexArray. "idRangeOffSetArray[key]" specifies |
|
842 |
* the number of bytes from that location in the table where |
|
843 |
* the subarray of glyphIndexes starting at "firstCode" begins. |
|
844 |
* Each entry in the subHeader table is 8 bytes, and the |
|
845 |
* idRangeOffSetArray field is at offset 6 in the entry. |
|
846 |
* The glyphIndexArray immediately follows the subHeaders. |
|
847 |
* So if there are "N" entries then the number of bytes to the |
|
848 |
* start of glyphIndexArray is (N-key)*8-6. |
|
849 |
* Subtract this from the idRangeOffSetArray value to get |
|
850 |
* the number of bytes into glyphIndexArray and divide by 2 to |
|
851 |
* get the (char) array index. |
|
852 |
*/ |
|
853 |
int glyphArrayOffset = ((idRangeOffSetArray.length-key)*8)-6; |
|
854 |
int glyphSubArrayStart = |
|
855 |
(idRangeOffSetArray[key] - glyphArrayOffset)/2; |
|
856 |
char glyphCode = glyphIndexArray[glyphSubArrayStart+mapMe]; |
|
857 |
if (glyphCode != 0) { |
|
858 |
glyphCode += idDeltaArray[key]; //idDelta |
|
859 |
return glyphCode; |
|
860 |
} |
|
861 |
} |
|
862 |
return 0; |
|
863 |
} |
|
864 |
} |
|
865 |
||
866 |
// Format 6: Trimmed table mapping |
|
867 |
static class CMapFormat6 extends CMap { |
|
868 |
||
869 |
char firstCode; |
|
870 |
char entryCount; |
|
871 |
char[] glyphIdArray; |
|
872 |
||
873 |
CMapFormat6(ByteBuffer bbuffer, int offset, char[] xlat) { |
|
874 |
||
875 |
bbuffer.position(offset+6); |
|
876 |
CharBuffer buffer = bbuffer.asCharBuffer(); |
|
877 |
firstCode = buffer.get(); |
|
878 |
entryCount = buffer.get(); |
|
879 |
glyphIdArray = new char[entryCount]; |
|
880 |
for (int i=0; i< entryCount; i++) { |
|
881 |
glyphIdArray[i] = buffer.get(); |
|
882 |
} |
|
883 |
} |
|
884 |
||
885 |
char getGlyph(int charCode) { |
|
886 |
int controlGlyph = getControlCodeGlyph(charCode, true); |
|
887 |
if (controlGlyph >= 0) { |
|
888 |
return (char)controlGlyph; |
|
889 |
} |
|
890 |
||
891 |
if (xlat != null) { |
|
892 |
charCode = xlat[charCode]; |
|
893 |
} |
|
894 |
||
895 |
charCode -= firstCode; |
|
896 |
if (charCode < 0 || charCode >= entryCount) { |
|
897 |
return 0; |
|
898 |
} else { |
|
899 |
return glyphIdArray[charCode]; |
|
900 |
} |
|
901 |
} |
|
902 |
} |
|
903 |
||
904 |
// Format 8: mixed 16-bit and 32-bit coverage |
|
905 |
// Seems unlikely this code will ever get tested as we look for |
|
906 |
// MS platform Cmaps and MS states (in the Opentype spec on their website) |
|
907 |
// that MS doesn't support this format |
|
908 |
static class CMapFormat8 extends CMap { |
|
909 |
byte[] is32 = new byte[8192]; |
|
910 |
int nGroups; |
|
911 |
int[] startCharCode; |
|
912 |
int[] endCharCode; |
|
913 |
int[] startGlyphID; |
|
914 |
||
915 |
CMapFormat8(ByteBuffer bbuffer, int offset, char[] xlat) { |
|
916 |
||
917 |
bbuffer.position(12); |
|
918 |
bbuffer.get(is32); |
|
919 |
nGroups = bbuffer.getInt(); |
|
920 |
startCharCode = new int[nGroups]; |
|
921 |
endCharCode = new int[nGroups]; |
|
922 |
startGlyphID = new int[nGroups]; |
|
923 |
} |
|
924 |
||
925 |
char getGlyph(int charCode) { |
|
926 |
if (xlat != null) { |
|
927 |
throw new RuntimeException("xlat array for cmap fmt=8"); |
|
928 |
} |
|
929 |
return 0; |
|
930 |
} |
|
931 |
||
932 |
} |
|
933 |
||
934 |
||
935 |
// Format 4-byte 10: Trimmed table mapping |
|
936 |
// Seems unlikely this code will ever get tested as we look for |
|
937 |
// MS platform Cmaps and MS states (in the Opentype spec on their website) |
|
938 |
// that MS doesn't support this format |
|
939 |
static class CMapFormat10 extends CMap { |
|
940 |
||
941 |
long firstCode; |
|
942 |
int entryCount; |
|
943 |
char[] glyphIdArray; |
|
944 |
||
945 |
CMapFormat10(ByteBuffer bbuffer, int offset, char[] xlat) { |
|
946 |
||
947 |
firstCode = bbuffer.getInt() & INTMASK; |
|
948 |
entryCount = bbuffer.getInt() & INTMASK; |
|
949 |
bbuffer.position(offset+20); |
|
950 |
CharBuffer buffer = bbuffer.asCharBuffer(); |
|
951 |
glyphIdArray = new char[entryCount]; |
|
952 |
for (int i=0; i< entryCount; i++) { |
|
953 |
glyphIdArray[i] = buffer.get(); |
|
954 |
} |
|
955 |
} |
|
956 |
||
957 |
char getGlyph(int charCode) { |
|
958 |
||
959 |
if (xlat != null) { |
|
960 |
throw new RuntimeException("xlat array for cmap fmt=10"); |
|
961 |
} |
|
962 |
||
963 |
int code = (int)(charCode - firstCode); |
|
964 |
if (code < 0 || code >= entryCount) { |
|
965 |
return 0; |
|
966 |
} else { |
|
967 |
return glyphIdArray[code]; |
|
968 |
} |
|
969 |
} |
|
970 |
} |
|
971 |
||
972 |
// Format 12: Segmented coverage for UCS-4 (fonts supporting |
|
973 |
// surrogate pairs) |
|
974 |
static class CMapFormat12 extends CMap { |
|
975 |
||
976 |
int numGroups; |
|
977 |
int highBit =0; |
|
978 |
int power; |
|
979 |
int extra; |
|
980 |
long[] startCharCode; |
|
981 |
long[] endCharCode; |
|
982 |
int[] startGlyphID; |
|
983 |
||
984 |
CMapFormat12(ByteBuffer buffer, int offset, char[] xlat) { |
|
985 |
if (xlat != null) { |
|
986 |
throw new RuntimeException("xlat array for cmap fmt=12"); |
|
987 |
} |
|
988 |
||
989 |
numGroups = buffer.getInt(offset+12); |
|
990 |
startCharCode = new long[numGroups]; |
|
991 |
endCharCode = new long[numGroups]; |
|
992 |
startGlyphID = new int[numGroups]; |
|
993 |
buffer.position(offset+16); |
|
994 |
buffer = buffer.slice(); |
|
995 |
IntBuffer ibuffer = buffer.asIntBuffer(); |
|
996 |
for (int i=0; i<numGroups; i++) { |
|
997 |
startCharCode[i] = ibuffer.get() & INTMASK; |
|
998 |
endCharCode[i] = ibuffer.get() & INTMASK; |
|
999 |
startGlyphID[i] = ibuffer.get() & INTMASK; |
|
1000 |
} |
|
1001 |
||
1002 |
/* Finds the high bit by binary searching through the bits */ |
|
1003 |
int value = numGroups; |
|
1004 |
||
1005 |
if (value >= 1 << 16) { |
|
1006 |
value >>= 16; |
|
1007 |
highBit += 16; |
|
1008 |
} |
|
1009 |
||
1010 |
if (value >= 1 << 8) { |
|
1011 |
value >>= 8; |
|
1012 |
highBit += 8; |
|
1013 |
} |
|
1014 |
||
1015 |
if (value >= 1 << 4) { |
|
1016 |
value >>= 4; |
|
1017 |
highBit += 4; |
|
1018 |
} |
|
1019 |
||
1020 |
if (value >= 1 << 2) { |
|
1021 |
value >>= 2; |
|
1022 |
highBit += 2; |
|
1023 |
} |
|
1024 |
||
1025 |
if (value >= 1 << 1) { |
|
1026 |
value >>= 1; |
|
1027 |
highBit += 1; |
|
1028 |
} |
|
1029 |
||
1030 |
power = 1 << highBit; |
|
1031 |
extra = numGroups - power; |
|
1032 |
} |
|
1033 |
||
1034 |
char getGlyph(int charCode) { |
|
1035 |
int controlGlyph = getControlCodeGlyph(charCode, false); |
|
1036 |
if (controlGlyph >= 0) { |
|
1037 |
return (char)controlGlyph; |
|
1038 |
} |
|
1039 |
int probe = power; |
|
1040 |
int range = 0; |
|
1041 |
||
1042 |
if (startCharCode[extra] <= charCode) { |
|
1043 |
range = extra; |
|
1044 |
} |
|
1045 |
||
1046 |
while (probe > 1) { |
|
1047 |
probe >>= 1; |
|
1048 |
||
1049 |
if (startCharCode[range+probe] <= charCode) { |
|
1050 |
range += probe; |
|
1051 |
} |
|
1052 |
} |
|
1053 |
||
1054 |
if (startCharCode[range] <= charCode && |
|
1055 |
endCharCode[range] >= charCode) { |
|
1056 |
return (char) |
|
1057 |
(startGlyphID[range] + (charCode - startCharCode[range])); |
|
1058 |
} |
|
1059 |
||
1060 |
return 0; |
|
1061 |
} |
|
1062 |
||
1063 |
} |
|
1064 |
||
1065 |
/* Used to substitute for bad Cmaps. */ |
|
1066 |
static class NullCMapClass extends CMap { |
|
1067 |
||
1068 |
char getGlyph(int charCode) { |
|
1069 |
return 0; |
|
1070 |
} |
|
1071 |
} |
|
1072 |
||
1073 |
public static final NullCMapClass theNullCmap = new NullCMapClass(); |
|
1074 |
||
1075 |
final int getControlCodeGlyph(int charCode, boolean noSurrogates) { |
|
1076 |
if (charCode < 0x0010) { |
|
1077 |
switch (charCode) { |
|
1078 |
case 0x0009: |
|
1079 |
case 0x000a: |
|
1080 |
case 0x000d: return CharToGlyphMapper.INVISIBLE_GLYPH_ID; |
|
1081 |
} |
|
1082 |
} else if (charCode >= 0x200c) { |
|
1083 |
if ((charCode <= 0x200f) || |
|
1084 |
(charCode >= 0x2028 && charCode <= 0x202e) || |
|
1085 |
(charCode >= 0x206a && charCode <= 0x206f)) { |
|
1086 |
return CharToGlyphMapper.INVISIBLE_GLYPH_ID; |
|
1087 |
} else if (noSurrogates && charCode >= 0xFFFF) { |
|
1088 |
return 0; |
|
1089 |
} |
|
1090 |
} |
|
1091 |
return -1; |
|
1092 |
} |
|
50840 | 1093 |
|
1094 |
static class UVS { |
|
1095 |
int numSelectors; |
|
1096 |
int[] selector; |
|
1097 |
||
1098 |
//for Non-Default UVS Table |
|
1099 |
int[] numUVSMapping; |
|
1100 |
int[][] unicodeValue; |
|
1101 |
char[][] glyphID; |
|
1102 |
||
1103 |
UVS(ByteBuffer buffer, int offset) { |
|
1104 |
numSelectors = buffer.getInt(offset+6); |
|
1105 |
selector = new int[numSelectors]; |
|
1106 |
numUVSMapping = new int[numSelectors]; |
|
1107 |
unicodeValue = new int[numSelectors][]; |
|
1108 |
glyphID = new char[numSelectors][]; |
|
1109 |
||
1110 |
for (int i = 0; i < numSelectors; i++) { |
|
1111 |
buffer.position(offset + 10 + i * 11); |
|
1112 |
selector[i] = (buffer.get() & 0xff) << 16; //UINT24 |
|
1113 |
selector[i] += (buffer.get() & 0xff) << 8; |
|
1114 |
selector[i] += buffer.get() & 0xff; |
|
1115 |
||
1116 |
//skip Default UVS Table |
|
1117 |
||
1118 |
//for Non-Default UVS Table |
|
1119 |
int tableOffset = buffer.getInt(offset + 10 + i * 11 + 7); |
|
1120 |
if (tableOffset == 0) { |
|
1121 |
numUVSMapping[i] = 0; |
|
1122 |
} else if (tableOffset > 0) { |
|
1123 |
buffer.position(offset+tableOffset); |
|
1124 |
numUVSMapping[i] = buffer.getInt() & INTMASK; |
|
1125 |
unicodeValue[i] = new int[numUVSMapping[i]]; |
|
1126 |
glyphID[i] = new char[numUVSMapping[i]]; |
|
1127 |
||
1128 |
for (int j = 0; j < numUVSMapping[i]; j++) { |
|
1129 |
int temp = (buffer.get() & 0xff) << 16; //UINT24 |
|
1130 |
temp += (buffer.get() & 0xff) << 8; |
|
1131 |
temp += buffer.get() & 0xff; |
|
1132 |
unicodeValue[i][j] = temp; |
|
1133 |
glyphID[i][j] = buffer.getChar(); |
|
1134 |
} |
|
1135 |
} |
|
1136 |
} |
|
1137 |
} |
|
1138 |
||
1139 |
static final int VS_NOGLYPH = 0; |
|
1140 |
private int getGlyph(int charCode, int variationSelector) { |
|
1141 |
int targetSelector = -1; |
|
1142 |
for (int i = 0; i < numSelectors; i++) { |
|
1143 |
if (selector[i] == variationSelector) { |
|
1144 |
targetSelector = i; |
|
1145 |
break; |
|
1146 |
} |
|
1147 |
} |
|
1148 |
if (targetSelector == -1) { |
|
1149 |
return VS_NOGLYPH; |
|
1150 |
} |
|
1151 |
if (numUVSMapping[targetSelector] > 0) { |
|
1152 |
int index = java.util.Arrays.binarySearch( |
|
1153 |
unicodeValue[targetSelector], charCode); |
|
1154 |
if (index >= 0) { |
|
1155 |
return glyphID[targetSelector][index]; |
|
1156 |
} |
|
1157 |
} |
|
1158 |
return VS_NOGLYPH; |
|
1159 |
} |
|
1160 |
} |
|
1161 |
||
1162 |
char getVariationGlyph(int charCode, int variationSelector) { |
|
1163 |
char glyph = 0; |
|
1164 |
if (uvs == null) { |
|
1165 |
glyph = getGlyph(charCode); |
|
1166 |
} else { |
|
1167 |
int result = uvs.getGlyph(charCode, variationSelector); |
|
1168 |
if (result > 0) { |
|
1169 |
glyph = (char)(result & 0xFFFF); |
|
1170 |
} else { |
|
1171 |
glyph = getGlyph(charCode); |
|
1172 |
} |
|
1173 |
} |
|
1174 |
return glyph; |
|
1175 |
} |
|
2 | 1176 |
} |