lyx_mirror/src/mathed/math_parser.C
André Pönitz f660211af3 Fix stupid parser bug
git-svn-id: svn://svn.lyx.org/lyx/lyx-devel/trunk@4496 a592a061-630c-0410-9148-cb99ea01b6c8
2002-06-27 11:31:32 +00:00

1155 lines
25 KiB
C
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

/** The math parser
\author André Pönitz (2001)
*/
/*
If someone desperately needs partial "structures" (such as a few
cells of an array inset or similar) (s)he could uses the
following hack as starting point to write some macros:
\newif\ifcomment
\commentfalse
\ifcomment
\def\makeamptab{\catcode`\&=4\relax}
\def\makeampletter{\catcode`\&=11\relax}
\def\b{\makeampletter\expandafter\makeamptab\bi}
\long\def\bi#1\e{}
\else
\def\b{}\def\e{}
\fi
...
\[\begin{array}{ccc}
1 & 2\b & 3^2\\
4 & 5\e & 6\\
7 & 8 & 9
\end{array}\]
*/
#include <config.h>
#ifdef __GNUG__
#pragma implementation
#endif
#include "math_parser.h"
#include "math_inset.h"
#include "math_arrayinset.h"
#include "math_braceinset.h"
#include "math_boxinset.h"
#include "math_charinset.h"
#include "math_deliminset.h"
#include "math_extern.h"
#include "math_factory.h"
#include "math_kerninset.h"
#include "math_macro.h"
#include "math_macrotable.h"
#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_sqrtinset.h"
#include "math_support.h"
#include "math_xyarrowinset.h"
#include "ref_inset.h"
#include "lyxlex.h"
#include "debug.h"
#include "support/LAssert.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;
using std::vector;
using std::atoi;
//#define FILEDEBUG
namespace {
bool stared(string const & s)
{
string::size_type const n = s.size();
return n && s[n - 1] == '*';
}
// These are TeX's catcodes
enum CatCode {
catEscape, // 0 backslash
catBegin, // 1 {
catEnd, // 2 }
catMath, // 3 $
catAlign, // 4 &
catNewline, // 5 ^^M
catParameter, // 6 #
catSuper, // 7 ^
catSub, // 8 _
catIgnore, // 9
catSpace, // 10 space
catLetter, // 11 a-zA-Z
catOther, // 12 none of the above
catActive, // 13 ~
catComment, // 14 %
catInvalid // 15 <delete>
};
CatCode theCatcode[256];
inline CatCode catcode(unsigned char c)
{
return theCatcode[c];
}
enum {
FLAG_BRACE_LAST = 1 << 1, // last closing brace ends the parsing
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
FLAG_TEXTMODE = 1 << 5, // we are in a box
FLAG_ITEM = 1 << 6, // read a (possibly braced token)
FLAG_LEAVE = 1 << 7, // leave the loop at the end
FLAG_SIMPLE = 1 << 8, // next $ leaves the loop
FLAG_EQUATION = 1 << 9, // next \] leaves the loop
FLAG_SIMPLE2 = 1 << 10 // next \) leaves the loop
};
void catInit()
{
fill(theCatcode, theCatcode + 256, catOther);
fill(theCatcode + 'a', theCatcode + 'z' + 1, catLetter);
fill(theCatcode + 'A', theCatcode + 'Z' + 1, catLetter);
theCatcode['\\'] = catEscape;
theCatcode['{'] = catBegin;
theCatcode['}'] = catEnd;
theCatcode['$'] = catMath;
theCatcode['&'] = catAlign;
theCatcode['\n'] = catNewline;
theCatcode['#'] = catParameter;
theCatcode['^'] = catSuper;
theCatcode['_'] = catSub;
theCatcode[''] = catIgnore;
theCatcode[' '] = catSpace;
theCatcode['\t'] = catSpace;
theCatcode['\r'] = catSpace;
theCatcode['~'] = catActive;
theCatcode['%'] = catComment;
}
//
// Helper class for parsing
//
class Token {
public:
///
Token() : cs_(), char_(0), cat_(catIgnore) {}
///
Token(char c, CatCode cat) : cs_(), char_(c), cat_(cat) {}
///
Token(string const & cs) : cs_(cs), char_(0), cat_(catIgnore) {}
///
string const & cs() const { return cs_; }
///
CatCode cat() const { return cat_; }
///
char character() const { return char_; }
///
string asString() const;
///
bool isCR() const;
private:
///
string cs_;
///
char char_;
///
CatCode cat_;
};
bool Token::isCR() const
{
return cs_ == "\\" || cs_ == "cr" || cs_ == "crcr";
}
string Token::asString() const
{
return cs_.size() ? cs_ : string(1, char_);
}
ostream & operator<<(ostream & os, Token const & t)
{
if (t.cs().size())
os << "\\" << t.cs();
else
os << "[" << t.character() << "," << t.cat() << "]";
return os;
}
class Parser {
public:
///
Parser(LyXLex & lex);
///
Parser(istream & is);
///
bool parse_macro(string & name);
///
bool parse_normal(MathAtom & at);
///
void parse_into(MathArray & array, unsigned flags, bool mathmode);
///
int lineno() const { return lineno_; }
///
void putback();
private:
///
void parse_into1(MathGridInset & grid, unsigned flags, bool mathmode, bool numbered);
///
void parse_into2(MathAtom & at, unsigned flags, bool mathmode, bool numbered);
/// get arg delimited by 'left' and 'right'
string getArg(char left, char right);
///
char getChar();
///
void error(string const & msg);
/// dump contents to screen
void dump() const;
private:
///
void tokenize(istream & is);
///
void tokenize(string const & s);
///
void skipSpaceTokens(istream & is, char c);
///
void push_back(Token const & t);
///
void pop_back();
///
Token const & prevToken() const;
///
Token const & nextToken() const;
///
Token const & getToken();
/// skips spaces if any
void skipSpaces();
///
void lex(string const & s);
///
bool good() const;
///
int lineno_;
///
vector<Token> tokens_;
///
unsigned pos_;
};
Parser::Parser(LyXLex & lexer)
: lineno_(lexer.getLineNo()), pos_(0)
{
tokenize(lexer.getStream());
lexer.eatLine();
}
Parser::Parser(istream & is)
: lineno_(0), pos_(0)
{
tokenize(is);
}
void Parser::push_back(Token const & t)
{
tokens_.push_back(t);
}
void Parser::pop_back()
{
tokens_.pop_back();
}
Token const & Parser::prevToken() const
{
static const Token dummy;
return pos_ > 0 ? tokens_[pos_ - 1] : dummy;
}
Token const & Parser::nextToken() const
{
static const Token dummy;
return good() ? tokens_[pos_] : dummy;
}
Token const & Parser::getToken()
{
static const Token dummy;
//lyxerr << "looking at token " << tokens_[pos_] << " pos: " << pos_ << '\n';
return good() ? tokens_[pos_++] : dummy;
}
void Parser::skipSpaces()
{
while (nextToken().cat() == catSpace)
getToken();
}
void Parser::putback()
{
--pos_;
}
bool Parser::good() const
{
return pos_ < tokens_.size();
}
char Parser::getChar()
{
if (!good())
error("The input stream is not well...");
return tokens_[pos_++].character();
}
string Parser::getArg(char left, char right)
{
skipSpaces();
string result;
char c = getChar();
if (c != left)
putback();
else
while ((c = getChar()) != right && good())
result += c;
return result;
}
void Parser::tokenize(istream & is)
{
// eat everything up to the next \end_inset or end of stream
// and store it in s for further tokenization
string s;
char c;
while (is.get(c)) {
s += c;
if (s.size() >= 10 && s.substr(s.size() - 10) == "\\end_inset") {
s = s.substr(0, s.size() - 10);
break;
}
}
// tokenize buffer
tokenize(s);
}
void Parser::skipSpaceTokens(istream & is, char c)
{
// skip trailing spaces
while (catcode(c) == catSpace || catcode(c) == catNewline)
if (!is.get(c))
break;
//lyxerr << "putting back: " << c << "\n";
is.putback(c);
}
void Parser::tokenize(string const & buffer)
{
static bool init_done = false;
if (!init_done) {
catInit();
init_done = true;
}
istringstream is(buffer.c_str(), ios::in | ios::binary);
char c;
while (is.get(c)) {
//lyxerr << "reading c: " << c << "\n";
switch (catcode(c)) {
case catNewline: {
++lineno_;
is.get(c);
if (catcode(c) == catNewline)
; //push_back(Token("par"));
else {
push_back(Token(' ', catSpace));
is.putback(c);
}
break;
}
case catComment: {
while (is.get(c) && catcode(c) != catNewline)
;
++lineno_;
break;
}
case catEscape: {
is.get(c);
if (!is) {
error("unexpected end of input");
} else {
string s(1, c);
if (catcode(c) == catLetter) {
// collect letters
while (is.get(c) && catcode(c) == catLetter)
s += c;
skipSpaceTokens(is, c);
}
push_back(Token(s));
}
break;
}
case catSuper:
case catSub: {
push_back(Token(c, catcode(c)));
is.get(c);
skipSpaceTokens(is, c);
break;
}
case catIgnore: {
lyxerr << "ignoring a char: " << int(c) << "\n";
break;
}
default:
push_back(Token(c, catcode(c)));
}
}
#ifdef FILEDEBUG
dump();
#endif
}
void Parser::dump() const
{
lyxerr << "\nTokens: ";
for (unsigned i = 0; i < tokens_.size(); ++i) {
if (i == pos_)
lyxerr << " <#> ";
lyxerr << tokens_[i];
}
lyxerr << " pos: " << pos_ << "\n";
}
void Parser::error(string const & msg)
{
lyxerr << "Line ~" << lineno_ << ": Math parse error: " << msg << endl;
dump();
//exit(1);
}
bool Parser::parse_macro(string & name)
{
int nargs = 0;
name = "{error}";
skipSpaces();
if (nextToken().cs() == "def") {
getToken();
name = getToken().cs();
string pars;
while (good() && nextToken().cat() != catBegin)
pars += getToken().cs();
if (!good()) {
error("bad stream in parse_macro\n");
return false;
}
//lyxerr << "read \\def parameter list '" << pars << "'\n";
if (!pars.empty()) {
error("can't handle non-empty parameter lists\n");
return false;
}
} else if (nextToken().cs() == "newcommand") {
getToken();
if (getToken().cat() != catBegin) {
error("'{' in \\newcommand expected (1) \n");
return false;
}
name = getToken().cs();
if (getToken().cat() != catEnd) {
error("'}' expected\n");
return false;
}
string arg = getArg('[', ']');
if (!arg.empty())
nargs = atoi(arg.c_str());
} else {
lyxerr << "\\newcommand or \\def expected\n";
return false;
}
if (getToken().cat() != catBegin) {
error("'{' in macro definition expected (2)\n");
return false;
}
MathArray ar1;
parse_into(ar1, FLAG_BRACE_LAST, true);
// we cannot handle recursive stuff at all
MathArray test;
test.push_back(createMathInset(name));
if (ar1.contains(test)) {
error("we cannot handle recursive macros at all.\n");
return false;
}
// is a version for display attached?
MathArray ar2;
parse_into(ar2, FLAG_ITEM, true);
MathMacroTable::create(name, nargs, ar1, ar2);
return true;
}
bool Parser::parse_normal(MathAtom & at)
{
MathArray ar;
parse_into(ar, false, false);
if (ar.size() != 1) {
lyxerr << "Unusual contents found: " << ar << endl;
at.reset(new MathParInset);
at->cell(0) = ar;
return true;
}
at = ar[0];
return true;
}
void Parser::parse_into(MathArray & array, unsigned flags, bool mathmode)
{
MathGridInset grid(1, 1);
parse_into1(grid, flags, mathmode, false);
array = grid.cell(0);
// remove 'unnecessary' braces:
if (array.size() == 1 && array.back()->asBraceInset()) {
lyxerr << "extra braces removed\n";
array = array.back()->asBraceInset()->cell(0);
}
}
void Parser::parse_into2(MathAtom & at, unsigned flags,
bool mathmode, bool numbered)
{
parse_into1(*(at->asGridInset()), flags, mathmode, numbered);
}
void Parser::parse_into1(MathGridInset & grid, unsigned flags,
bool mathmode, bool numbered)
{
int limits = 0;
MathGridInset::row_type cellrow = 0;
MathGridInset::col_type cellcol = 0;
MathArray * cell = &grid.cell(grid.index(cellrow, cellcol));
if (grid.asHullInset())
grid.asHullInset()->numbered(cellrow, numbered);
//dump();
//lyxerr << "grid: " << grid << endl;
while (good()) {
Token const & t = getToken();
#ifdef FILEDEBUG
lyxerr << "t: " << t << " flags: " << flags << "\n";
//cell->dump();
lyxerr << "\n";
#endif
if (flags & FLAG_ITEM) {
if (t.cat() == catSpace)
continue;
flags &= ~FLAG_ITEM;
if (t.cat() == catBegin) {
// skip the brace and collect everything to the next matching
// closing brace
flags |= FLAG_BRACE_LAST;
continue;
}
// handle only this single token, leave the loop if done
flags |= FLAG_LEAVE;
}
//
// cat codes
//
if (t.cat() == catMath) {
if (!mathmode) {
// we are inside some text mode thingy, so opening new math is allowed
Token const & n = getToken();
if (n.cat() == catMath) {
// TeX's $$...$$ syntax for displayed math
cell->push_back(MathAtom(new MathHullInset(LM_OT_EQUATION)));
parse_into2(cell->back(), FLAG_SIMPLE, true, false);
getToken(); // skip the second '$' token
} else {
// simple $...$ stuff
putback();
cell->push_back(MathAtom(new MathHullInset(LM_OT_SIMPLE)));
parse_into2(cell->back(), FLAG_SIMPLE, true, false);
}
}
else if (flags & FLAG_SIMPLE) {
// this is the end of the formula
return;
}
else {
error("something strange in the parser\n");
break;
}
}
else if (t.cat() == catLetter)
cell->push_back(MathAtom(new MathCharInset(t.character())));
else if (t.cat() == catSpace && !mathmode)
cell->push_back(MathAtom(new MathCharInset(t.character())));
else if (t.cat() == catParameter) {
Token const & n = getToken();
cell->push_back(MathAtom(new MathMacroArgument(n.character()-'0')));
}
else if (t.cat() == catBegin) {
MathArray ar;
parse_into(ar, FLAG_BRACE_LAST, mathmode);
#ifndef WITH_WARNINGS
#warning this might be wrong in general!
#endif
// ignore braces around simple items
if ((ar.size() == 1 && !ar.front()->needsBraces()
|| (ar.size() == 2 && !ar.front()->needsBraces()
&& ar.back()->asScriptInset()))
|| (ar.size() == 0 && cell->size() == 0))
{
cell->push_back(ar);
} else {
cell->push_back(MathAtom(new MathBraceInset));
cell->back()->cell(0).swap(ar);
}
}
else if (t.cat() == catEnd) {
if (flags & FLAG_BRACE_LAST)
return;
error("found '}' unexpectedly");
//lyx::Assert(0);
//add(cell, '}', LM_TC_TEX);
}
else if (t.cat() == catAlign) {
++cellcol;
//lyxerr << " column now " << cellcol << " max: " << grid.ncols() << "\n";
if (cellcol == grid.ncols()) {
lyxerr << "adding column " << cellcol << "\n";
grid.addCol(cellcol - 1);
}
cell = &grid.cell(grid.index(cellrow, cellcol));
}
else if (t.cat() == catSuper || t.cat() == catSub) {
bool up = (t.cat() == catSuper);
MathScriptInset * p = 0;
if (cell->size())
p = cell->back()->asScriptInset();
if (!p || p->has(up)) {
cell->push_back(MathAtom(new MathScriptInset(up)));
p = cell->back()->asScriptInset();
}
p->ensure(up);
parse_into(p->cell(up), FLAG_ITEM, mathmode);
p->limits(limits);
limits = 0;
}
else if (t.character() == ']' && (flags & FLAG_BRACK_END))
return;
else if (t.cat() == catOther)
cell->push_back(MathAtom(new MathCharInset(t.character())));
//
// control sequences
//
else if (t.cs() == "(") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_SIMPLE)));
parse_into2(cell->back(), FLAG_SIMPLE2, true, false);
}
else if (t.cs() == "[") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_EQUATION)));
parse_into2(cell->back(), FLAG_EQUATION, true, false);
}
else if (t.cs() == "protect")
// ignore \\protect, will hopefully be re-added during output
;
else if (t.cs() == "end") {
if (flags & FLAG_END) {
// eat environment name
//string const name =
getArg('{', '}');
// FIXME: check that we ended the correct environment
return;
}
error("found 'end' unexpectedly");
}
else if (t.cs() == ")") {
if (flags & FLAG_SIMPLE2)
return;
error("found '\\)' unexpectedly");
}
else if (t.cs() == "]") {
if (flags & FLAG_EQUATION)
return;
error("found '\\]' unexpectedly");
}
else if (t.cs() == "\\") {
grid.vcrskip(LyXLength(getArg('[', ']')), cellrow);
++cellrow;
cellcol = 0;
if (cellrow == grid.nrows())
grid.addRow(cellrow - 1);
if (grid.asHullInset())
grid.asHullInset()->numbered(cellrow, numbered);
cell = &grid.cell(grid.index(cellrow, cellcol));
}
#if 1
else if (t.cs() == "multicolumn") {
// extract column count and insert dummy cells
MathArray count;
parse_into(count, FLAG_ITEM, mathmode);
int cols = 1;
if (!extractNumber(count, cols)) {
lyxerr << " can't extract number of cells from " << count << "\n";
}
// resize the table if necessary
for (int i = 0; i < cols; ++i) {
++cellcol;
if (cellcol == grid.ncols()) {
lyxerr << "adding column " << cellcol << "\n";
grid.addCol(cellcol - 1);
}
cell = &grid.cell(grid.index(cellrow, cellcol));
// mark this as dummy
grid.cellinfo(grid.index(cellrow, cellcol)).dummy_ = true;
}
// the last cell is the real thng, not a dummy
grid.cellinfo(grid.index(cellrow, cellcol)).dummy_ = false;
// read special alignment
MathArray align;
parse_into(align, FLAG_ITEM, mathmode);
//grid.cellinfo(grid.index(cellrow, cellcol)).align_ = extractString(align);
// parse the remaining contents into the "real" cell
parse_into(*cell, FLAG_ITEM, mathmode);
}
#endif
else if (t.cs() == "limits")
limits = 1;
else if (t.cs() == "nolimits")
limits = -1;
else if (t.cs() == "nonumber") {
if (grid.asHullInset())
grid.asHullInset()->numbered(cellrow, false);
}
else if (t.cs() == "number") {
if (grid.asHullInset())
grid.asHullInset()->numbered(cellrow, true);
}
else if (t.cs() == "hline") {
if (grid.asHullInset())
grid.asHullInset()->rowinfo(cellrow + 1);
}
else if (t.cs() == "sqrt") {
char c = getChar();
if (c == '[') {
cell->push_back(MathAtom(new MathRootInset));
parse_into(cell->back()->cell(0), FLAG_BRACK_END, mathmode);
parse_into(cell->back()->cell(1), FLAG_ITEM, mathmode);
} else {
putback();
cell->push_back(MathAtom(new MathSqrtInset));
parse_into(cell->back()->cell(0), FLAG_ITEM, mathmode);
}
}
else if (t.cs() == "ref") {
cell->push_back(MathAtom(new RefInset));
char c = getChar();
if (c == '[')
parse_into(cell->back()->cell(1), FLAG_BRACK_END, mathmode);
else
putback();
parse_into(cell->back()->cell(0), FLAG_ITEM, mathmode);
}
else if (t.cs() == "left") {
string l = getToken().asString();
MathArray ar;
parse_into(ar, FLAG_RIGHT, mathmode);
string r = getToken().asString();
cell->push_back(MathAtom(new MathDelimInset(l, r, ar)));
}
else if (t.cs() == "right") {
if (flags & FLAG_RIGHT)
return;
//lyxerr << "got so far: '" << cell << "'\n";
error("Unmatched right delimiter");
return;
}
else if (t.cs() == "begin") {
string const name = getArg('{', '}');
if (name == "array" || name == "subarray") {
string const valign = getArg('[', ']') + 'c';
string const halign = getArg('{', '}');
cell->push_back(MathAtom(new MathArrayInset(name, valign[0], halign)));
parse_into2(cell->back(), FLAG_END, mathmode, false);
}
else if (name == "split" || name == "cases" ||
name == "gathered" || name == "aligned") {
cell->push_back(createMathInset(name));
parse_into2(cell->back(), FLAG_END, mathmode, false);
}
else if (name == "math") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_SIMPLE)));
parse_into2(cell->back(), FLAG_SIMPLE, true, true);
}
else if (name == "equation" || name == "equation*"
|| name == "displaymath") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_EQUATION)));
parse_into2(cell->back(), FLAG_END, true, (name == "equation"));
}
else if (name == "eqnarray" || name == "eqnarray*") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_EQNARRAY)));
parse_into2(cell->back(), FLAG_END, true, !stared(name));
}
else if (name == "align" || name == "align*") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_ALIGN)));
parse_into2(cell->back(), FLAG_END, true, !stared(name));
}
else if (name == "alignat" || name == "alignat*") {
// ignore this for a while
getArg('{', '}');
cell->push_back(MathAtom(new MathHullInset(LM_OT_ALIGNAT)));
parse_into2(cell->back(), FLAG_END, true, !stared(name));
}
else if (name == "xalignat" || name == "xalignat*") {
// ignore this for a while
getArg('{', '}');
cell->push_back(MathAtom(new MathHullInset(LM_OT_XALIGNAT)));
parse_into2(cell->back(), FLAG_END, true, !stared(name));
}
else if (name == "xxalignat") {
// ignore this for a while
getArg('{', '}');
cell->push_back(MathAtom(new MathHullInset(LM_OT_XXALIGNAT)));
parse_into2(cell->back(), FLAG_END, true, !stared(name));
}
else if (name == "multline" || name == "multline*") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_MULTLINE)));
parse_into2(cell->back(), FLAG_END, true, !stared(name));
}
else if (name == "gather" || name == "gather*") {
cell->push_back(MathAtom(new MathHullInset(LM_OT_GATHER)));
parse_into2(cell->back(), FLAG_END, true, !stared(name));
}
else {
latexkeys const * l = in_word_set(name);
if (l) {
if (l->inset == "matrix") {
cell->push_back(createMathInset(name));
parse_into2(cell->back(), FLAG_END, mathmode, false);
}
} else {
lyxerr << "unknow math inset begin '" << name << "'\n";
}
}
}
else if (t.cs() == "kern") {
#ifdef WITH_WARNINGS
#warning A hack...
#endif
string s;
while (1) {
Token const & t = getToken();
if (!good()) {
putback();
break;
}
s += t.character();
if (isValidLength(s))
break;
}
cell->push_back(MathAtom(new MathKernInset(s)));
}
else if (t.cs() == "label") {
if (grid.asHullInset())
grid.asHullInset()->label(cellrow, getArg('{', '}'));
}
else if (t.cs() == "choose" || t.cs() == "over" || t.cs() == "atop") {
MathAtom p = createMathInset(t.cs());
cell->swap(p->cell(0));
parse_into(p->cell(1), flags, mathmode);
cell->push_back(p);
return;
}
else if (t.cs() == "substack") {
cell->push_back(createMathInset(t.cs()));
parse_into2(cell->back(), FLAG_ITEM, mathmode, false);
}
else if (t.cs() == "xymatrix") {
cell->push_back(createMathInset(t.cs()));
parse_into2(cell->back(), FLAG_ITEM, mathmode, false);
}
#if 0
// Disabled
else if (1 && t.cs() == "ar") {
MathXYArrowInset * p = new MathXYArrowInset;
// try to read target
char c = getChar();
if (c == '[') {
parse_into(p->cell(0), FLAG_BRACK_END, mathmode);
//lyxerr << "read target: " << p->cell(0) << "\n";
} else {
putback();
}
// try to read label
if (nextToken().cat() == catSuper || nextToken().cat() == catSub) {
p->up_ = nextToken().cat() == catSuper;
getToken();
parse_into(p->cell(1), FLAG_ITEM, mathmode);
//lyxerr << "read label: " << p->cell(1) << "\n";
}
cell->push_back(MathAtom(p));
//lyxerr << "read cell: " << cell << "\n";
}
#endif
else if (t.cs().size()) {
latexkeys const * l = in_word_set(t.cs());
if (l) {
if (l->inset == "font") {
lyxerr << "starting font " << t.cs() << "\n";
MathAtom p = createMathInset(t.cs());
bool textmode = (t.cs()[0] == 't');
parse_into(p->cell(0), FLAG_ITEM, !textmode);
cell->push_back(p);
//lyxerr << "ending font\n";
}
else if (l->inset == "oldfont") {
cell->push_back(createMathInset(t.cs()));
parse_into(cell->back()->cell(0), flags, l->extra == "mathmode");
return;
}
else if (l->inset == "box") {
// switch to text mode
cell->push_back(createMathInset(t.cs()));
parse_into(cell->back()->cell(0), FLAG_ITEM, mathmode);
}
else if (l->inset == "style") {
cell->push_back(createMathInset(t.cs()));
parse_into(cell->back()->cell(0), flags, mathmode);
return;
}
else {
MathAtom p = createMathInset(t.cs());
for (MathInset::idx_type i = 0; i < p->nargs(); ++i)
parse_into(p->cell(i), FLAG_ITEM, l->extra == "mathmode");
cell->push_back(p);
}
}
else {
MathAtom p = createMathInset(t.cs());
for (MathInset::idx_type i = 0; i < p->nargs(); ++i)
parse_into(p->cell(i), FLAG_ITEM, mathmode);
cell->push_back(p);
}
}
if (flags & FLAG_LEAVE) {
flags &= ~FLAG_LEAVE;
break;
}
}
}
} // anonymous namespace
void mathed_parse_cell(MathArray & ar, string const & str)
{
istringstream is(str.c_str());
mathed_parse_cell(ar, is);
}
void mathed_parse_cell(MathArray & ar, istream & is)
{
Parser(is).parse_into(ar, 0, true);
}
bool mathed_parse_macro(string & name, string const & str)
{
istringstream is(str.c_str());
Parser parser(is);
return parser.parse_macro(name);
}
bool mathed_parse_macro(string & name, istream & is)
{
Parser parser(is);
return parser.parse_macro(name);
}
bool mathed_parse_macro(string & name, LyXLex & lex)
{
Parser parser(lex);
return parser.parse_macro(name);
}
bool mathed_parse_normal(MathAtom & t, string const & str)
{
istringstream is(str.c_str());
Parser parser(is);
return parser.parse_normal(t);
}
bool mathed_parse_normal(MathAtom & t, istream & is)
{
Parser parser(is);
return parser.parse_normal(t);
}
bool mathed_parse_normal(MathAtom & t, LyXLex & lex)
{
Parser parser(lex);
return parser.parse_normal(t);
}