X-Git-Url: https://git.lyx.org/gitweb/?a=blobdiff_plain;f=src%2FLexer.cpp;h=c4b738f119bcf13fe903d8a74df6e53e1514bc28;hb=f5b829bc8abf2e777d8dfc93400ae603d26d3a22;hp=530e4f12f041e72ebe24a927b74e15845b03d1d0;hpb=11a0458d3f66bce83b90e23f1c85175a8ffcae45;p=lyx.git diff --git a/src/Lexer.cpp b/src/Lexer.cpp index 530e4f12f0..c4b738f119 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 * @@ -14,12 +14,14 @@ #include #include "Lexer.h" +#include "Format.h" #include "support/convert.h" #include "support/debug.h" #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" @@ -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(); /// @@ -102,6 +104,8 @@ public: string pushTok; /// char commentChar; + /// used for error messages + string context; private: /// non-copyable Pimpl(Pimpl const &); @@ -171,7 +175,9 @@ 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; } @@ -233,9 +239,7 @@ 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") { + if (formats.isZippedFile(filename)) { 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 @@ -245,9 +249,10 @@ bool Lexer::Pimpl::setFile(FileName const & filename) "file or stream already set."); gz_.open(filename.toFilesystemEncoding().c_str(), ios::in); is.rdbuf(&gz_); - name = filename.absFilename(); + name = filename.absFileName(); lineno = 0; - return gz_.is_open() && is.good(); + if (!gz_.is_open() || !is.good()) + return false; } else { LYXERR(Debug::LYXLEX, "lyxlex: UNcompressed"); @@ -258,12 +263,25 @@ bool Lexer::Pimpl::setFile(FileName const & filename) LYXERR(Debug::LYXLEX, "Error in Lexer::setFile: " "file or stream already set."); } - fb_.open(filename.toFilesystemEncoding().c_str(), ios::in); + fb_.open(filename.toSafeFilesystemEncoding().c_str(), ios::in); is.rdbuf(&fb_); - name = filename.absFilename(); + name = filename.absFileName(); lineno = 0; - return fb_.is_open() && is.good(); + if (!fb_.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; } @@ -300,121 +318,35 @@ bool Lexer::Pimpl::next(bool esc /* = false */) status = LEX_TOKEN; return true; } - if (!esc) { - unsigned char c = 0; // getc() returns an int - char cc = 0; - status = 0; - while (is && !status) { - is.get(cc); - c = cc; - if (c == commentChar) { - // Read rest of line (fast :-) -#if 1 - // That is not fast... (Lgb) - string dummy; - getline(is, dummy); - - LYXERR(Debug::LYXLEX, "Comment read: `" << c << dummy << '\''); -#else - // unfortunately ignore is buggy (Lgb) - is.ignore(100, '\n'); -#endif - ++lineno; - continue; - } - - if (c == '\"') { - buff.clear(); - - do { - is.get(cc); - c = cc; - if (c != '\r') - buff.push_back(c); - } while (c != '\"' && c != '\n' && is); - - if (c != '\"') { - printError("Missing quote"); - if (c == '\n') - ++lineno; - } - - buff.resize(buff.size()-1); - status = LEX_DATA; - break; - } - - 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 - if (c > ' ' && is) { - buff.clear(); - - do { - buff.push_back(c); - is.get(cc); - c = cc; - } while (c > ' ' && c != ',' && is); - - status = LEX_TOKEN; - } - 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 - // when it expected to find a '\n' - is.get(cc); - c = cc; - } - if (c == '\n') - ++lineno; - - } - if (status) - return true; - - status = is.eof() ? LEX_FEOF: LEX_UNDEF; - buff.clear(); - return false; - } else { - unsigned char c = 0; // getc() returns an int - char cc = 0; - - status = 0; - while (is && !status) { - is.get(cc); - c = cc; - - // skip ','s - if (c == ',') - continue; + unsigned char c = 0; // getc() returns an int + char cc = 0; + status = 0; + while (is && !status) { + is.get(cc); + c = cc; - if (c == commentChar) { - // Read rest of line (fast :-) + if (c == commentChar) { + // Read rest of line (fast :-) #if 1 - // That is still not fast... (Lgb) - string dummy; - getline(is, dummy); + // That is not fast... (Lgb) + string dummy; + getline(is, dummy); - LYXERR(Debug::LYXLEX, "Comment read: `" << c << dummy << '\''); + LYXERR(Debug::LYXLEX, "Comment read: `" << c << dummy << '\''); #else - // but ignore is also still buggy (Lgb) - // This is fast (Lgb) - is.ignore(100, '\n'); + // unfortunately ignore is buggy (Lgb) + is.ignore(100, '\n'); #endif - ++lineno; - continue; - } + ++lineno; + continue; + } + + if (c == '\"') { + buff.clear(); - // string - if (c == '\"') { - buff.clear(); + if (esc) { bool escaped = false; do { @@ -437,50 +369,75 @@ bool Lexer::Pimpl::next(bool esc /* = false */) break; } while (c != '\n' && is); - if (c != '\"') { - printError("Missing quote"); - if (c == '\n') - ++lineno; - } + } else { + + do { + is.get(cc); + c = cc; + if (c != '\r') + buff.push_back(c); + } while (c != '\"' && c != '\n' && is); - buff.resize(buff.size() -1); - status = LEX_DATA; - break; } - if (c > ' ' && is) { - buff.clear(); + if (c != '\"') { + printError("Missing quote"); + if (c == '\n') + ++lineno; + } - do { - if (c == '\\') { - // escape the next char - is.get(cc); - c = cc; - //escaped = true; - } - buff.push_back(c); + buff.resize(buff.size() - 1); + status = LEX_DATA; + break; + } + + 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 + if (c > ' ' && is) { + buff.clear(); + + do { + if (esc && c == '\\') { + // escape the next char is.get(cc); c = cc; - } while (c > ' ' && c != ',' && is); + //escaped = true; + } + buff.push_back(c); + is.get(cc); + c = cc; + } while (c > ' ' && c != ',' && is); + status = LEX_TOKEN; + } - status = LEX_TOKEN; - } - // new line - if (c == '\n') - ++lineno; + 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 + // when it expected to find a '\n' + is.get(cc); + c = cc; } - if (status) - return true; + if (c == '\n') + ++lineno; - status = is.eof() ? LEX_FEOF : LEX_UNDEF; - buff.clear(); - return false; } + if (status) + return true; + + status = is.eof() ? LEX_FEOF: LEX_UNDEF; + buff.clear(); + return 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 = @@ -500,7 +457,7 @@ 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; } @@ -515,7 +472,7 @@ 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); } @@ -556,7 +513,7 @@ 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 == '\\' @@ -570,7 +527,7 @@ 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 == '\\') @@ -612,7 +569,7 @@ void Lexer::Pimpl::pushToken(string const & pt) ////////////////////////////////////////////////////////////////////// Lexer::Lexer() - : pimpl_(new Pimpl(0, 0)) + : pimpl_(new Pimpl(0, 0)), lastReadOk_(false) {} @@ -693,6 +650,7 @@ void Lexer::setCommentChar(char c) pimpl_->setCommentChar(c); } + int Lexer::lex() { return pimpl_->lex(); @@ -777,7 +735,7 @@ string const Lexer::getLongString(string const & endtoken) LYXERR(Debug::PARSER, "LongString: `" << getString() << '\''); - // We do a case independent comparison, like search_kw does. + // We do a case independent comparison, like searchKeyword does. if (compare_ascii_no_case(token, endtoken) == 0) break; @@ -807,15 +765,18 @@ 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; } @@ -932,9 +893,19 @@ 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 += '"'; @@ -944,4 +915,46 @@ 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 & str) +{ + pimpl_->context = str; +} + + } // namespace lyx