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"
31 using support::FileName;
33 #ifndef CXX_GLOBAL_CSTD
45 char_type arabic_table2[63][4] = {
46 {0x41, 0x41, 0x41, 0x41}, // 0xc1 = hamza
47 {0x42, 0xa1, 0x42, 0xa1}, // 0xc2 = ligature madda on alef
48 {0x43, 0xa2, 0x43, 0xa2}, // 0xc3 = ligature hamza on alef
49 {0x44, 0xa3, 0x44, 0xa3}, // 0xc4 = ligature hamza on waw
50 {0x45, 0xa4, 0x45, 0xa4}, // 0xc5 = ligature hamza under alef
51 {0x46, 0xf9, 0xf8, 0xa0}, // 0xc6 = ligature hamza on ya
52 {0x47, 0xa5, 0x47, 0xa5}, // 0xc7 = alef
53 {0x48, 0xae, 0xac, 0xad}, // 0xc8 = baa
54 {0x49, 0xb1, 0x49, 0xb1}, // 0xc9 = taa marbuta
55 {0x4a, 0xb4, 0xb2, 0xb3}, // 0xca = taa
56 {0x4b, 0xb7, 0xb5, 0xb6}, // 0xcb = thaa
57 {0x4c, 0xba, 0xb8, 0xb9}, // 0xcc = jeem
58 {0x4d, 0xbd, 0xbb, 0xbc}, // 0xcd = haa
59 {0x4e, 0xc0, 0xbe, 0xbf}, // 0xce = khaa
60 {0x4f, 0xa6, 0x4f, 0xa6}, // 0xcf = dal
62 {0x50, 0xa7, 0x50, 0xa7}, // 0xd0 = thal
63 {0x51, 0xa8, 0x51, 0xa8}, // 0xd1 = ra
64 {0x52, 0xa9, 0x52, 0xa9}, // 0xd2 = zain
65 {0x53, 0xc3, 0xc1, 0xc2}, // 0xd3 = seen
66 {0x54, 0xc6, 0xc4, 0xc5}, // 0xd4 = sheen
67 {0x55, 0xc9, 0xc7, 0xc8}, // 0xd5 = sad
68 {0x56, 0xcc, 0xca, 0xcb}, // 0xd6 = dad
69 {0x57, 0xcf, 0xcd, 0xce}, // 0xd7 = tah
70 {0x58, 0xd2, 0xd0, 0xd1}, // 0xd8 = zah
71 {0x59, 0xd5, 0xd3, 0xd4}, // 0xd9 = ain
72 {0x5a, 0xd8, 0xd6, 0xd7}, // 0xda = ghain
80 {0x61, 0xdb, 0xd9, 0xda}, // 0xe1 = fa
81 {0x62, 0xde, 0xdc, 0xdd}, // 0xe2 = qaf
82 {0x63, 0xe1, 0xdf, 0xe0}, // 0xe3 = kaf
83 {0x64, 0xe4, 0xe2, 0xe3}, // 0xe4 = lam
84 {0x65, 0xe7, 0xe5, 0xe6}, // 0xe5 = meem
85 {0x66, 0xea, 0xe8, 0xe9}, // 0xe6 = noon
86 {0x67, 0xed, 0xeb, 0xec}, // 0xe7 = ha
87 {0x68, 0xaa, 0x68, 0xaa}, // 0xe8 = waw
88 {0x69, 0xab, 0x69, 0xab}, // 0xe9 = alef maksura
89 {0x6a, 0xf0, 0xee, 0xef}, // 0xea = ya
90 {0x6b, 0x6b, 0x6b, 0x6b}, // 0xeb = fathatan
91 {0x6c, 0x6c, 0x6c, 0x6c}, // 0xec = dammatan
92 {0x6d, 0x6d, 0x6d, 0x6d}, // 0xed = kasratan
93 {0x6e, 0x6e, 0x6e, 0x6e}, // 0xee = fatha
94 {0x6f, 0x6f, 0x6f, 0x6f}, // 0xef = damma
96 {0x70, 0x70, 0x70, 0x70}, // 0xf0 = kasra
97 {0x71, 0x71, 0x71, 0x71}, // 0xf1 = shadda
98 {0x72, 0x72, 0x72, 0x72}, // 0xf2 = sukun
115 char_type arabic_table[63][2] = {
116 {0xc1, 0xc1}, // 0xc1 = hamza
117 {0xc2, 0xc2}, // 0xc2 = ligature madda on alef
118 {0xc3, 0xc3}, // 0xc3 = ligature hamza on alef
119 {0xc4, 0xc4}, // 0xc4 = ligature hamza on waw
120 {0xc5, 0xc5}, // 0xc5 = ligature hamza under alef
121 {0xc6, 0xc0}, // 0xc6 = ligature hamza on ya
122 {0xc7, 0xc7}, // 0xc7 = alef
123 {0xc8, 0xeb}, // 0xc8 = baa
124 {0xc9, 0xc9}, // 0xc9 = taa marbuta
125 {0xca, 0xec}, // 0xca = taa
126 {0xcb, 0xed}, // 0xcb = thaa
127 {0xcc, 0xee}, // 0xcc = jeem
128 {0xcd, 0xef}, // 0xcd = haa
129 {0xce, 0xf0}, // 0xce = khaa
130 {0xcf, 0xcf}, // 0xcf = dal
132 {0xd0, 0xd0}, // 0xd0 = thal
133 {0xd1, 0xd1}, // 0xd1 = ra
134 {0xd2, 0xd2}, // 0xd2 = zain
135 {0xd3, 0xf1}, // 0xd3 = seen
136 {0xd4, 0xf2}, // 0xd4 = sheen
137 {0xd5, 0xf3}, // 0xd5 = sad
138 {0xd6, 0xf4}, // 0xd6 = dad
139 {0xd7, 0xd7}, // 0xd7 = tah
140 {0xd8, 0xd8}, // 0xd8 = zah
141 {0xd9, 0xf5}, // 0xd9 = ain
142 {0xda, 0xf6}, // 0xda = ghain
150 {0xe1, 0xf7}, // 0xe1 = fa
151 {0xe2, 0xf8}, // 0xe2 = qaf
152 {0xe3, 0xf9}, // 0xe3 = kaf
153 {0xe4, 0xfa}, // 0xe4 = lam
154 {0xe5, 0xfb}, // 0xe5 = meem
155 {0xe6, 0xfc}, // 0xe6 = noon
156 {0xe7, 0xfd}, // 0xe7 = ha
157 {0xe8, 0xe8}, // 0xe8 = waw
158 {0xe9, 0xe9}, // 0xe9 = alef maksura
159 {0xea, 0xfe}, // 0xea = ya
160 {0xa8, 0xa8}, // 0xeb = fathatan
161 {0xa9, 0xa9}, // 0xec = dammatan
162 {0xaa, 0xaa}, // 0xed = kasratan
163 {0xab, 0xab}, // 0xee = fatha
164 {0xac, 0xac}, // 0xef = damma
166 {0xad, 0xad}, // 0xf0 = kasra
167 {0xae, 0xae}, // 0xf1 = shadda
168 {0xaf, 0xaf}, // 0xf2 = sukun
185 char_type const arabic_start = 0xc1;
188 /// Information about a single UCS4 character
190 /// LaTeX command for this character
192 /// Needed LaTeX preamble (or feature)
194 /// Is this a combining character?
196 /// Is \c preamble a feature known by LaTeXFeatures, or a raw LaTeX
199 /// Always force the LaTeX command, even if the encoding contains
205 typedef std::map<char_type, CharInfo> CharInfoMap;
206 CharInfoMap unicodesymbols;
211 Encoding::Encoding(string const & n, string const & l, string const & i)
212 : Name_(n), LatexName_(l), iconvName_(i)
215 // ASCII can encode 128 code points and nothing else
216 start_encodable_ = 128;
217 else if (i == "UTF-8")
218 // UTF8 can encode all 1<<20 + 1<<16 UCS4 code points
219 start_encodable_ = 0x110000;
221 start_encodable_ = 0;
222 // temporarily switch off lyxerr, since we will generate iconv errors
224 for (unsigned short j = 0; j < 256; ++j) {
226 std::vector<char_type> const ucs4 = eightbit_to_ucs4(&c, 1, i);
227 if (ucs4.size() == 1) {
228 char_type const c = ucs4[0];
229 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
230 if (it == unicodesymbols.end() || !it->second.force)
231 encodable_.insert(c);
235 CharSet::iterator it = encodable_.find(start_encodable_);
236 while (it != encodable_.end()) {
237 encodable_.erase(it);
239 it = encodable_.find(start_encodable_);
245 docstring const Encoding::latexChar(char_type c) const
247 if (c < start_encodable_)
248 return docstring(1, c);
249 if (encodable_.find(c) == encodable_.end()) {
250 // c cannot be encoded in this encoding
251 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
252 if (it == unicodesymbols.end())
253 lyxerr << "Could not find LaTeX command for character 0x"
254 << std::hex << c << ".\nLaTeX export will fail."
257 return it->second.command;
259 return docstring(1, c);
263 void Encoding::validate(char_type c, LaTeXFeatures & features) const
265 // Add the preamble stuff even if c can be encoded in this encoding,
266 // since the inputenc package only maps the code point c to a command,
267 // it does not make this command available.
268 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
269 if (it != unicodesymbols.end() && !it->second.preamble.empty()) {
270 if (it->second.feature)
271 features.require(it->second.preamble);
273 features.addPreambleSnippet(it->second.preamble);
278 bool Encodings::isComposeChar_hebrew(char_type c)
280 return c <= 0xd2 && c >= 0xc0 &&
281 c != 0xce && c != 0xd0;
285 // Special Arabic letters are ones that do not get connected from left
286 // they are hamza, alef_madda, alef_hamza, waw_hamza, alef_hamza_under,
287 // alef, tah_marbota, dal, thal, rah, zai, wow, alef_maksoura
289 bool Encodings::is_arabic_special(char_type c)
291 return (c >= 0xc1 && c <= 0xc5) ||
292 c == 0xc7 || c == 0xc9 ||
293 c == 0xcf || c == 0xe8 ||
294 (c >= 0xd0 && c <= 0xd2) ||
298 bool Encodings::isComposeChar_arabic(char_type c)
300 return c >= 0xeb && c <= 0xf2;
304 bool Encodings::is_arabic(char_type c)
306 return c >= arabic_start && arabic_table[c-arabic_start][0];
310 char_type Encodings::transformChar(char_type c,
311 Encodings::Letter_Form form)
316 if (lyxrc.font_norm_type == LyXRC::ISO_10646_1)
317 return arabic_table2[c-arabic_start][form];
319 return arabic_table[c-arabic_start][form >> 1];
323 bool Encodings::isCombiningChar(char_type c)
325 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
326 if (it != unicodesymbols.end())
327 return it->second.combining;
332 Encoding const * Encodings::getFromLyXName(string const & name) const
334 EncodingList::const_iterator it = encodinglist.find(name);
335 if (it != encodinglist.end())
342 Encoding const * Encodings::getFromLaTeXName(string const & name) const
344 // We don't use std::find_if because it makes copies of the pairs in
346 // This linear search is OK since we don't have many encodings.
347 // Users could even optimize it by putting the encodings they use
348 // most at the top of lib/encodings.
349 EncodingList::const_iterator const end = encodinglist.end();
350 for (EncodingList::const_iterator it = encodinglist.begin(); it != end; ++it)
351 if (it->second.latexName() == name)
357 Encodings::Encodings()
362 void Encodings::read(FileName const & encfile, FileName const & symbolsfile)
364 // We must read the symbolsfile first, because the Encoding
365 // constructor depends on it.
366 LyXLex symbolslex(0, 0);
367 symbolslex.setFile(symbolsfile);
368 while (symbolslex.isOK()) {
373 if (symbolslex.next(true)) {
374 std::istringstream is(symbolslex.getString());
375 // reading symbol directly does not work if
376 // char_type == std::wchar_t.
378 if(!(is >> std::hex >> tmp))
383 if (symbolslex.next(true))
384 info.command = symbolslex.getDocString();
387 if (symbolslex.next(true))
388 info.preamble = symbolslex.getString();
391 if (symbolslex.next(true))
392 flags = symbolslex.getString();
396 info.combining = false;
397 info.feature = false;
399 while (!flags.empty()) {
401 flags = support::split(flags, flag, ',');
402 if (flag == "combining")
403 info.combining = true;
404 else if (flag == "force")
407 lyxerr << "Ignoring unknown flag `" << flag
408 << "' for symbol `0x" << std::hex
409 << symbol << "'." << endl;
412 if (!info.preamble.empty())
413 info.feature = info.preamble[0] != '\\';
416 << "Read unicode symbol " << symbol << " '"
417 << to_utf8(info.command) << "' '" << info.preamble
418 << "' " << info.combining << ' ' << info.feature
420 unicodesymbols[symbol] = info;
423 // Now read the encodings
430 struct keyword_item encodingtags[et_last - 1] = {
431 { "encoding", et_encoding },
435 LyXLex lex(encodingtags, et_last - 1);
436 lex.setFile(encfile);
442 string const name = lex.getString();
444 string const latexname = lex.getString();
446 string const iconvname = lex.getString();
447 lyxerr[Debug::INFO] << "Reading encoding " << name << endl;
448 encodinglist[name] = Encoding(name, latexname, iconvname);
449 if (lex.lex() != et_end)
450 lex.printError("Encodings::read: "
455 lex.printError("Encodings::read: Misplaced end");
457 case LyXLex::LEX_FEOF:
460 lex.printError("Encodings::read: "
461 "Unknown tag: `$$Token'");