4 * This file is part of LyX, the document processor.
5 * Licence details can be found in the file COPYING.
9 * Full author contact details are available in file CREDITS.
19 #include "support/docstream.h"
24 enum mode_type {UNDECIDED_MODE, TEXT_MODE, MATH_MODE, MATHTEXT_MODE, TABLE_MODE};
26 mode_type asMode(mode_type oldmode, std::string const & str);
29 // These are TeX's catcodes
31 catEscape, // 0 backslash
42 catLetter, // 11 a-zA-Z
43 catOther, // 12 none of the above
46 catInvalid // 15 <delete>
57 FLAG_BRACE_LAST = 1 << 1, // last closing brace ends the parsing
58 FLAG_RIGHT = 1 << 2, // next \\right ends the parsing process
59 FLAG_END = 1 << 3, // next \\end ends the parsing process
60 FLAG_BRACK_LAST = 1 << 4, // next closing bracket ends the parsing
61 FLAG_TEXTMODE = 1 << 5, // we are in a box
62 FLAG_ITEM = 1 << 6, // read a (possibly braced token)
63 FLAG_LEAVE = 1 << 7, // leave the loop at the end
64 FLAG_SIMPLE = 1 << 8, // next $ leaves the loop
65 FLAG_EQUATION = 1 << 9, // next \] leaves the loop
66 FLAG_SIMPLE2 = 1 << 10, // next \) leaves the loop
67 FLAG_OPTION = 1 << 11, // read [...] style option
68 FLAG_BRACED = 1 << 12, // read {...} style argument
69 FLAG_CELL = 1 << 13, // read table cell
70 FLAG_TABBING = 1 << 14 // We are inside a tabbing environment
76 // Helper class for parsing
82 Token() : cs_(), cat_(catIgnore) {}
84 Token(docstring const & cs, CatCode cat) : cs_(to_utf8(cs)), cat_(cat) {}
86 /// Returns the token as string
87 std::string const & cs() const { return cs_; }
88 /// Returns the catcode of the token
89 CatCode cat() const { return cat_; }
90 /** Get the character of tokens that were constructed from a single
91 * character input or a two character input and cat_ == catEscape.
92 * FIXME: The intended usage is not clear. The Token class in
93 * ../mathed/MathParser.cpp (which is the anchestor of this
94 * class) uses a separate char member for this method. I
95 * believe that the intended usage is to not cover tokens with
96 * catEscape or catComment, e.g. \code
97 * return (cs_.empty() || cat_ == catEscape || cat_ == catComment) ? 0 : cs_[0];
99 * All usages of this method should be checked. gb 2011-01-05
101 char character() const { return cs_.empty() ? 0 : cs_[0]; }
102 /// Returns the token verbatim
103 std::string asInput() const;
104 /// Is the token an alphanumerical character?
105 bool isAlnumASCII() const;
114 std::ostream & operator<<(std::ostream & os, Token const & t);
117 extern void debugToken(std::ostream & os, Token const & t, unsigned int flags);
122 * Actual parser class
124 * The parser parses every character of the inputstream into a token
125 * and classifies the token.
126 * The following transformations are done:
127 * - Consecutive spaces are combined into one single token with CatCode catSpace
128 * - Consecutive newlines are combined into one single token with CatCode catNewline
129 * - Comments and %\n combinations are parsed into one token with CatCode catComment
134 Parser(Parser const & p);
135 Parser & operator=(Parser const & p);
138 Parser(idocstream & is);
140 Parser(std::string const & s);
145 CatCode catcode(char_type c) const;
147 void setCatcode(char c, CatCode cat);
148 /// set parser to normal or verbatim mode
149 void setCatcodes(cat_type t);
151 /// change the iconv encoding of the input stream
152 /// according to the latex encoding and package
153 void setEncoding(std::string const & encoding, int const & package);
154 /// change the iconv encoding of the input stream
155 void setEncoding(std::string const & encoding);
156 /// get the current iconv encoding of the input stream
157 std::string getEncoding() const { return encoding_iconv_; }
160 int lineno() const { return lineno_; }
163 /// store current position
165 /// restore previous position
167 /// dump contents to screen
170 /// Does an optional argument follow after the current token?
173 typedef std::pair<bool, std::string> Arg;
175 * Get an argument enclosed by \p left and \p right.
176 * If \p allow_escaping is true, a right delimiter escaped by a
177 * backslash does not count as delimiter, but is included in the
179 * \returns wether an argument was found in \p Arg.first and the
180 * argument in \p Arg.second. \see getArg().
182 Arg getFullArg(char left, char right, bool allow_escaping = true);
184 * Get an argument enclosed by \p left and \p right.
185 * If \p allow_escaping is true, a right delimiter escaped by a
186 * backslash does not count as delimiter, but is included in the
188 * \returns the argument (without \p left and \p right) or the empty
189 * string if the next non-space token is not \p left. Use
190 * getFullArg() if you need to know wether there was an empty
191 * argument or no argument at all.
193 std::string getArg(char left, char right, bool allow_escaping = true);
195 * Like getOpt(), but distinguishes between a missing argument ""
196 * and an empty argument "[]".
198 std::string getFullOpt(bool keepws = false);
200 * \returns getArg('[', ']') including the brackets or the
201 * empty string if there is no such argument.
202 * No whitespace is eaten if \p keepws is true and no optional
203 * argument exists. This is important if an optional argument is
204 * parsed that would go after a command in ERT: In this case the
205 * whitespace is needed to separate the ERT from the subsequent
206 * word. Without it, the ERT and the next word would be concatenated
207 * during .tex export, thus creating an invalid command.
209 std::string getOpt(bool keepws = false);
211 * \returns getFullArg('(', ')') including the parentheses or the
212 * empty string if there is no such argument.
214 std::string getFullParentheseArg();
216 * \returns the contents of the environment \p name.
217 * <tt>\begin{name}</tt> must be parsed already, <tt>\end{name}</tt>
218 * is parsed but not returned. This parses nested environments properly.
220 std::string const ertEnvironment(std::string const & name);
222 * The same as ertEnvironment(std::string const & name) but
223 * \begin and \end commands inside the name environment are not parsed.
224 * This function is designed to parse verbatim environments.
226 std::string const plainEnvironment(std::string const & name);
228 * Basically the same as plainEnvironment(std::string const & name) but
229 * instead of \begin and \end commands the parsing is started/stopped
230 * at given characters.
231 * This function is designed to parse verbatim commands.
233 std::string const plainCommand(char left, char right, std::string const & name);
235 * Basically the same as plainEnvironment() but the parsing is
236 * stopped at string \p end_string. Contrary to the other
237 * methods, this uses proper catcode setting. This function is
238 * designed to parse verbatim environments and command. The
239 * intention is to eventually replace all of its siblings.
241 std::string const verbatimStuff(std::string const & end_string);
243 * \returns the contents of the environment \p name.
244 * <tt>\begin{name}</tt> must be parsed already,
245 * <tt>\end{name}</tt> is parsed but not returned. The string
246 * is parsed with proper verbatim catcodes and one newline is
247 * removed from head and tail of the string if applicable.
249 std::string const verbatimEnvironment(std::string const & end_string);
251 * Returns the character of the current token and increments
252 * the token position.
256 void error(std::string const & msg);
257 /// Parses one token from \p is
260 void push_back(Token const & t);
261 /// The previous token.
262 Token const prev_token() const;
263 /// The current token.
264 Token const curr_token() const;
266 Token const next_token();
267 /// The next but one token.
268 Token const next_next_token();
269 /// Make the next token current and return that.
270 Token const get_token();
271 /// \return whether the current token starts a new paragraph
273 /// skips spaces (and comments if \p skip_comments is true)
274 /// \return whether whitespace was skipped (not comments)
275 bool skip_spaces(bool skip_comments = false);
276 /// puts back spaces (and comments if \p skip_comments is true)
277 void unskip_spaces(bool skip_comments = false);
279 void lex(std::string const & s);
283 std::string verbatim_item();
285 std::string verbatimOption();
286 /// resets the parser to initial state
290 /// Setup catcode table
295 std::vector<Token> tokens_;
299 std::vector<unsigned> positions_;
301 idocstringstream * iss_;
304 /// iconv name of the current encoding
305 std::string encoding_iconv_;
307 CatCode theCatcode_[256];
309 cat_type theCatcodesType_;