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?
52 /// Represent any of the above packages
57 Encoding(std::string const & n, std::string const & l,
58 std::string const & g, std::string const & i,
59 bool f, bool u, Package p);
63 std::string const & name() const { return name_; }
65 std::string const & latexName() const { return latexName_; }
67 std::string const & guiName() const { return guiName_; }
69 std::string const & iconvName() const { return iconvName_; }
71 bool hasFixedWidth() const { return fixedwidth_; }
73 bool unsafe() const { return unsafe_; }
74 /// \p c is representable in this encoding without a LaTeX macro
75 bool encodable(char_type c) const;
77 * Convert \p c to something that LaTeX can understand.
78 * This is either the character itself (if it is representable
79 * in this encoding), or a LaTeX macro.
80 * If the character is not representable in this encoding, but no
81 * LaTeX macro is known, a warning is given of lyxerr, and the
82 * character is returned.
83 * \return the converted character and a flag indicating whether
84 * the command needs to be terminated by {} or a space.
86 std::pair<docstring, bool> latexChar(char_type c) const;
88 * Convert \p input to something that LaTeX can understand.
89 * This is either the string itself (if it is representable
90 * in this encoding), or a LaTeX macro.
91 * If a character is not representable in this encoding, but no
92 * LaTeX macro is known, a warning is given of lyxerr, and the
93 * character is returned in the second string of the pair and
94 * omitted in the first.
95 * \p dryrun specifies whether the string is used within source
96 * preview (which yields a special warning).
98 std::pair<docstring, docstring> latexString(docstring const input,
99 bool dryrun = false) const;
100 /// Which LaTeX package handles this encoding?
101 Package package() const { return package_; }
102 /// A list of all characters usable in this encoding
103 std::vector<char_type> symbolsList() const;
106 * Do we have to output this character as LaTeX command in any case?
107 * This is true if the "force" flag is set.
108 * We need this if the inputencoding does not support a certain glyph.
110 bool isForced(char_type c) const;
114 std::string latexName_;
116 std::string guiName_;
118 std::string iconvName_;
119 /// Is this a fixed width encoding?
121 /// Is this encoding TeX unsafe, e.g. control characters like {, }
122 /// and \\ may appear in high bytes?
125 typedef std::set<char_type> CharSet;
126 /// Set of UCS4 characters that we can encode (for singlebyte
128 mutable CharSet encodable_;
129 /// Set of UCS4 characters that we can't encode
130 CharSet const * forced_;
131 /// All code points below this are encodable. This helps us to avoid
132 /// lokup of ASCII characters in encodable_ and gives about 1 sec
133 /// speedup on export of the Userguide.
134 mutable char_type start_encodable_;
135 /// Which LaTeX package handles this encoding?
138 * If this is true the stored information about the encoding covers
139 * all encodable characters. We set this to false initially so that
140 * we only need to query iconv for the actually used encodings.
141 * This is needed especially for the multibyte encodings, if we
142 * complete all encoding info on startup it takes 2-3 minutes.
144 mutable bool complete_;
150 typedef std::set<char_type> MathCommandSet;
152 typedef std::set<char_type> TextCommandSet;
154 typedef std::set<char_type> MathSymbolSet;
156 typedef std::map<std::string, Encoding> EncodingList;
157 /// iterator to iterate over all encodings.
158 /// We hide the fact that our encoding list is implemented as a map.
159 class const_iterator : public EncodingList::const_iterator {
160 typedef EncodingList::const_iterator base;
162 const_iterator() : base() {}
163 const_iterator(base const & b) : base(b) {}
164 Encoding const & operator*() const { return base::operator*().second; }
165 Encoding const * operator->() const { return &(base::operator*().second); }
169 /// Read the encodings.
170 /// \param encfile encodings definition file
171 /// \param symbolsfile unicode->LaTeX mapping file
172 void read(support::FileName const & encfile,
173 support::FileName const & symbolsfile);
174 /// Get encoding from LyX name \p name
176 fromLyXName(std::string const & name, bool allowUnsafe = false) const;
177 /// Get encoding from LaTeX name \p name and package \p package
178 Encoding const * fromLaTeXName(std::string const & name,
179 int const & package = Encoding::any, bool allowUnsafe = false) const;
180 /// Get encoding from iconv name \p name and package \p package
181 Encoding const * fromIconvName(std::string const & name,
182 int const & package = Encoding::any, bool allowUnsafe = false) const;
185 const_iterator begin() const { return encodinglist.begin(); }
187 const_iterator end() const { return encodinglist.end(); }
201 static bool isHebrewComposeChar(char_type c);
203 static bool isArabicComposeChar(char_type c);
205 static bool isArabicSpecialChar(char_type c);
207 static bool isArabicChar(char_type c);
209 static char_type transformChar(char_type c, LetterForm form);
210 /// Is this a combining char?
211 static bool isCombiningChar(char_type c);
212 /// Return the TIPA shortcut
213 static std::string const TIPAShortcut(char_type c);
215 * Is this a known char from some language?
216 * If \p preamble is empty and code point \p c is known to belong
217 * to a supported script, true is returned and \p preamble is set
218 * to the corresponding entry in the unicodesymbols file.
219 * If \p preamble is not empty, a check is made whether code point
220 * \p c is a known character matching the preamble entry.
222 static bool isKnownScriptChar(char_type const c, std::string & preamble);
224 * Do we have to display in italics this character when in mathmode?
225 * This is true if the "mathalpha" flag is set. We use this for
226 * letters and accented characters that are output as math commands.
228 static bool isMathAlpha(char_type c);
230 * Register \p c as a mathmode command.
232 static void addMathCmd(char_type c) { mathcmd.insert(c); }
234 * Register \p c as a textmode command.
236 static void addTextCmd(char_type c) { textcmd.insert(c); }
238 * Register \p c as a mathmode symbol.
240 static void addMathSym(char_type c) { mathsym.insert(c); }
242 * Tell whether \p c is registered as a mathmode command.
244 static bool isMathCmd(char_type c) { return mathcmd.count(c); }
246 * Tell whether \p c is registered as a textmode command.
248 static bool isTextCmd(char_type c) { return textcmd.count(c); }
250 * Tell whether \p c is registered as a mathmode symbol.
252 static bool isMathSym(char_type c) { return mathsym.count(c); }
254 * Initialize mathcmd, textcmd, and mathsym sets.
256 static void initUnicodeMath(Buffer const & buffer, bool for_master = true);
258 * If \p c cannot be encoded in the given \p encoding, convert
259 * it to something that LaTeX can understand in mathmode.
260 * \p needsTermination indicates whether the command needs to be
261 * terminated by {} or a space.
262 * \return whether \p command is a mathmode command
264 static bool latexMathChar(char_type c, bool mathmode,
265 Encoding const * encoding, docstring & command,
266 bool & needsTermination);
268 * Convert the LaTeX command in \p cmd to the corresponding unicode
269 * point and set \p combining to true if it is a combining symbol.
270 * \p needsTermination indicates whether the command needs to be
271 * terminated by {} or a space.
273 static char_type fromLaTeXCommand(docstring const & cmd, int cmdtype,
274 bool & combining, bool & needsTermination,
275 std::set<std::string> * req = 0);
284 * Convert the LaTeX commands in \p cmd and \return a docstring
285 * of corresponding unicode points. The conversion stops at the
286 * first command which could not be converted, and the remaining
287 * unconverted commands are returned in \p rem.
288 * The \p cmdtype parameter can be used to limit recognized
289 * commands to math or text mode commands only.
290 * \p needsTermination indicates whether the command needs to be
291 * terminated by {} or a space.
293 static docstring fromLaTeXCommand(docstring const & cmd, int cmdtype,
294 bool & needsTermination, docstring & rem,
295 std::set<std::string> * req = 0);
297 * Add the preamble snippet needed for the output of \p c to
299 * This does not depend on the used encoding, since the inputenc
300 * package only maps the code point \p c to a command, it does not
301 * make this command available.
303 static void validate(char_type c, LaTeXFeatures & features, bool for_mathed = false);
307 EncodingList encodinglist;
309 static MathCommandSet mathcmd;
311 static TextCommandSet textcmd;
313 static MathSymbolSet mathsym;
316 extern Encodings encodings;