X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FLexer.cpp;h=c6402aec6f39f69e966b6afe17710d44aebeb5e8;hb=8f43f83ddfbe6adbb04a2ba86e62797c0a313324;hp=a069c93e471117d19dc7860c344c773009ede7d5;hpb=13b5a9e3ceaa63b1adf3727d56d3d9b3890c00f5;p=lyx.git diff --git a/src/Lexer.cpp b/src/Lexer.cpp index a069c93e47..c6402aec6f 100644 --- a/src/Lexer.cpp +++ b/src/Lexer.cpp @@ -4,7 +4,7 @@ * Licence details can be found in the file COPYING. * * \author Alejandro Aguilar Sierra - * \author Lars Gullik Bjønnes + * \author Lars Gullik Bjønnes * \author Jean-Marc Lasgouttes * \author John Levon * @@ -20,11 +20,13 @@ #include "support/FileName.h" #include "support/filetools.h" #include "support/gzstream.h" +#include "support/lassert.h" #include "support/lstrings.h" #include "support/lyxalgo.h" -#include "support/types.h" +#include // sort, lower_bound #include +#include #include #include #include @@ -67,7 +69,7 @@ public: /// bool next(bool esc = false); /// - int search_kw(char const * const tag) const; + int searchKeyword(char const * const tag) const; /// int lex(); /// @@ -78,9 +80,6 @@ public: bool inputAvailable(); /// void pushToken(string const &); - /// fb_ is only used to open files, the stream is accessed through is. - filebuf fb_; - /// gz_ is only used to open files, the stream is accessed through is. gz::gzstreambuf gz_; @@ -100,6 +99,8 @@ public: int lineno; /// string pushTok; + /// used for error messages + string context; /// char commentChar; private: @@ -114,7 +115,7 @@ private: public: /// PushedTable() - : table_elem(0), table_siz(0) {} + : table_elem(nullptr), table_siz(0) {} /// PushedTable(LexerKeyword * ki, int siz) : table_elem(ki), table_siz(siz) {} @@ -128,27 +129,23 @@ private: }; - namespace { -class CompareTags - : public binary_function { -public: - // used by lower_bound, sort and sorted - bool operator()(LexerKeyword const & a, LexerKeyword const & b) const - { - // we use the ascii version, because in turkish, 'i' - // is not the lowercase version of 'I', and thus - // turkish locale breaks parsing of tags. - return compare_ascii_no_case(a.tag, b.tag) < 0; - } -}; +// used by lower_bound, sort and sorted +bool compareTags(LexerKeyword const & a, LexerKeyword const & b) +{ + // we use the ascii version, because in turkish, 'i' + // is not the lowercase version of 'I', and thus + // turkish locale breaks parsing of tags. + return compare_ascii_no_case(a.tag, b.tag) < 0; +} + +} // namespace -} // end of anon namespace Lexer::Pimpl::Pimpl(LexerKeyword * tab, int num) - : is(&fb_), table(tab), no_items(num), + : is(&gz_), table(tab), no_items(num), status(0), lineno(0), commentChar('#') { verifyTable(); @@ -171,7 +168,9 @@ void Lexer::Pimpl::printError(string const & message) const { string const tmpmsg = subst(message, "$$Token", getString()); lyxerr << "LyX: " << tmpmsg << " [around line " << lineno - << " of file " << to_utf8(makeDisplayPath(name)) << ']' << endl; + << " of file " << to_utf8(makeDisplayPath(name)) + << " current token: '" << getString() << "'" + << " context: '" << context << "']" << endl; } @@ -190,14 +189,14 @@ void Lexer::Pimpl::verifyTable() { // Check if the table is sorted and if not, sort it. if (table - && !lyx::sorted(table, table + no_items, CompareTags())) { + && !lyx::sorted(table, table + no_items, &compareTags)) { lyxerr << "The table passed to Lexer is not sorted!\n" << "Tell the developers to fix it!" << endl; // We sort it anyway to avoid problems. lyxerr << "\nUnsorted:" << endl; printTable(lyxerr); - sort(table, table + no_items, CompareTags()); + sort(table, table + no_items, &compareTags); lyxerr << "\nSorted:" << endl; printTable(lyxerr); } @@ -232,47 +231,33 @@ void Lexer::Pimpl::popTable() bool Lexer::Pimpl::setFile(FileName const & filename) { - // Check the format of the file. - string const format = filename.guessFormatFromContents(); - - if (format == "gzip" || format == "zip" || format == "compress") { - LYXERR(Debug::LYXLEX, "lyxlex: compressed"); - // The check only outputs a debug message, because it triggers - // a bug in compaq cxx 6.2, where is_open() returns 'true' for - // a fresh new filebuf. (JMarc) if (gz_.is_open() || istream::off_type(is.tellg()) > -1) - LYXERR(Debug::LYXLEX, "Error in LyXLex::setFile: " - "file or stream already set."); + LYXERR0("Error in LyXLex::setFile: file or stream already set."); gz_.open(filename.toFilesystemEncoding().c_str(), ios::in); is.rdbuf(&gz_); - name = filename.absFilename(); - lineno = 0; - return gz_.is_open() && is.good(); - } else { - LYXERR(Debug::LYXLEX, "lyxlex: UNcompressed"); - - // The check only outputs a debug message, because it triggers - // a bug in compaq cxx 6.2, where is_open() returns 'true' for - // a fresh new filebuf. (JMarc) - if (fb_.is_open() || istream::off_type(is.tellg()) > 0) { - LYXERR(Debug::LYXLEX, "Error in Lexer::setFile: " - "file or stream already set."); - } - fb_.open(filename.toFilesystemEncoding().c_str(), ios::in); - is.rdbuf(&fb_); - name = filename.absFilename(); + name = filename.absFileName(); lineno = 0; - return fb_.is_open() && is.good(); + if (!gz_.is_open() || !is.good()) + return false; + + // Skip byte order mark. + if (is.peek() == 0xef) { + is.get(); + if (is.peek() == 0xbb) { + is.get(); + LASSERT(is.get() == 0xbf, /**/); + } else + is.unget(); } + + return true; } void Lexer::Pimpl::setStream(istream & i) { - if (fb_.is_open() || istream::off_type(is.tellg()) > 0) { - LYXERR(Debug::LYXLEX, "Error in Lexer::setStream: " - "file or stream already set."); - } + if (gz_.is_open() || istream::off_type(is.tellg()) > 0) + LYXERR0("Error in Lexer::setStream: file or stream already set."); is.rdbuf(i.rdbuf()); lineno = 0; } @@ -302,18 +287,12 @@ bool Lexer::Pimpl::next(bool esc /* = false */) } - unsigned char c = 0; // getc() returns an int char cc = 0; status = 0; while (is && !status) { is.get(cc); - c = cc; - - // skip ','s - if (esc && c == ',') - continue; + unsigned char c = cc; - if (c == commentChar) { // Read rest of line (fast :-) #if 1 @@ -335,9 +314,8 @@ bool Lexer::Pimpl::next(bool esc /* = false */) if (esc) { - bool escaped = false; do { - escaped = false; + bool escaped = false; is.get(cc); c = cc; if (c == '\r') continue; @@ -378,13 +356,13 @@ bool Lexer::Pimpl::next(bool esc /* = false */) break; } - if (!esc && c == ',') + if (c == ',') continue; /* Skip ','s */ - // using relational operators with chars other - // than == and != is not safe. And if it is done - // the type _have_ to be unsigned. It usually a - // lot better to use the functions from cctype + // using relational operators with chars other + // than == and != is not safe. And if it is done + // the type _have_ to be unsigned. It usually a + // lot better to use the functions from cctype if (c > ' ' && is) { buff.clear(); @@ -402,7 +380,7 @@ bool Lexer::Pimpl::next(bool esc /* = false */) status = LEX_TOKEN; } - if (!esc && c == '\r' && is) { + if (c == '\r' && is) { // The Windows support has lead to the // possibility of "\r\n" at the end of // a line. This will stop LyX choking @@ -424,12 +402,12 @@ bool Lexer::Pimpl::next(bool esc /* = false */) } -int Lexer::Pimpl::search_kw(char const * const tag) const +int Lexer::Pimpl::searchKeyword(char const * const tag) const { LexerKeyword search_tag = { tag, 0 }; LexerKeyword * res = lower_bound(table, table + no_items, - search_tag, CompareTags()); + search_tag, &compareTags); // use the compare_ascii_no_case instead of compare_no_case, // because in turkish, 'i' is not the lowercase version of 'I', // and thus turkish locale breaks parsing of tags. @@ -444,7 +422,7 @@ int Lexer::Pimpl::lex() { //NOTE: possible bug. if (next() && status == LEX_TOKEN) - return search_kw(getString().c_str()); + return searchKeyword(getString().c_str()); return status; } @@ -459,7 +437,7 @@ bool Lexer::Pimpl::eatLine() is.get(cc); c = cc; //LYXERR(Debug::LYXLEX, "Lexer::EatLine read char: `" << c << '\''); - if (c != '\r') + if (c != '\r' && is) buff.push_back(c); } @@ -500,7 +478,7 @@ bool Lexer::Pimpl::nextToken() char cc = 0; is.get(cc); c = cc; - if (c >= ' ' && is) { + if ((c >= ' ' || c == '\t') && is) { buff.clear(); if (c == '\\') { // first char == '\\' @@ -514,7 +492,7 @@ bool Lexer::Pimpl::nextToken() buff.push_back(c); is.get(cc); c = cc; - } while (c >= ' ' && c != '\\' && is); + } while ((c >= ' ' || c == '\t') && c != '\\' && is); } if (c == '\\') @@ -556,7 +534,7 @@ void Lexer::Pimpl::pushToken(string const & pt) ////////////////////////////////////////////////////////////////////// Lexer::Lexer() - : pimpl_(new Pimpl(0, 0)) + : pimpl_(new Pimpl(nullptr, 0)), lastReadOk_(false) {} @@ -637,6 +615,7 @@ void Lexer::setCommentChar(char c) pimpl_->setCommentChar(c); } + int Lexer::lex() { return pimpl_->lex(); @@ -681,23 +660,23 @@ double Lexer::getFloat() const } -string const Lexer::getString() const +string const Lexer::getString(bool trim) const { lastReadOk_ = pimpl_->status == LEX_DATA || pimpl_->status == LEX_TOKEN; if (lastReadOk_) - return pimpl_->getString(); + return trim ? support::trim(pimpl_->getString(), "\t ") : pimpl_->getString(); return string(); } -docstring const Lexer::getDocString() const +docstring const Lexer::getDocString(bool trim) const { lastReadOk_ = pimpl_->status == LEX_DATA || pimpl_->status == LEX_TOKEN; if (lastReadOk_) - return pimpl_->getDocString(); + return trim ? support::trim(pimpl_->getDocString(), "\t ") : pimpl_->getDocString(); return docstring(); } @@ -706,28 +685,27 @@ docstring const Lexer::getDocString() const // I would prefer to give a tag number instead of an explicit token // here, but it is not possible because Buffer::readDocument uses // explicit tokens (JMarc) -string const Lexer::getLongString(string const & endtoken) +docstring Lexer::getLongString(docstring const & endtoken) { - string str; - string prefix; + docstring str; + docstring prefix; bool firstline = true; while (pimpl_->is) { //< eatLine only reads from is, not from pushTok if (!eatLine()) // blank line in the file being read continue; + docstring tmpstr = getDocString(); + docstring const token = trim(tmpstr, " \t"); - string const token = trim(getString(), " \t"); - - LYXERR(Debug::PARSER, "LongString: `" << getString() << '\''); + LYXERR(Debug::PARSER, "LongString: `" << tmpstr << '\''); - // We do a case independent comparison, like search_kw does. - if (compare_ascii_no_case(token, endtoken) == 0) + // We do a case independent comparison, like searchKeyword does. + if (compare_no_case(token, endtoken) == 0) break; - string tmpstr = getString(); if (firstline) { - size_t i = tmpstr.find_first_not_of(' '); + size_t i = tmpstr.find_first_not_of(from_ascii(" \t")); if (i != string::npos) prefix = tmpstr.substr(0, i); firstline = false; @@ -736,14 +714,14 @@ string const Lexer::getLongString(string const & endtoken) // further lines in long strings may have the same // whitespace prefix as the first line. Remove it. - if (prefix.length() && prefixIs(tmpstr, prefix)) - tmpstr.erase(0, prefix.length() - 1); + if (!prefix.empty() && prefixIs(tmpstr, prefix)) + tmpstr.erase(0, prefix.length()); - str += ltrim(tmpstr, "\t") + '\n'; + str += tmpstr + '\n'; } if (!pimpl_->is) - printError("Long string not ended by `" + endtoken + '\''); + printError("Long string not ended by `" + to_utf8(endtoken) + '\''); return str; } @@ -751,15 +729,18 @@ string const Lexer::getLongString(string const & endtoken) bool Lexer::getBool() const { - if (pimpl_->getString() == "true") { + string const s = pimpl_->getString(); + if (s == "false" || s == "0") { + lastReadOk_ = true; + return false; + } + if (s == "true" || s == "1") { lastReadOk_ = true; return true; - } else if (pimpl_->getString() != "false") { - pimpl_->printError("Bad boolean `$$Token'. " - "Use \"false\" or \"true\""); - lastReadOk_ = false; } - lastReadOk_ = true; + pimpl_->printError("Bad boolean `$$Token'. " + "Use \"false\" or \"true\""); + lastReadOk_ = false; return false; } @@ -794,7 +775,7 @@ Lexer::operator void const *() const // use fail() here. However, our implementation of getString() et al. // can cause the eof() and fail() bits to be set, even though we // haven't tried to read 'em. - return lastReadOk_? this : 0; + return lastReadOk_? this : nullptr; } @@ -876,9 +857,19 @@ Lexer & Lexer::operator>>(bool & s) } +Lexer & Lexer::operator>>(char & c) +{ + string s; + operator>>(s); + if (!s.empty()) + c = s[0]; + return *this; +} + + // quotes a string, e.g. for use in preferences files or as an argument // of the "log" dialog -string const Lexer::quoteString(string const & arg) +string Lexer::quoteString(string const & arg) { string res; res += '"'; @@ -888,4 +879,46 @@ string const Lexer::quoteString(string const & arg) } +// same for docstring +docstring Lexer::quoteString(docstring const & arg) +{ + docstring res; + res += '"'; + res += subst(subst(arg, from_ascii("\\"), from_ascii("\\\\")), + from_ascii("\""), from_ascii("\\\"")); + res += '"'; + return res; +} + + +Lexer & Lexer::operator>>(char const * required) +{ + string token; + *this >> token; + if (token != required) { + LYXERR0("Missing '" << required << "'-tag in " << pimpl_->context + << ". Got " << token << " instead. Line: " << lineNumber()); + pushToken(token); + } + return *this; +} + + +bool Lexer::checkFor(char const * required) +{ + string token; + *this >> token; + if (token == required) + return true; + pushToken(token); + return false; +} + + +void Lexer::setContext(std::string const & functionName) +{ + pimpl_->context = functionName; +} + + } // namespace lyx