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_table[50][4] = {
46 {0xfe80, 0xfe80, 0xfe80, 0xfe80}, // 0x0621 = hamza
47 {0xfe81, 0xfe82, 0xfe81, 0xfe82}, // 0x0622 = ligature madda on alef
48 {0xfe83, 0xfe84, 0xfe83, 0xfe84}, // 0x0623 = ligature hamza on alef
49 {0xfe85, 0xfe86, 0xfe85, 0xfe86}, // 0x0624 = ligature hamza on waw
50 {0xfe87, 0xfe88, 0xfe87, 0xfe88}, // 0x0625 = ligature hamza under alef
51 {0xfe89, 0xfe8a, 0xfe8b, 0xfe8c}, // 0x0626 = ligature hamza on ya
52 {0xfe8d, 0xfe8e, 0xfe8d, 0xfe8e}, // 0x0627 = alef
53 {0xfe8f, 0xfe90, 0xfe91, 0xfe92}, // 0x0628 = baa
54 {0xfe93, 0xfe94, 0xfe93, 0xfe94}, // 0x0629 = taa marbuta
55 {0xfe95, 0xfe96, 0xfe97, 0xfe98}, // 0x062a = taa
56 {0xfe99, 0xfe9a, 0xfe9b, 0xfe9c}, // 0x062b = thaa
57 {0xfe9d, 0xfe9e, 0xfe9f, 0xfea0}, // 0x062c = jeem
58 {0xfea1, 0xfea2, 0xfea3, 0xfea4}, // 0x062d = haa
59 {0xfea5, 0xfea6, 0xfea7, 0xfea8}, // 0x062e = khaa
60 {0xfea9, 0xfeaa, 0xfea9, 0xfeaa}, // 0x062f = dal
62 {0xfeab, 0xfeac, 0xfeab, 0xfeac}, // 0x0630 = thal
63 {0xfead, 0xfeae, 0xfead, 0xfeae}, // 0x0631 = ra
64 {0xfeaf, 0xfeb0, 0xfeaf, 0xfeb0}, // 0x0632 = zain
65 {0xfeb1, 0xfeb2, 0xfeb3, 0xfeb4}, // 0x0633 = seen
66 {0xfeb5, 0xfeb6, 0xfeb7, 0xfeb8}, // 0x0634 = sheen
67 {0xfeb9, 0xfeba, 0xfebb, 0xfebc}, // 0x0635 = sad
68 {0xfebd, 0xfebe, 0xfebf, 0xfec0}, // 0x0636 = dad
69 {0xfec1, 0xfec2, 0xfec3, 0xfec4}, // 0x0637 = tah
70 {0xfec5, 0xfec6, 0xfec7, 0xfec8}, // 0x0638 = zah
71 {0xfec9, 0xfeca, 0xfecb, 0xfecc}, // 0x0639 = ain
72 {0xfecd, 0xfece, 0xfecf, 0xfed0}, // 0x063a = ghain
73 {0, 0, 0, 0}, // 0x063b
74 {0, 0, 0, 0}, // 0x063c
75 {0, 0, 0, 0}, // 0x063d
76 {0, 0, 0, 0}, // 0x063e
77 {0, 0, 0, 0}, // 0x063f
79 {0, 0, 0, 0}, // 0x0640
80 {0xfed1, 0xfed2, 0xfed3, 0xfed4}, // 0x0641 = fa
81 {0xfed5, 0xfed6, 0xfed7, 0xfed8}, // 0x0642 = qaf
82 {0xfed9, 0xfeda, 0xfedb, 0xfedc}, // 0x0643 = kaf
83 {0xfedd, 0xfede, 0xfedf, 0xfee0}, // 0x0644 = lam
84 {0xfee1, 0xfee2, 0xfee3, 0xfee4}, // 0x0645 = meem
85 {0xfee5, 0xfee6, 0xfee7, 0xfee8}, // 0x0646 = noon
86 {0xfee9, 0xfeea, 0xfeeb, 0xfeec}, // 0x0647 = ha
87 {0xfeed, 0xfeee, 0xfeed, 0xfeee}, // 0x0648 = waw
88 {0xfeef, 0xfef0, 0xfeef, 0xfef0}, // 0x0649 = alef maksura
89 {0xfef1, 0xfef2, 0xfef3, 0xfef4}, // 0x064a = ya
90 {0x065b, 0x065b, 0x065b, 0x065b}, // 0x064b = fathatan
91 {0x065c, 0x065c, 0x065c, 0x065c}, // 0x064c = dammatan
92 {0x064d, 0x064d, 0x064d, 0x064d}, // 0x064d = kasratan
93 {0x064e, 0x064e, 0x064e, 0x064e}, // 0x064e = fatha
94 {0x064f, 0x064f, 0x064f, 0x064f}, // 0x064f = damma
96 {0x0650, 0x0650, 0x0650, 0x0650}, // 0x0650 = kasra
97 {0x0651, 0x0651, 0x0651, 0x0651}, // 0x0651 = shadda
98 {0x0652, 0x0652, 0x0652, 0x0652}, // 0x0652 = sukun
102 char_type const arabic_start = 0x0621;
103 char_type const arabic_end = 0x0652;
106 /// Information about a single UCS4 character
108 /// LaTeX command for this character
110 /// Needed LaTeX preamble (or feature)
112 /// Is this a combining character?
114 /// Is \c preamble a feature known by LaTeXFeatures, or a raw LaTeX
117 /// Always force the LaTeX command, even if the encoding contains
123 typedef std::map<char_type, CharInfo> CharInfoMap;
124 CharInfoMap unicodesymbols;
127 /// The highest code point in UCS4 encoding (1<<20 + 1<<16)
128 char_type const max_ucs4 = 0x110000;
133 Encoding::Encoding(string const & n, string const & l, string const & i,
134 bool f, Encoding::Package p)
135 : Name_(n), LatexName_(l), iconvName_(i), fixedwidth_(f), package_(p)
138 // ASCII can encode 128 code points and nothing else
139 start_encodable_ = 128;
141 } else if (i == "UTF-8") {
142 // UTF8 can encode all UCS4 code points
143 start_encodable_ = max_ucs4;
151 void Encoding::init() const
153 start_encodable_ = 0;
154 // temporarily switch off lyxerr, since we will generate iconv errors
157 // We do not need to check all UCS4 code points, it is enough
158 // if we check all 256 code points of this encoding.
159 for (unsigned short j = 0; j < 256; ++j) {
161 std::vector<char_type> const ucs4 = eightbit_to_ucs4(&c, 1, iconvName_);
162 if (ucs4.size() == 1) {
163 char_type const c = ucs4[0];
164 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
165 if (it == unicodesymbols.end() || !it->second.force)
166 encodable_.insert(c);
170 // We do not know how many code points this encoding has, and
171 // they do not have a direct representation as a single byte,
172 // therefore we need to check all UCS4 code points.
173 // This is expensive!
174 for (char_type c = 0; c < max_ucs4; ++c) {
175 std::vector<char> const eightbit = ucs4_to_eightbit(&c, 1, iconvName_);
176 if (!eightbit.empty()) {
177 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
178 if (it == unicodesymbols.end() || !it->second.force)
179 encodable_.insert(c);
184 CharSet::iterator it = encodable_.find(start_encodable_);
185 while (it != encodable_.end()) {
186 encodable_.erase(it);
188 it = encodable_.find(start_encodable_);
194 docstring const Encoding::latexChar(char_type c) const
196 // validate() should have been called before
197 //BOOST_ASSERT(complete_);
199 if (c < start_encodable_)
200 return docstring(1, c);
201 if (encodable_.find(c) == encodable_.end()) {
202 // c cannot be encoded in this encoding
203 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
204 if (it == unicodesymbols.end())
205 lyxerr << "Could not find LaTeX command for character 0x"
206 << std::hex << c << std::dec
207 << ".\nLaTeX export will fail."
210 return it->second.command;
212 return docstring(1, c);
216 void Encodings::validate(char_type c, LaTeXFeatures & features)
218 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
219 if (it != unicodesymbols.end() && !it->second.preamble.empty()) {
220 if (it->second.feature)
221 features.require(it->second.preamble);
223 features.addPreambleSnippet(it->second.preamble);
228 bool Encodings::isComposeChar_hebrew(char_type c)
230 return c <= 0x05c2 && c >= 0x05b0 &&
231 c != 0x05be && c != 0x05c0;
235 // Special Arabic letters are ones that do not get connected from left
236 // they are hamza, alef_madda, alef_hamza, waw_hamza, alef_hamza_under,
237 // alef, tah_marbota, dal, thal, rah, zai, wow, alef_maksoura
239 bool Encodings::is_arabic_special(char_type c)
241 return (c >= 0x0621 && c <= 0x0625) ||
242 c == 0x0627 || c == 0x0629 ||
243 c == 0x062f || c == 0x0648 ||
244 (c >= 0x0630 && c <= 0x0632) ||
249 bool Encodings::isComposeChar_arabic(char_type c)
251 return c >= 0x064b && c <= 0x0652;
255 bool Encodings::is_arabic(char_type c)
257 return c >= arabic_start && c <= arabic_end &&
258 arabic_table[c-arabic_start][0];
262 char_type Encodings::transformChar(char_type c,
263 Encodings::Letter_Form form)
268 return arabic_table[c-arabic_start][form];
272 bool Encodings::isCombiningChar(char_type c)
274 CharInfoMap::const_iterator const it = unicodesymbols.find(c);
275 if (it != unicodesymbols.end())
276 return it->second.combining;
281 Encoding const * Encodings::getFromLyXName(string const & name) const
283 EncodingList::const_iterator it = encodinglist.find(name);
284 if (it != encodinglist.end())
291 Encoding const * Encodings::getFromLaTeXName(string const & name) const
293 // We don't use std::find_if because it makes copies of the pairs in
295 // This linear search is OK since we don't have many encodings.
296 // Users could even optimize it by putting the encodings they use
297 // most at the top of lib/encodings.
298 EncodingList::const_iterator const end = encodinglist.end();
299 for (EncodingList::const_iterator it = encodinglist.begin(); it != end; ++it)
300 if (it->second.latexName() == name)
306 Encodings::Encodings()
311 void Encodings::read(FileName const & encfile, FileName const & symbolsfile)
313 // We must read the symbolsfile first, because the Encoding
314 // constructor depends on it.
315 Lexer symbolslex(0, 0);
316 symbolslex.setFile(symbolsfile);
317 while (symbolslex.isOK()) {
322 if (symbolslex.next(true)) {
323 std::istringstream is(symbolslex.getString());
324 // reading symbol directly does not work if
325 // char_type == std::wchar_t.
327 if(!(is >> std::hex >> tmp))
332 if (symbolslex.next(true))
333 info.command = symbolslex.getDocString();
336 if (symbolslex.next(true))
337 info.preamble = symbolslex.getString();
340 if (symbolslex.next(true))
341 flags = symbolslex.getString();
345 info.combining = false;
346 info.feature = false;
348 while (!flags.empty()) {
350 flags = support::split(flags, flag, ',');
351 if (flag == "combining")
352 info.combining = true;
353 else if (flag == "force")
356 lyxerr << "Ignoring unknown flag `" << flag
357 << "' for symbol `0x"
358 << std::hex << symbol << std::dec
362 if (!info.preamble.empty())
363 info.feature = info.preamble[0] != '\\';
366 << "Read unicode symbol " << symbol << " '"
367 << to_utf8(info.command) << "' '" << info.preamble
368 << "' " << info.combining << ' ' << info.feature
370 unicodesymbols[symbol] = info;
373 // Now read the encodings
380 struct keyword_item encodingtags[et_last - 1] = {
381 { "encoding", et_encoding },
385 Lexer lex(encodingtags, et_last - 1);
386 lex.setFile(encfile);
392 string const name = lex.getString();
394 string const latexname = lex.getString();
396 string const iconvname = lex.getString();
398 string const width = lex.getString();
400 if (width == "fixed")
402 else if (width == "variable")
405 lex.printError("Encodings::read: "
406 "Unknown width: `$$Token'");
408 string const p = lex.getString();
409 Encoding::Package package;
411 package = Encoding::none;
412 else if (p == "inputenc")
413 package = Encoding::inputenc;
415 package = Encoding::CJK;
417 lex.printError("Encodings::read: "
418 "Unknown package: `$$Token'");
419 LYXERR(Debug::INFO) << "Reading encoding " << name << endl;
420 encodinglist[name] = Encoding(name, latexname,
421 iconvname, fixedwidth,
423 if (lex.lex() != et_end)
424 lex.printError("Encodings::read: "
429 lex.printError("Encodings::read: Misplaced end");
431 case Lexer::LEX_FEOF:
434 lex.printError("Encodings::read: "
435 "Unknown tag: `$$Token'");