X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FEncoding.h;h=1f367366a82dde25abb282b00ec83db45873b02f;hb=35058f35a9864f8adf9ae225d0747cd19d84122c;hp=6d10d3bf67c41d46f910c5c2459ed7ea94f9de00;hpb=f630be890494c849981e4fb52ea4740506e92bed;p=lyx.git diff --git a/src/Encoding.h b/src/Encoding.h index 6d10d3bf67..1f367366a8 100644 --- a/src/Encoding.h +++ b/src/Encoding.h @@ -4,7 +4,7 @@ * This file is part of LyX, the document processor. * Licence details can be found in the file COPYING. * - * \author Lars Gullik Bjønnes + * \author Lars Gullik Bjønnes * \author Jean-Marc Lasgouttes * * Full author contact details are available in file CREDITS. @@ -14,31 +14,59 @@ #define ENCODING_H #include "support/docstring.h" +#include "support/types.h" #include #include +#include namespace lyx { namespace support { class FileName; } +class Buffer; class LaTeXFeatures; +class EncodingException : public std::exception { +public: + EncodingException(char_type c); + virtual ~EncodingException() throw() {} + virtual const char * what() const throw(); + + char_type failed_char; + int par_id; + pos_type pos; +}; + /// class Encoding { public: + /// Which LaTeX package handles this encoding? + enum Package { + none, + inputenc, + CJK, + japanese + }; /// Encoding() {} /// Encoding(std::string const & n, std::string const & l, - std::string const & i); + std::string const & g, std::string const & i, + bool f, Package p); + /// + void init() const; + /// + std::string const & name() const { return name_; } /// - std::string const & name() const { return Name_; } + std::string const & latexName() const { return latexName_; } /// - std::string const & latexName() const { return LatexName_; } + std::string const & guiName() const { return guiName_; } /// std::string const & iconvName() const { return iconvName_; } + /// + bool const & hasFixedWidth() const { return fixedwidth_; } /** * Convert \p c to something that LaTeX can understand. * This is either the character itself (if it is representable @@ -47,27 +75,51 @@ public: * LaTeX macro is known, a warning is given of lyxerr, and the * character is returned. */ - docstring const latexChar(char_type c) const; + docstring latexChar(char_type c, bool no_commands = false) const; + /// Which LaTeX package handles this encoding? + Package package() const { return package_; } + /// A list of all characters usable in this encoding + std::vector symbolsList() const; private: /// - std::string Name_; + std::string name_; + /// + std::string latexName_; /// - std::string LatexName_; + std::string guiName_; /// std::string iconvName_; + /// Is this a fixed width encoding? + bool fixedwidth_; /// typedef std::set CharSet; /// Set of UCS4 characters that we can encode (for singlebyte /// encodings only) - CharSet encodable_; + mutable CharSet encodable_; /// All code points below this are encodable. This helps us to avoid /// lokup of ASCII characters in encodable_ and gives about 1 sec /// speedup on export of the Userguide. - char_type start_encodable_; + mutable char_type start_encodable_; + /// Which LaTeX package handles this encoding? + Package package_; + /** + * If this is true the stored information about the encoding covers + * all encodable characters. We set this to false initially so that + * we only need to query iconv for the actually used encodings. + * This is needed especially for the multibyte encodings, if we + * complete all encoding info on startup it takes 2-3 minutes. + */ + mutable bool complete_; }; class Encodings { public: + /// + typedef std::set MathCommandSet; + /// + typedef std::set TextCommandSet; + /// + typedef std::set MathSymbolSet; /// typedef std::map EncodingList; /// iterator to iterate over all encodings. @@ -86,11 +138,11 @@ public: /// \param encfile encodings definition file /// \param symbolsfile unicode->LaTeX mapping file void read(support::FileName const & encfile, - support::FileName const & symbolsfile); + support::FileName const & symbolsfile); /// Get encoding from LyX name \p name - Encoding const * getFromLyXName(std::string const & name) const; + Encoding const * fromLyXName(std::string const & name) const; /// Get encoding from LaTeX name \p name - Encoding const * getFromLaTeXName(std::string const & name) const; + Encoding const * fromLaTeXName(std::string const & name) const; /// const_iterator begin() const { return encodinglist.begin(); } @@ -98,7 +150,7 @@ public: const_iterator end() const { return encodinglist.end(); } /// - enum Letter_Form { + enum LetterForm { /// FORM_ISOLATED, /// @@ -109,17 +161,95 @@ public: FORM_MEDIAL }; /// - static bool isComposeChar_hebrew(char_type c); + static bool isHebrewComposeChar(char_type c); /// - static bool isComposeChar_arabic(char_type c); + static bool isArabicComposeChar(char_type c); /// - static bool is_arabic_special(char_type c); + static bool isArabicSpecialChar(char_type c); /// - static bool is_arabic(char_type c); + static bool isArabicChar(char_type c); /// - static char_type transformChar(char_type c, Letter_Form form); + static char_type transformChar(char_type c, LetterForm form); /// Is this a combining char? static bool isCombiningChar(char_type c); + /** + * Is this a known char from some language? + * If \p preamble is empty and code point \p c is known to belong + * to a supported script, true is returned and \p preamble is set + * to the corresponding entry in the unicodesymbols file. + * If \p preamble is not empty, a check is made whether code point + * \p c is a known character matching the preamble entry. + */ + static bool isKnownScriptChar(char_type const c, std::string & preamble); + /** + * Do we have to output this character as LaTeX command in any case? + * This is true if the "force" flag is set. + * We need this if the inputencoding does not support a certain glyph. + */ + static bool isForced(char_type c); + /** + * Do we have to display in italics this character when in mathmode? + * This is true if the "mathalpha" flag is set. We use this for + * letters and accented characters that are output as math commands. + */ + static bool isMathAlpha(char_type c); + /** + * Register \p c as a mathmode command. + */ + static void addMathCmd(char_type c) { mathcmd.insert(c); } + /** + * Register \p c as a textmode command. + */ + static void addTextCmd(char_type c) { textcmd.insert(c); } + /** + * Register \p c as a mathmode symbol. + */ + static void addMathSym(char_type c) { mathsym.insert(c); } + /** + * Tell whether \p c is registered as a mathmode command. + */ + static bool isMathCmd(char_type c) { return mathcmd.count(c); } + /** + * Tell whether \p c is registered as a textmode command. + */ + static bool isTextCmd(char_type c) { return textcmd.count(c); } + /** + * Tell whether \p c is registered as a mathmode symbol. + */ + static bool isMathSym(char_type c) { return mathsym.count(c); } + /** + * Initialize mathcmd, textcmd, and mathsym sets. + */ + static void initUnicodeMath(Buffer const & buffer, bool clear_sets = true); + /** + * If \p c cannot be encoded in the given \p encoding, convert + * it to something that LaTeX can understand in mathmode. + * \return whether \p command is a mathmode command + */ + static bool latexMathChar(char_type c, bool mathmode, + Encoding const * encoding, docstring & command); + /** + * Convert the LaTeX command in \p cmd to the corresponding unicode + * point and set \p combining to true if it is a combining symbol + */ + static char_type fromLaTeXCommand(docstring const & cmd, bool & combining); + /// + enum LatexCmd { + /// + MATH_CMD = 1, + /// + TEXT_CMD = 2 + }; + /** + * Convert the LaTeX commands in \p cmd and \return a docstring + * of corresponding unicode points. The conversion stops at the + * first command which could not be converted, and the remaining + * unconverted commands are returned in \p rem. + * The \p cmdtype parameter can be used to limit recognized + * commands to math or text mode commands only. + */ + static docstring fromLaTeXCommand(docstring const & cmd, + docstring & rem, int cmdtype = MATH_CMD | TEXT_CMD); /** * Add the preamble snippet needed for the output of \p c to * \p features. @@ -127,11 +257,17 @@ public: * package only maps the code point \p c to a command, it does not * make this command available. */ - static void validate(char_type c, LaTeXFeatures & features); + static void validate(char_type c, LaTeXFeatures & features, bool for_mathed = false); private: /// EncodingList encodinglist; + /// + static MathCommandSet mathcmd; + /// + static TextCommandSet textcmd; + /// + static MathSymbolSet mathsym; }; extern Encodings encodings;