3 * This file is part of LyX, the document processor.
4 * Licence details can be found in the file COPYING.
6 * \author Lars Gullik Bjønnes
7 * \author Jean-Marc Lasgouttes
10 * Full author contact details are available in file CREDITS.
18 #include "LaTeXFeatures.h"
22 #include "support/filename.h"
23 #include "support/lstrings.h"
24 #include "support/unicode.h"
29 using support::FileName;
31 #ifndef CXX_GLOBAL_CSTD
43 char_type arabic_table2[63][4] = {
44 {0x41, 0x41, 0x41, 0x41}, // 0xc1 = hamza
45 {0x42, 0xa1, 0x42, 0xa1}, // 0xc2 = ligature madda on alef
46 {0x43, 0xa2, 0x43, 0xa2}, // 0xc3 = ligature hamza on alef
47 {0x44, 0xa3, 0x44, 0xa3}, // 0xc4 = ligature hamza on waw
48 {0x45, 0xa4, 0x45, 0xa4}, // 0xc5 = ligature hamza under alef
49 {0x46, 0xf9, 0xf8, 0xa0}, // 0xc6 = ligature hamza on ya
50 {0x47, 0xa5, 0x47, 0xa5}, // 0xc7 = alef
51 {0x48, 0xae, 0xac, 0xad}, // 0xc8 = baa
52 {0x49, 0xb1, 0x49, 0xb1}, // 0xc9 = taa marbuta
53 {0x4a, 0xb4, 0xb2, 0xb3}, // 0xca = taa
54 {0x4b, 0xb7, 0xb5, 0xb6}, // 0xcb = thaa
55 {0x4c, 0xba, 0xb8, 0xb9}, // 0xcc = jeem
56 {0x4d, 0xbd, 0xbb, 0xbc}, // 0xcd = haa
57 {0x4e, 0xc0, 0xbe, 0xbf}, // 0xce = khaa
58 {0x4f, 0xa6, 0x4f, 0xa6}, // 0xcf = dal
60 {0x50, 0xa7, 0x50, 0xa7}, // 0xd0 = thal
61 {0x51, 0xa8, 0x51, 0xa8}, // 0xd1 = ra
62 {0x52, 0xa9, 0x52, 0xa9}, // 0xd2 = zain
63 {0x53, 0xc3, 0xc1, 0xc2}, // 0xd3 = seen
64 {0x54, 0xc6, 0xc4, 0xc5}, // 0xd4 = sheen
65 {0x55, 0xc9, 0xc7, 0xc8}, // 0xd5 = sad
66 {0x56, 0xcc, 0xca, 0xcb}, // 0xd6 = dad
67 {0x57, 0xcf, 0xcd, 0xce}, // 0xd7 = tah
68 {0x58, 0xd2, 0xd0, 0xd1}, // 0xd8 = zah
69 {0x59, 0xd5, 0xd3, 0xd4}, // 0xd9 = ain
70 {0x5a, 0xd8, 0xd6, 0xd7}, // 0xda = ghain
78 {0x61, 0xdb, 0xd9, 0xda}, // 0xe1 = fa
79 {0x62, 0xde, 0xdc, 0xdd}, // 0xe2 = qaf
80 {0x63, 0xe1, 0xdf, 0xe0}, // 0xe3 = kaf
81 {0x64, 0xe4, 0xe2, 0xe3}, // 0xe4 = lam
82 {0x65, 0xe7, 0xe5, 0xe6}, // 0xe5 = meem
83 {0x66, 0xea, 0xe8, 0xe9}, // 0xe6 = noon
84 {0x67, 0xed, 0xeb, 0xec}, // 0xe7 = ha
85 {0x68, 0xaa, 0x68, 0xaa}, // 0xe8 = waw
86 {0x69, 0xab, 0x69, 0xab}, // 0xe9 = alef maksura
87 {0x6a, 0xf0, 0xee, 0xef}, // 0xea = ya
88 {0x6b, 0x6b, 0x6b, 0x6b}, // 0xeb = fathatan
89 {0x6c, 0x6c, 0x6c, 0x6c}, // 0xec = dammatan
90 {0x6d, 0x6d, 0x6d, 0x6d}, // 0xed = kasratan
91 {0x6e, 0x6e, 0x6e, 0x6e}, // 0xee = fatha
92 {0x6f, 0x6f, 0x6f, 0x6f}, // 0xef = damma
94 {0x70, 0x70, 0x70, 0x70}, // 0xf0 = kasra
95 {0x71, 0x71, 0x71, 0x71}, // 0xf1 = shadda
96 {0x72, 0x72, 0x72, 0x72}, // 0xf2 = sukun
113 char_type arabic_table[63][2] = {
114 {0xc1, 0xc1}, // 0xc1 = hamza
115 {0xc2, 0xc2}, // 0xc2 = ligature madda on alef
116 {0xc3, 0xc3}, // 0xc3 = ligature hamza on alef
117 {0xc4, 0xc4}, // 0xc4 = ligature hamza on waw
118 {0xc5, 0xc5}, // 0xc5 = ligature hamza under alef
119 {0xc6, 0xc0}, // 0xc6 = ligature hamza on ya
120 {0xc7, 0xc7}, // 0xc7 = alef
121 {0xc8, 0xeb}, // 0xc8 = baa
122 {0xc9, 0xc9}, // 0xc9 = taa marbuta
123 {0xca, 0xec}, // 0xca = taa
124 {0xcb, 0xed}, // 0xcb = thaa
125 {0xcc, 0xee}, // 0xcc = jeem
126 {0xcd, 0xef}, // 0xcd = haa
127 {0xce, 0xf0}, // 0xce = khaa
128 {0xcf, 0xcf}, // 0xcf = dal
130 {0xd0, 0xd0}, // 0xd0 = thal
131 {0xd1, 0xd1}, // 0xd1 = ra
132 {0xd2, 0xd2}, // 0xd2 = zain
133 {0xd3, 0xf1}, // 0xd3 = seen
134 {0xd4, 0xf2}, // 0xd4 = sheen
135 {0xd5, 0xf3}, // 0xd5 = sad
136 {0xd6, 0xf4}, // 0xd6 = dad
137 {0xd7, 0xd7}, // 0xd7 = tah
138 {0xd8, 0xd8}, // 0xd8 = zah
139 {0xd9, 0xf5}, // 0xd9 = ain
140 {0xda, 0xf6}, // 0xda = ghain
148 {0xe1, 0xf7}, // 0xe1 = fa
149 {0xe2, 0xf8}, // 0xe2 = qaf
150 {0xe3, 0xf9}, // 0xe3 = kaf
151 {0xe4, 0xfa}, // 0xe4 = lam
152 {0xe5, 0xfb}, // 0xe5 = meem
153 {0xe6, 0xfc}, // 0xe6 = noon
154 {0xe7, 0xfd}, // 0xe7 = ha
155 {0xe8, 0xe8}, // 0xe8 = waw
156 {0xe9, 0xe9}, // 0xe9 = alef maksura
157 {0xea, 0xfe}, // 0xea = ya
158 {0xa8, 0xa8}, // 0xeb = fathatan
159 {0xa9, 0xa9}, // 0xec = dammatan
160 {0xaa, 0xaa}, // 0xed = kasratan
161 {0xab, 0xab}, // 0xee = fatha
162 {0xac, 0xac}, // 0xef = damma
164 {0xad, 0xad}, // 0xf0 = kasra
165 {0xae, 0xae}, // 0xf1 = shadda
166 {0xaf, 0xaf}, // 0xf2 = sukun
183 char_type const arabic_start = 0xc1;
186 /// Information about a single UCS4 character
188 /// LaTeX command for this character
190 /// Needed LaTeX preamble (or feature)
192 /// Is this a combining character?
194 /// Is \c preamble a feature known by LaTeXFeatures, or a raw LaTeX
197 /// Always force the LaTeX command, even if the encoding contains
203 typedef std::map<char_type, CharInfo> CharInfoMap;
204 CharInfoMap unicodesymbols;
209 Encoding::Encoding(string const & n, string const & l, string const & i)
210 : Name_(n), LatexName_(l), iconvName_(i)
213 // UTF8 can encode all 1<<20 + 1<<16 UCS4 code points
214 start_encodable_ = 0x110000;
216 start_encodable_ = 0;
217 // temporarily switch off lyxerr, since we will generate iconv errors
219 for (unsigned short j = 0; j < 256; ++j) {
221 std::vector<char_type> const ucs4 = eightbit_to_ucs4(&c, 1, i);
222 if (ucs4.size() == 1) {
223 char_type const c = ucs4[0];
224 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
225 if (it == unicodesymbols.end() || !it->second.force)
226 encodable_.insert(c);
230 CharSet::iterator it = encodable_.find(start_encodable_);
231 while (it != encodable_.end()) {
232 encodable_.erase(it);
234 it = encodable_.find(start_encodable_);
240 docstring const Encoding::latexChar(char_type c) const
242 if (c < start_encodable_)
243 return docstring(1, c);
244 if (encodable_.find(c) == encodable_.end()) {
245 // c cannot be encoded in this encoding
246 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
247 if (it == unicodesymbols.end())
248 lyxerr << "Could not find LaTeX command for character 0x"
249 << std::hex << c << ".\nLaTeX export will fail."
252 return it->second.command;
254 return docstring(1, c);
258 void Encoding::validate(char_type c, LaTeXFeatures & features) const
260 if (c < start_encodable_)
263 if (encodable_.find(c) != encodable_.end())
266 // c cannot be encoded in this encoding
267 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
268 if (it != unicodesymbols.end() && !it->second.preamble.empty()) {
269 if (it->second.feature)
270 features.require(it->second.preamble);
272 features.addPreambleSnippet(it->second.preamble);
277 bool Encodings::isComposeChar_hebrew(char_type c)
279 return c <= 0xd2 && c >= 0xc0 &&
280 c != 0xce && c != 0xd0;
284 // Special Arabic letters are ones that do not get connected from left
285 // they are hamza, alef_madda, alef_hamza, waw_hamza, alef_hamza_under,
286 // alef, tah_marbota, dal, thal, rah, zai, wow, alef_maksoura
288 bool Encodings::is_arabic_special(char_type c)
290 return (c >= 0xc1 && c <= 0xc5) ||
291 c == 0xc7 || c == 0xc9 ||
292 c == 0xcf || c == 0xe8 ||
293 (c >= 0xd0 && c <= 0xd2) ||
297 bool Encodings::isComposeChar_arabic(char_type c)
299 return c >= 0xeb && c <= 0xf2;
303 bool Encodings::is_arabic(char_type c)
305 return c >= arabic_start && arabic_table[c-arabic_start][0];
309 char_type Encodings::transformChar(char_type c,
310 Encodings::Letter_Form form)
315 if (lyxrc.font_norm_type == LyXRC::ISO_10646_1)
316 return arabic_table2[c-arabic_start][form];
318 return arabic_table[c-arabic_start][form >> 1];
322 bool Encodings::isCombiningChar(char_type c)
324 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
325 if (it != unicodesymbols.end())
326 return it->second.combining;
331 Encoding const * Encodings::getFromLyXName(string const & name) const
333 EncodingList::const_iterator it = encodinglist.find(name);
334 if (it != encodinglist.end())
341 Encoding const * Encodings::getFromLaTeXName(string const & name) const
343 // We don't use std::find_if because it makes copies of the pairs in
345 // This linear search is OK since we don't have many encodings.
346 // Users could even optimize it by putting the encodings they use
347 // most at the top of lib/encodings.
348 EncodingList::const_iterator const end = encodinglist.end();
349 for (EncodingList::const_iterator it = encodinglist.begin(); it != end; ++it)
350 if (it->second.latexName() == name)
356 Encodings::Encodings()
361 void Encodings::read(FileName const & encfile, FileName const & symbolsfile)
363 // We must read the symbolsfile first, because the Encoding
364 // constructor depends on it.
365 LyXLex symbolslex(0, 0);
366 symbolslex.setFile(symbolsfile);
367 while (symbolslex.isOK()) {
372 if (symbolslex.next(true)) {
373 std::istringstream is(symbolslex.getString());
374 // reading symbol directly does not work if
375 // char_type == std::wchar_t.
377 if(!(is >> std::hex >> tmp))
382 if (symbolslex.next(true))
383 info.command = symbolslex.getDocString();
386 if (symbolslex.next(true))
387 info.preamble = symbolslex.getString();
390 if (symbolslex.next(true))
391 flags = symbolslex.getString();
395 info.combining = false;
397 while (!flags.empty()) {
399 flags = support::split(flags, flag, ',');
400 if (flag == "combining")
401 info.combining = true;
402 else if (flag == "force")
405 lyxerr << "Ignoring unknown flag `" << flag
406 << "' for symbol `0x" << std::hex
407 << symbol << "'." << endl;
410 if (!info.preamble.empty())
411 info.feature = info.preamble[0] != '\\';
414 << "Read unicode symbol " << symbol << " '"
415 << to_utf8(info.command) << "' '" << info.preamble
416 << "' " << info.combining << ' ' << info.feature
418 unicodesymbols[symbol] = info;
421 // Now read the encodings
428 struct keyword_item encodingtags[et_last - 1] = {
429 { "encoding", et_encoding },
433 LyXLex lex(encodingtags, et_last - 1);
434 lex.setFile(encfile);
440 string const name = lex.getString();
442 string const latexname = lex.getString();
444 string const iconvname = lex.getString();
445 lyxerr[Debug::INFO] << "Reading encoding " << name << endl;
446 encodinglist[name] = Encoding(name, latexname, iconvname);
447 if (lex.lex() != et_end)
448 lex.printError("Encodings::read: "
453 lex.printError("Encodings::read: Misplaced end");
455 case LyXLex::LEX_FEOF:
458 lex.printError("Encodings::read: "
459 "Unknown tag: `$$Token'");