X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FEncoding.h;h=4913ddc238346193941fc5d2e5addc8e7ad12ac9;hb=5cb80b867f4a59c3253487652ba74a29ad5b3f0f;hp=fb36e726511736be548ee43b7d185785748bb7c3;hpb=2eea1590b1016d873608e6b86d2a8dd08f01ed65;p=lyx.git diff --git a/src/Encoding.h b/src/Encoding.h index fb36e72651..4913ddc238 100644 --- a/src/Encoding.h +++ b/src/Encoding.h @@ -14,6 +14,7 @@ #define ENCODING_H #include "support/docstring.h" +#include "support/trivstring.h" #include "support/types.h" #include @@ -24,22 +25,114 @@ 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(); - + virtual ~EncodingException() noexcept {} + virtual const char * what() const noexcept override; + char_type failed_char; int par_id; pos_type pos; }; -/// +enum CharInfoFlags { + /// + CharInfoCombining = 1, + /// + CharInfoTextFeature = 2, + /// + CharInfoMathFeature = 4, + /// + CharInfoForce = 8, + /// + CharInfoTextNoTermination = 16, + /// + CharInfoMathNoTermination = 32, + /// + CharInfoForceSelected = 64, + /// + CharInfoDeprecated = 128 +}; + + +/// Information about a single UCS4 character +class CharInfo { +public: + CharInfo() : flags_(0) {} + CharInfo( + docstring const & text_command, docstring const & math_command, + std::string const & text_preamble, std::string const & math_preamble, + std::string const & tipa_shortcut, unsigned int flags); + CharInfo( + std::vector const & text_commands, std::vector const & math_commands, + std::string const & text_preamble, std::string const & math_preamble, + std::string const & tipa_shortcut, unsigned int flags); + // Add a new text command for this symbol. + void addTextCommand(const docstring& newTextCommand) { text_commands_.emplace_back(newTextCommand); } + // Add a new math command for this symbol. + void addMathCommand(const docstring& newMathCommand) { math_commands_.emplace_back(newMathCommand); } + + // we assume that at least one command is nonempty when using unicodesymbols + bool isUnicodeSymbol() const { return !text_commands_.empty() || !math_commands_.empty(); } + /// LaTeX command (text mode) for this character + docstring textCommand() const { return text_commands_[0]; } + /// All known LaTeX commands (text mode) for this character + const std::vector& textCommands() const { return text_commands_; } + /// LaTeX command (math mode) for this character + docstring mathCommand() const { return math_commands_[0]; } + /// All known LaTeX commands (math mode) for this character + const std::vector& mathCommands() const { return math_commands_; } + /// Needed LaTeX preamble (or feature) for text mode + std::string textPreamble() const { return text_preamble_; } + /// Needed LaTeX preamble (or feature) for math mode + std::string mathPreamble() const { return math_preamble_; } + /// Is this a combining character? + bool combining() const { return flags_ & CharInfoCombining; } + /// Is \c textPreamble a feature known by LaTeXFeatures, or a raw LaTeX + /// command? + bool textFeature() const { return flags_ & CharInfoTextFeature; } + /// Is \c mathPreamble a feature known by LaTeXFeatures, or a raw LaTeX + /// command? + bool mathFeature() const { return flags_ & CharInfoMathFeature; } + /// Always force the LaTeX command, even if the encoding contains + /// this character? + bool force() const { return flags_ & CharInfoForce; } + /// Force the LaTeX command for some encodings? + bool forceSelected() const { return flags_ & CharInfoForceSelected; } + /// Disable LaTeX command => char_type conversion for this deprecated symbol? + bool deprecated() const { return flags_ & CharInfoDeprecated; } + /// TIPA shortcut + std::string const tipaShortcut() const { return tipa_shortcut_; } + /// \c textCommand needs no termination (such as {} or space). + bool textNoTermination() const { return flags_ & CharInfoTextNoTermination; } + /// \c mathCommand needs no termination (such as {} or space). + bool mathNoTermination() const { return flags_ & CharInfoMathNoTermination; } + /// +private: + /// LaTeX commands (text mode) for this character. The first one is the default, the others + /// are only present for compatibility other ways users may encode the character + std::vector text_commands_; + /// LaTeX command (math mode) for this character. The first one is the default, the others + // /// are only present for compatibility other ways users may encode the character + std::vector math_commands_; + /// Needed LaTeX preamble (or feature) for text mode + trivstring text_preamble_; + /// Needed LaTeX preamble (or feature) for math mode + trivstring math_preamble_; + /// TIPA shortcut + trivstring tipa_shortcut_; + /// feature flags + unsigned int flags_; +}; + + +/** + * An encoding as defined in lib/encodings. + * All const methods are thread-safe, so the caller does not need any locking. + * This property must be kept when changing the class. + */ class Encoding { public: /// Which LaTeX package handles this encoding? @@ -50,9 +143,10 @@ public: japanese = 8 }; /// Represent any of the above packages - static int const any = -1; + static int const any; /// - Encoding() {} + Encoding() : fixedwidth_(true), unsafe_(false), forced_(nullptr), + start_encodable_(0), package_(none), complete_(false) {} /// Encoding(std::string const & n, std::string const & l, std::string const & g, std::string const & i, @@ -60,13 +154,13 @@ public: /// 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 guiName() const { return guiName_; } /// - std::string const & iconvName() const { return iconvName_; } + std::string const iconvName() const { return iconvName_; } /// bool hasFixedWidth() const { return fixedwidth_; } /// @@ -95,7 +189,7 @@ public: * \p dryrun specifies whether the string is used within source * preview (which yields a special warning). */ - std::pair latexString(docstring const input, + std::pair latexString(docstring const & input, bool dryrun = false) const; /// Which LaTeX package handles this encoding? Package package() const { return package_; } @@ -109,13 +203,13 @@ private: */ bool isForced(char_type c) const; /// - std::string name_; + trivstring name_; /// - std::string latexName_; + trivstring latexName_; /// - std::string guiName_; + trivstring guiName_; /// - std::string iconvName_; + trivstring iconvName_; /// Is this a fixed width encoding? bool fixedwidth_; /// Is this encoding TeX unsafe, e.g. control characters like {, } @@ -125,13 +219,13 @@ private: typedef std::set CharSet; /// Set of UCS4 characters that we can encode (for singlebyte /// encodings only) - mutable CharSet encodable_; + CharSet encodable_; /// Set of UCS4 characters that we can't encode CharSet const * forced_; /// 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. - mutable char_type start_encodable_; + char_type start_encodable_; /// Which LaTeX package handles this encoding? Package package_; /** @@ -141,7 +235,7 @@ private: * This is needed especially for the multibyte encodings, if we * complete all encoding info on startup it takes 2-3 minutes. */ - mutable bool complete_; + bool complete_; }; class Encodings { @@ -153,7 +247,7 @@ public: /// typedef std::set MathSymbolSet; /// - typedef std::map EncodingList; + typedef std::map EncodingList; /// iterator to iterate over all encodings. /// We hide the fact that our encoding list is implemented as a map. class const_iterator : public EncodingList::const_iterator { @@ -176,56 +270,42 @@ public: fromLyXName(std::string const & name, bool allowUnsafe = false) const; /// Get encoding from LaTeX name \p name and package \p package Encoding const * fromLaTeXName(std::string const & name, - int const & package = Encoding::any, bool allowUnsafe = false) const; + int package = Encoding::any, bool allowUnsafe = false) const; /// Get encoding from iconv name \p name and package \p package Encoding const * fromIconvName(std::string const & name, - int const & package = Encoding::any, bool allowUnsafe = false) const; + int package = Encoding::any, bool allowUnsafe = false) const; /// const_iterator begin() const { return encodinglist.begin(); } /// const_iterator end() const { return encodinglist.end(); } - /// - enum LetterForm { - /// - FORM_ISOLATED, - /// - FORM_FINAL, - /// - FORM_INITIAL, - /// - FORM_MEDIAL - }; - /// - static bool isHebrewComposeChar(char_type c); - /// - static bool isArabicComposeChar(char_type c); - /// - static bool isArabicSpecialChar(char_type c); - /// - static bool isArabicChar(char_type c); - /// - static char_type transformChar(char_type c, LetterForm form); + /// Accessor for the unicode information table. + static CharInfo const & unicodeCharInfo(char_type c); /// Is this a combining char? static bool isCombiningChar(char_type c); /// Return the TIPA shortcut static std::string const TIPAShortcut(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. + * Test, if \p c is a supported Greek or Cyrillic letter. + * Return script macro name or empty string. */ - static bool isKnownScriptChar(char_type const c, std::string & preamble); + static std::string const isKnownScriptChar(char_type const c); + /// Does \p fontenc support characters in \p script? + static bool fontencSupportsScript(std::string const & fontenc, + std::string const & script); /** * 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); + /** + * Do we have to wrap in \text this character when in mathmode? + * This is true if \p c is not ascii and the "mathalpha" flag is not + * set and a mathCommand is not defined in the unicodesymbols file. + */ + static bool isUnicodeTextOnly(char_type c); /** * Register \p c as a mathmode command. */ @@ -250,10 +330,6 @@ public: * 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 for_master = true); /** * If \p c cannot be encoded in the given \p encoding, convert * it to something that LaTeX can understand in mathmode. @@ -272,7 +348,7 @@ public: */ static char_type fromLaTeXCommand(docstring const & cmd, int cmdtype, bool & combining, bool & needsTermination, - std::set * req = 0); + std::set * req = nullptr); /// enum LatexCmd { /// @@ -292,17 +368,9 @@ public: */ static docstring fromLaTeXCommand(docstring const & cmd, int cmdtype, bool & needsTermination, docstring & rem, - std::set * req = 0); - /** - * Add the preamble snippet needed for the output of \p c to - * \p features. - * This does not depend on the used encoding, since the inputenc - * 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, bool for_mathed = false); + std::set * req = nullptr); -private: +protected: /// EncodingList encodinglist; ///