#define ENCODING_H
#include "support/docstring.h"
+#include "support/trivstring.h"
#include "support/types.h"
#include <map>
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;
};
-///
+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 & textcommand, docstring const & mathcommand,
+ std::string const & textpreamble, std::string const & mathpreamble,
+ std::string const & tipashortcut, unsigned int flags);
+ // we assume that at least one command is nonempty when using unicodesymbols
+ bool isUnicodeSymbol() const { return !textcommand_.empty() || !mathcommand_.empty(); }
+ /// LaTeX command (text mode) for this character
+ docstring const textcommand() const { return textcommand_; }
+ /// LaTeX command (math mode) for this character
+ docstring mathcommand() const { return mathcommand_; }
+ /// Needed LaTeX preamble (or feature) for text mode
+ std::string textpreamble() const { return textpreamble_; }
+ /// Needed LaTeX preamble (or feature) for math mode
+ std::string mathpreamble() const { return mathpreamble_; }
+ /// Is this a combining character?
+ bool combining() const { return flags_ & CharInfoCombining ? true : false; }
+ /// Is \c textpreamble a feature known by LaTeXFeatures, or a raw LaTeX
+ /// command?
+ bool textfeature() const { return flags_ & CharInfoTextFeature ? true : false; }
+ /// Is \c mathpreamble a feature known by LaTeXFeatures, or a raw LaTeX
+ /// command?
+ bool mathfeature() const { return flags_ & CharInfoMathFeature ? true : false; }
+ /// Always force the LaTeX command, even if the encoding contains
+ /// this character?
+ bool force() const { return flags_ & CharInfoForce ? true : false; }
+ /// Force the LaTeX command for some encodings?
+ bool forceselected() const { return flags_ & CharInfoForceSelected ? true : false; }
+ /// Disable LaTeX command => char_type conversion for this deprecated symbol?
+ bool deprecated() const { return flags_ & CharInfoDeprecated ? true : false; }
+ /// TIPA shortcut
+ std::string const tipashortcut() const { return tipashortcut_; }
+ /// \c textcommand needs no termination (such as {} or space).
+ bool textnotermination() const { return flags_ & CharInfoTextNoTermination ? true : false; }
+ /// \c mathcommand needs no termination (such as {} or space).
+ bool mathnotermination() const { return flags_ & CharInfoMathNoTermination ? true : false; }
+ ///
+private:
+ /// LaTeX command (text mode) for this character
+ trivdocstring textcommand_;
+ /// LaTeX command (math mode) for this character
+ trivdocstring mathcommand_;
+ /// Needed LaTeX preamble (or feature) for text mode
+ trivstring textpreamble_;
+ /// Needed LaTeX preamble (or feature) for math mode
+ trivstring mathpreamble_;
+ /// TIPA shortcut
+ trivstring tipashortcut_;
+ /// 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?
enum Package {
- none,
- inputenc,
- CJK,
- japanese
+ none = 1,
+ inputenc = 2,
+ CJK = 4,
+ japanese = 8
};
+ /// Represent any of the above packages
+ static int const any;
///
- Encoding() {}
+ Encoding() : fixedwidth_(true), unsafe_(false), forced_(0),
+ start_encodable_(0), package_(none), complete_(false) {}
///
Encoding(std::string const & n, std::string const & l,
std::string const & g, std::string const & i,
- bool f, Package p);
+ bool f, bool u, 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 guiName() const { return guiName_; }
///
- std::string const & iconvName() const { return iconvName_; }
+ std::string const iconvName() const { return iconvName_; }
///
- bool const & hasFixedWidth() const { return fixedwidth_; }
+ bool hasFixedWidth() const { return fixedwidth_; }
+ ///
+ bool unsafe() const { return unsafe_; }
/// \p c is representable in this encoding without a LaTeX macro
bool encodable(char_type c) const;
/**
* the command needs to be terminated by {} or a space.
*/
std::pair<docstring, bool> latexChar(char_type c) const;
+ /**
+ * Convert \p input to something that LaTeX can understand.
+ * This is either the string itself (if it is representable
+ * in this encoding), or a LaTeX macro.
+ * If a character is not representable in this encoding, but no
+ * LaTeX macro is known, a warning is given of lyxerr, and the
+ * character is returned in the second string of the pair and
+ * omitted in the first.
+ * \p dryrun specifies whether the string is used within source
+ * preview (which yields a special warning).
+ */
+ std::pair<docstring, docstring> latexString(docstring const & input,
+ bool dryrun = false) const;
/// Which LaTeX package handles this encoding?
Package package() const { return package_; }
/// A list of all characters usable in this encoding
std::vector<char_type> symbolsList() const;
private:
+ /**
+ * 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.
+ */
+ 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 {, }
+ /// and \\ may appear in high bytes?
+ bool unsafe_;
///
typedef std::set<char_type> 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_;
/**
* 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 {
///
typedef std::set<char_type> MathSymbolSet;
///
- typedef std::map<std::string, Encoding> EncodingList;
+ typedef std::map<trivstring, Encoding> 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 {
void read(support::FileName const & encfile,
support::FileName const & symbolsfile);
/// Get encoding from LyX name \p name
- Encoding const * fromLyXName(std::string const & name) const;
- /// Get encoding from LaTeX name \p name
- Encoding const * fromLaTeXName(std::string const & name) const;
+ Encoding const *
+ 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;
+ /// 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;
///
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
* \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);
+ /**
+ * 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.
*/
* 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.
static docstring fromLaTeXCommand(docstring const & cmd, int cmdtype,
bool & needsTermination, docstring & rem,
std::set<std::string> * 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);
-private:
+protected:
///
EncodingList encodinglist;
///