]> git.lyx.org Git - lyx.git/blobdiff - src/mathed/math_parser.C
Finxing Umlauts, Part I
[lyx.git] / src / mathed / math_parser.C
index fe3244226cfd53b7c39498439965a5cab11cd0cb..08d06ca3afbff5017fa543af73e3c7d98355597c 100644 (file)
@@ -45,9 +45,6 @@ point to write some macros:
 
 #include <config.h>
 
-#include <cctype>
-#include <stack>
-
 #ifdef __GNUG__
 #pragma implementation
 #endif
@@ -67,28 +64,36 @@ point to write some macros:
 #include "math_macrotemplate.h"
 #include "math_hullinset.h"
 #include "math_rootinset.h"
+#include "math_sizeinset.h"
 #include "math_sqrtinset.h"
 #include "math_scriptinset.h"
 #include "math_specialcharinset.h"
 #include "math_splitinset.h"
 #include "math_sqrtinset.h"
 #include "math_support.h"
+
 #include "lyxlex.h"
 #include "debug.h"
+
 #include "support/lstrings.h"
 
+#include <cctype>
+#include <stack>
+#include <algorithm>
+
 using std::istream;
 using std::ostream;
 using std::ios;
 using std::endl;
 using std::stack;
+using std::fill;
 
 
 namespace {
 
 bool stared(string const & s)
 {
-       unsigned n = s.size();
+       string::size_type const n = s.size();
        return n && s[n - 1] == '*';
 }
 
@@ -133,20 +138,19 @@ enum {
        FLAG_RIGHT      = 1 << 2,  //  next \\right ends the parsing process
        FLAG_END        = 1 << 3,  //  next \\end ends the parsing process
        FLAG_BRACK_END  = 1 << 4,  //  next closing bracket ends the parsing process
-       FLAG_ITEM       = 1 << 7,  //  read a (possibly braced token)
-       FLAG_BLOCK      = 1 << 8,  //  next block ends the parsing process
+       FLAG_BOX        = 1 << 5,  //  we are in a box
+       FLAG_ITEM       = 1 << 6,  //  read a (possibly braced token)
+       FLAG_BLOCK      = 1 << 7,  //  next block ends the parsing process
+       FLAG_BLOCK2     = 1 << 8,  //  next block2 ends the parsing process
        FLAG_LEAVE      = 1 << 9   //  leave the loop at the end
 };
 
 
 void catInit()
 {
-       for (int i = 0; i <= 255; ++i) 
-               theCatcode[i] = catOther;
-       for (int i = 'a'; i <= 'z'; ++i) 
-               theCatcode[i] = catLetter;
-       for (int i = 'A'; i <= 'Z'; ++i) 
-               theCatcode[i] = catLetter;
+       fill(theCatcode, theCatcode + 256, catOther);
+       fill(theCatcode + 'a', theCatcode + 'z' + 1, catLetter);
+       fill(theCatcode + 'A', theCatcode + 'Z' + 1, catLetter);
 
        theCatcode['\\'] = catEscape;   
        theCatcode['{']  = catBegin;    
@@ -178,7 +182,7 @@ public:
        ///
        Token(char c, CatCode cat) : cs_(), char_(c), cat_(cat) {}
        ///
-       Token(const string & cs) : cs_(cs), char_(0), cat_(catIgnore) {}
+       Token(string const & cs) : cs_(cs), char_(0), cat_(catIgnore) {}
 
        ///
        string const & cs() const { return cs_; }
@@ -210,16 +214,17 @@ string Token::asString() const
        return cs_.size() ? cs_ : string(1, char_);
 }
 
-bool operator==(Token const & s, Token const & t)
-{
-       return s.character() == t.character()
-               && s.cat() == t.cat() && s.cs() == t.cs(); 
-}
-
-bool operator!=(Token const & s, Token const & t)
-{
-       return !(s == t);
-}
+// Angus' compiler says these are not needed
+//bool operator==(Token const & s, Token const & t)
+//{
+//     return s.character() == t.character()
+//             && s.cat() == t.cat() && s.cs() == t.cs(); 
+//}
+//
+//bool operator!=(Token const & s, Token const & t)
+//{
+//     return !(s == t);
+//}
 
 ostream & operator<<(ostream & os, Token const & t)
 {
@@ -240,7 +245,7 @@ public:
        Parser(istream & is);
 
        ///
-       string parse_macro();
+       bool parse_macro(string & name);
        ///
        bool parse_normal(MathAtom &);
        ///
@@ -251,6 +256,8 @@ public:
        void putback();
 
 private:
+       ///
+       void parse_into1(MathArray & array, unsigned flags, MathTextCodes);
        ///
        string getArg(char lf, char rf);
        ///
@@ -259,6 +266,8 @@ private:
        void error(string const & msg);
        ///
        bool parse_lines(MathAtom & t, bool numbered, bool outmost);
+       /// parses {... & ... \\ ... & ... }
+       bool parse_lines2(MathAtom & t);
 
 private:
        ///
@@ -277,6 +286,10 @@ private:
        Token const & getToken();
        /// skips spaces if any
        void skipSpaces();
+       /// skips opening brace
+       void skipBegin();
+       /// skips closing brace
+       void skipEnd();
        /// counts a sequence of hlines
        int readHLines();
        ///
@@ -355,6 +368,24 @@ void Parser::skipSpaces()
 }
 
 
+void Parser::skipBegin()
+{
+       if (nextToken().cat() == catBegin)
+               getToken();
+       else
+               lyxerr << "'{' expected\n";
+}
+
+
+void Parser::skipEnd()
+{
+       if (nextToken().cat() == catEnd)
+               getToken();
+       else
+               lyxerr << "'}' expected\n";
+}
+
+
 int Parser::readHLines()
 {
        int num = 0;
@@ -492,6 +523,7 @@ void Parser::error(string const & msg)
 }
 
 
+
 bool Parser::parse_lines(MathAtom & t, bool numbered, bool outmost)
 {      
        MathGridInset * p = t->asGridInset();
@@ -500,8 +532,6 @@ bool Parser::parse_lines(MathAtom & t, bool numbered, bool outmost)
                return false;
        }
 
-       const int cols = p->ncols();
-
        // save global variables
        bool   const saved_num   = curr_num_;
        string const saved_label = curr_label_;
@@ -515,9 +545,14 @@ bool Parser::parse_lines(MathAtom & t, bool numbered, bool outmost)
                curr_label_.erase();
 
                // reading a row
-               for (int col = 0; col < cols; ++col) {
+               for (MathInset::col_type col = 0; col < p->ncols(); ++col) {
                        //lyxerr << "reading cell " << row << " " << col << "\n";
-                       parse_into(p->cell(col + row * cols), FLAG_BLOCK);
+               
+                       MathArray & ar = p->cell(col + row * p->ncols());
+                       parse_into(ar, FLAG_BLOCK);
+                       // remove 'unnecessary' braces:
+                       if (ar.size() == 1 && ar.back()->asBraceInset())
+                               ar = ar.back()->asBraceInset()->cell(0);
 
                        // break if cell is not followed by an ampersand
                        if (nextToken().cat() != catAlign) {
@@ -578,26 +613,82 @@ bool Parser::parse_lines(MathAtom & t, bool numbered, bool outmost)
 }
 
 
-string Parser::parse_macro()
+bool Parser::parse_lines2(MathAtom & t)
+{      
+       MathGridInset * p = t->asGridInset();
+       if (!p) {
+               lyxerr << "error in Parser::parse_lines() 1\n";
+               return false;
+       }
+
+       skipBegin();
+
+       for (int row = 0; true; ++row) {
+               // reading a row
+               for (MathInset::col_type col = 0; true; ++col) {
+                       //lyxerr << "reading cell " << row << " " << col << " " << p->ncols() << "\n";
+               
+                       if (col >= p->ncols()) {
+                               //lyxerr << "adding col " << col << "\n";
+                               p->addCol(p->ncols());
+                       }
+
+                       parse_into(p->cell(col + row * p->ncols()), FLAG_BLOCK2);
+                       //lyxerr << "read cell: " << p->cell(col + row * p->ncols()) << "\n";
+
+                       // break if cell is not followed by an ampersand
+                       if (nextToken().cat() != catAlign) {
+                               //lyxerr << "less cells read than normal in row/col: " << row << " " << col << "\n";
+                               break;
+                       }
+                       
+                       // skip the ampersand
+                       getToken();
+               }
+
+               // is a \\ coming?
+               if (nextToken().isCR()) {
+                       // skip the cr-token
+                       getToken();
+               }
+
+               // we are finished if the next token is an '}'
+               if (nextToken().cat() == catEnd) {
+                       // skip the end-token
+                       getToken();
+                       // leave the 'read a line'-loop
+                       break;
+               }
+
+               // otherwise, we have to start a new row
+               p->appendRow();
+       }
+
+       return true;
+}
+
+
+
+bool Parser::parse_macro(string & name)
 {
-       string name = "{error}";
+       name = "{error}";
        skipSpaces();
 
        if (getToken().cs() != "newcommand") {
                lyxerr << "\\newcommand expected\n";
-               return name;
+               return false;
        }
 
        if (getToken().cat() != catBegin) {
                lyxerr << "'{' in \\newcommand expected (1)\n";
-               return name;
+               return false;
        }
 
        name = getToken().cs();
 
        if (getToken().cat() != catEnd) {
                lyxerr << "'}' expected\n";
-               return name;
+               return false;
        }
 
        string    arg  = getArg('[', ']');
@@ -605,13 +696,22 @@ string Parser::parse_macro()
 
        if (getToken().cat() != catBegin) {
                lyxerr << "'{' in \\newcommand expected (2)\n";
-               return name;
+               return false;
        }
 
        MathArray ar;
        parse_into(ar, FLAG_BRACE_LAST);
+
+       // we cannot handle recursive stuff at all
+       MathArray test;
+       test.push_back(createMathInset(name));
+       if (ar.contains(test)) {
+               lyxerr << "we cannot handle recursive macros at all.\n";
+               return false;
+       }
+
        MathMacroTable::create(name, narg, ar);
-       return name;
+       return true;
 }
 
 
@@ -669,6 +769,12 @@ bool Parser::parse_normal(MathAtom & matrix)
 
        string const name = getArg('{', '}');
 
+       if (name == "math") {
+               matrix = MathAtom(new MathHullInset(LM_OT_SIMPLE));
+               parse_into(matrix->cell(0), 0);
+               return true;
+       }
+
        if (name == "equation" || name == "equation*" || name == "displaymath") {
                curr_num_ = (name == "equation");
                curr_label_.erase();
@@ -725,6 +831,15 @@ bool Parser::parse_normal(MathAtom & matrix)
 
 
 void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
+{
+       parse_into1(array, flags, code);
+       // remove 'unnecessary' braces:
+       if (array.size() == 1 && array.back()->asBraceInset())
+               array = array.back()->asBraceInset()->cell(0);
+}
+
+
+void Parser::parse_into1(MathArray & array, unsigned flags, MathTextCodes code)
 {
        bool panic  = false;
        int  limits = 0;
@@ -732,7 +847,7 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
        while (good()) {
                Token const & t = getToken();
        
-               //lyxerr << "t: " << t << " flags: " << flags << "'\n";
+               //lyxerr << "t: " << t << " flags: " << flags << "\n";
                //array.dump(lyxerr);
                //lyxerr << "\n";
 
@@ -756,11 +871,27 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
                        }
                }
 
+               if (flags & FLAG_BLOCK2) {
+                       if (t.cat() == catAlign || t.isCR() || t.cs() == "end"
+                                       || t.cat() == catEnd) {
+                               putback();
+                               return;
+                       }
+               }
+
                //
                // cat codes
                //
-               if (t.cat() == catMath)
-                       break;
+               if (t.cat() == catMath) {
+                       if (flags & FLAG_BOX) {
+                               // we are inside an mbox, so opening new math is allowed
+                               array.push_back(MathAtom(new MathHullInset(LM_OT_SIMPLE)));
+                               parse_into(array.back()->cell(0), 0);
+                       } else {
+                               // otherwise this is the end of the formula
+                               break;
+                       }
+               }
 
                else if (t.cat() == catLetter)
                        add(array, t.character(), code);
@@ -770,7 +901,7 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
 
                else if (t.cat() == catParameter) {
                        Token const & n = getToken();
-                       array.push_back(MathAtom(new MathMacroArgument(n.character() - '0')));
+                       array.push_back(MathAtom(new MathMacroArgument(n.character()-'0', code)));
                }
 
                else if (t.cat() == catBegin) {
@@ -780,7 +911,11 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
 #warning this might be wrong in general!
 #endif
                        // ignore braces around simple items
-                       if (ar.size() == 1 || (ar.size() == 2 && ar.back()->asScriptInset())) {
+                       if ((ar.size() == 1 && !ar.front()->needsBraces()
+       || (ar.size() == 2 && !ar.front()->needsBraces()
+                                           && ar.back()->asScriptInset()))
+       || (ar.size() == 0 && array.size() == 0))
+                       {
                                array.push_back(ar);
                        } else {
                                array.push_back(MathAtom(new MathBraceInset));
@@ -889,19 +1024,6 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
                        return;
                }
 
-/*             
-               case LM_TK_STY:
-               {
-                       lyxerr[Debug::MATHED] << "LM_TK_STY not implemented\n";
-                       //MathArray tmp = array;
-                       //MathSizeInset * p = new MathSizeInset(MathStyles(lval_->id));
-                       //array.push_back(p);
-                       //parse_into(p->cell(0), FLAG_BRACE_FONT);
-                       break; 
-               }
-
-*/
-               
                else if (t.cs() == "begin") {
                        string const name = getArg('{', '}');   
                        if (name == "array") {
@@ -949,8 +1071,18 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
                        return;
                }
 
-/*
+               else if (t.cs() == "xymatrix") {
+                       array.push_back(createMathInset(t.cs()));
+                       parse_lines2(array.back());
+               }
+
                // Disabled
+#if 0
+               else if (0 && t.cs() == "ar") {
+                       array.push_back(createMathInset(t.cs()));
+                       parse_lines2(array.back());
+               }
+
                else if (t.cs() == "mbox") {
                        array.push_back(createMathInset(t.cs()));
                        // slurp in the argument of mbox
@@ -958,7 +1090,8 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
                        MathBoxInset * p = array.back()->asBoxInset();
                        //lyx::assert(p);
                }
-*/
+#endif
+
        
                else if (t.cs().size()) {
                        latexkeys const * l = in_word_set(t.cs());
@@ -984,6 +1117,19 @@ void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
                                        code = static_cast<MathTextCodes>(l->id);
                                }
 
+                               else if (l->token == LM_TK_BOX) {
+                                       MathAtom p = createMathInset(t.cs());
+                                       parse_into(p->cell(0), FLAG_ITEM | FLAG_BOX, LM_TC_BOX);
+                                       array.push_back(p);
+                               }
+
+                               else if (l->token == LM_TK_STY) {
+                                       MathAtom p = createMathInset(t.cs());
+                                       parse_into(p->cell(0), flags, code);
+                                       array.push_back(p);
+                                       return;
+                               }
+
                                else {
                                        MathAtom p = createMathInset(t.cs());
                                        for (MathInset::idx_type i = 0; i < p->nargs(); ++i) 
@@ -1036,23 +1182,23 @@ void mathed_parse_cell(MathArray & ar, istream & is)
 
 
 
-string mathed_parse_macro(string const & str)
+bool mathed_parse_macro(string & name, string const & str)
 {
        istringstream is(str.c_str());
        Parser parser(is);
-       return parser.parse_macro();
+       return parser.parse_macro(name);
 }
 
-string mathed_parse_macro(istream & is)
+bool mathed_parse_macro(string & name, istream & is)
 {
        Parser parser(is);
-       return parser.parse_macro();
+       return parser.parse_macro(name);
 }
 
-string mathed_parse_macro(LyXLex & lex)
+bool mathed_parse_macro(string & name, LyXLex & lex)
 {
        Parser parser(lex);
-       return parser.parse_macro();
+       return parser.parse_macro(name);
 }