4 * This file is part of LyX, the document processor.
5 * Licence details can be found in the file COPYING.
7 * \author Lars Gullik Bjønnes
8 * \author Jean-Marc Lasgouttes
10 * Full author contact details are available in file CREDITS.
16 #include "support/docstring.h"
17 #include "support/types.h"
25 namespace support { class FileName; }
30 class EncodingException : public std::exception {
32 EncodingException(char_type c);
33 virtual ~EncodingException() throw() {}
34 virtual const char * what() const throw();
36 char_type failed_char;
45 /// Which LaTeX package handles this encoding?
55 Encoding(std::string const & n, std::string const & l,
56 std::string const & g, std::string const & i,
57 bool f, bool u, Package p);
61 std::string const & name() const { return name_; }
63 std::string const & latexName() const { return latexName_; }
65 std::string const & guiName() const { return guiName_; }
67 std::string const & iconvName() const { return iconvName_; }
69 bool hasFixedWidth() const { return fixedwidth_; }
71 bool unsafe() const { return unsafe_; }
72 /// \p c is representable in this encoding without a LaTeX macro
73 bool encodable(char_type c) const;
75 * Convert \p c to something that LaTeX can understand.
76 * This is either the character itself (if it is representable
77 * in this encoding), or a LaTeX macro.
78 * If the character is not representable in this encoding, but no
79 * LaTeX macro is known, a warning is given of lyxerr, and the
80 * character is returned.
81 * \return the converted character and a flag indicating whether
82 * the command needs to be terminated by {} or a space.
84 std::pair<docstring, bool> latexChar(char_type c) const;
86 * Convert \p input to something that LaTeX can understand.
87 * This is either the string itself (if it is representable
88 * in this encoding), or a LaTeX macro.
89 * If a character is not representable in this encoding, but no
90 * LaTeX macro is known, a warning is given of lyxerr, and the
91 * character is returned in the second string of the pair and
92 * omitted in the first.
93 * \p dryrun specifies whether the string is used within source
94 * preview (which yields a special warning).
96 std::pair<docstring, docstring> latexString(docstring const input,
97 bool dryrun = false) const;
98 /// Which LaTeX package handles this encoding?
99 Package package() const { return package_; }
100 /// A list of all characters usable in this encoding
101 std::vector<char_type> symbolsList() const;
104 * Do we have to output this character as LaTeX command in any case?
105 * This is true if the "force" flag is set.
106 * We need this if the inputencoding does not support a certain glyph.
108 bool isForced(char_type c) const;
112 std::string latexName_;
114 std::string guiName_;
116 std::string iconvName_;
117 /// Is this a fixed width encoding?
119 /// Is this encoding TeX unsafe, e.g. control characters like {, }
120 /// and \\ may appear in high bytes?
123 typedef std::set<char_type> CharSet;
124 /// Set of UCS4 characters that we can encode (for singlebyte
126 mutable CharSet encodable_;
127 /// Set of UCS4 characters that we can't encode
128 CharSet const * forced_;
129 /// All code points below this are encodable. This helps us to avoid
130 /// lokup of ASCII characters in encodable_ and gives about 1 sec
131 /// speedup on export of the Userguide.
132 mutable char_type start_encodable_;
133 /// Which LaTeX package handles this encoding?
136 * If this is true the stored information about the encoding covers
137 * all encodable characters. We set this to false initially so that
138 * we only need to query iconv for the actually used encodings.
139 * This is needed especially for the multibyte encodings, if we
140 * complete all encoding info on startup it takes 2-3 minutes.
142 mutable bool complete_;
148 typedef std::set<char_type> MathCommandSet;
150 typedef std::set<char_type> TextCommandSet;
152 typedef std::set<char_type> MathSymbolSet;
154 typedef std::map<std::string, Encoding> EncodingList;
155 /// iterator to iterate over all encodings.
156 /// We hide the fact that our encoding list is implemented as a map.
157 class const_iterator : public EncodingList::const_iterator {
158 typedef EncodingList::const_iterator base;
160 const_iterator() : base() {}
161 const_iterator(base const & b) : base(b) {}
162 Encoding const & operator*() const { return base::operator*().second; }
163 Encoding const * operator->() const { return &(base::operator*().second); }
167 /// Read the encodings.
168 /// \param encfile encodings definition file
169 /// \param symbolsfile unicode->LaTeX mapping file
170 void read(support::FileName const & encfile,
171 support::FileName const & symbolsfile);
172 /// Get encoding from LyX name \p name
174 fromLyXName(std::string const & name, bool allowUnsafe = false) const;
175 /// Get encoding from LaTeX name \p name
177 fromLaTeXName(std::string const & name, bool allowUnsafe = false) const;
180 const_iterator begin() const { return encodinglist.begin(); }
182 const_iterator end() const { return encodinglist.end(); }
196 static bool isHebrewComposeChar(char_type c);
198 static bool isArabicComposeChar(char_type c);
200 static bool isArabicSpecialChar(char_type c);
202 static bool isArabicChar(char_type c);
204 static char_type transformChar(char_type c, LetterForm form);
205 /// Is this a combining char?
206 static bool isCombiningChar(char_type c);
207 /// Return the TIPA shortcut
208 static std::string const TIPAShortcut(char_type c);
210 * Is this a known char from some language?
211 * If \p preamble is empty and code point \p c is known to belong
212 * to a supported script, true is returned and \p preamble is set
213 * to the corresponding entry in the unicodesymbols file.
214 * If \p preamble is not empty, a check is made whether code point
215 * \p c is a known character matching the preamble entry.
217 static bool isKnownScriptChar(char_type const c, std::string & preamble);
219 * Do we have to display in italics this character when in mathmode?
220 * This is true if the "mathalpha" flag is set. We use this for
221 * letters and accented characters that are output as math commands.
223 static bool isMathAlpha(char_type c);
225 * Register \p c as a mathmode command.
227 static void addMathCmd(char_type c) { mathcmd.insert(c); }
229 * Register \p c as a textmode command.
231 static void addTextCmd(char_type c) { textcmd.insert(c); }
233 * Register \p c as a mathmode symbol.
235 static void addMathSym(char_type c) { mathsym.insert(c); }
237 * Tell whether \p c is registered as a mathmode command.
239 static bool isMathCmd(char_type c) { return mathcmd.count(c); }
241 * Tell whether \p c is registered as a textmode command.
243 static bool isTextCmd(char_type c) { return textcmd.count(c); }
245 * Tell whether \p c is registered as a mathmode symbol.
247 static bool isMathSym(char_type c) { return mathsym.count(c); }
249 * Initialize mathcmd, textcmd, and mathsym sets.
251 static void initUnicodeMath(Buffer const & buffer, bool for_master = true);
253 * If \p c cannot be encoded in the given \p encoding, convert
254 * it to something that LaTeX can understand in mathmode.
255 * \p needsTermination indicates whether the command needs to be
256 * terminated by {} or a space.
257 * \return whether \p command is a mathmode command
259 static bool latexMathChar(char_type c, bool mathmode,
260 Encoding const * encoding, docstring & command,
261 bool & needsTermination);
263 * Convert the LaTeX command in \p cmd to the corresponding unicode
264 * point and set \p combining to true if it is a combining symbol.
265 * \p needsTermination indicates whether the command needs to be
266 * terminated by {} or a space.
268 static char_type fromLaTeXCommand(docstring const & cmd, int cmdtype,
269 bool & combining, bool & needsTermination,
270 std::set<std::string> * req = 0);
279 * Convert the LaTeX commands in \p cmd and \return a docstring
280 * of corresponding unicode points. The conversion stops at the
281 * first command which could not be converted, and the remaining
282 * unconverted commands are returned in \p rem.
283 * The \p cmdtype parameter can be used to limit recognized
284 * commands to math or text mode commands only.
285 * \p needsTermination indicates whether the command needs to be
286 * terminated by {} or a space.
288 static docstring fromLaTeXCommand(docstring const & cmd, int cmdtype,
289 bool & needsTermination, docstring & rem,
290 std::set<std::string> * req = 0);
292 * Add the preamble snippet needed for the output of \p c to
294 * This does not depend on the used encoding, since the inputenc
295 * package only maps the code point \p c to a command, it does not
296 * make this command available.
298 static void validate(char_type c, LaTeXFeatures & features, bool for_mathed = false);
302 EncodingList encodinglist;
304 static MathCommandSet mathcmd;
306 static TextCommandSet textcmd;
308 static MathSymbolSet mathsym;
311 extern Encodings encodings;