-/** The .tex to .lyx converter
- \author André Pönitz (2003)
+/**
+ * \file tex2lyx.C
+ * This file is part of LyX, the document processor.
+ * Licence details can be found in the file COPYING.
+ *
+ * \author André Pönitz
+ *
+ * Full author contact details are available in file CREDITS.
*/
// {[(
+#include <config.h>
+
#include "tex2lyx.h"
+#include "context.h"
#include "debug.h"
-#include "lyx_main.h"
#include "lyxtextclass.h"
-#include "support/path_defines.h"
+
+#include "support/convert.h"
+#include "support/filetools.h"
+#include "support/fs_extras.h"
+#include "support/lstrings.h"
+#include "support/lyxlib.h"
#include "support/os.h"
+#include "support/package.h"
+
+#include <boost/function.hpp>
+#include <boost/filesystem/operations.hpp>
+#include <boost/filesystem/path.hpp>
#include <cctype>
#include <fstream>
#include <string>
#include <sstream>
#include <vector>
+#include <map>
+using std::endl;
using std::cout;
using std::cerr;
-using std::endl;
using std::getline;
-using std::istream;
+
using std::ifstream;
+using std::ofstream;
using std::istringstream;
-using std::ostream;
using std::ostringstream;
using std::stringstream;
using std::string;
using std::vector;
+using std::map;
-// Hacks to allow the thing to link in the lyxlayout stuff
-Debug::type const Debug::ANY = Debug::type(0);
-DebugStream lyxerr;
+using lyx::support::changeExtension;
+using lyx::support::isStrUnsignedInt;
+using lyx::support::ltrim;
+using lyx::support::makeAbsPath;
+using lyx::support::onlyPath;
+using lyx::support::rtrim;
+using lyx::support::isFileReadable;
-void LyX::emergencyCleanup() {}
+namespace fs = boost::filesystem;
-void handle_comment(Parser & p)
-{
- string s;
- while (p.good()) {
- Token const & t = p.get_token();
- if (t.cat() == catNewline)
- break;
- s += t.asString();
- }
- //cerr << "comment: " << s << "\n";
- p.skip_spaces();
-}
+// Hacks to allow the thing to link in the lyxlayout stuff
+LyXErr lyxerr(std::cerr.rdbuf());
string const trim(string const & a, char const * p)
{
- // lyx::Assert(p);
+ // BOOST_ASSERT(p);
if (a.empty() || !*p)
return a;
}
-char const ** is_known(string const & str, char const ** what)
+char const * const * is_known(string const & str, char const * const * what)
{
for ( ; *what; ++what)
if (str == *what)
}
-void clean_layouts(istream & is, ostream & os)
+CommandMap known_commands;
+CommandMap known_environments;
+CommandMap known_math_environments;
+
+
+void add_known_command(string const & command, string const & o1,
+ bool o2)
{
- string last;
- string line;
- bool eating = false;
- while (getline(is, line)) {
- string tline = trim(line, " ");
- if (line.substr(0, 8) == "\\layout ") {
- //cerr << "layout: " << line << "\n";
- last = line;
- eating = true;
- } else if (eating && tline.empty()) {
- //cerr << "eat empty line\n";
- } else if (line.substr(0, 13) == "\\begin_deeper") {
- os << line << "\n";
+ // We have to handle the following cases:
+ // definition o1 o2 invocation result
+ // \newcommand{\foo}{bar} "" false \foo bar
+ // \newcommand{\foo}[1]{bar #1} "[1]" false \foo{x} bar x
+ // \newcommand{\foo}[1][]{bar #1} "[1]" true \foo bar
+ // \newcommand{\foo}[1][]{bar #1} "[1]" true \foo[x] bar x
+ // \newcommand{\foo}[1][x]{bar #1} "[1]" true \foo[x] bar x
+ unsigned int nargs = 0;
+ vector<ArgumentType> arguments;
+ string const opt1 = rtrim(ltrim(o1, "["), "]");
+ if (isStrUnsignedInt(opt1)) {
+ // The command has arguments
+ nargs = convert<unsigned int>(opt1);
+ if (nargs > 0 && o2) {
+ // The first argument is optional
+ arguments.push_back(optional);
+ --nargs;
+ }
+ }
+ for (unsigned int i = 0; i < nargs; ++i)
+ arguments.push_back(required);
+ known_commands[command] = arguments;
+}
+
+
+namespace {
+
+
+/*!
+ * Read one command definition from the syntax file
+ */
+void read_command(Parser & p, string command, CommandMap & commands) {
+ if (p.next_token().asInput() == "*") {
+ p.get_token();
+ command += '*';
+ }
+ vector<ArgumentType> arguments;
+ while (p.next_token().cat() == catBegin ||
+ p.next_token().asInput() == "[") {
+ if (p.next_token().cat() == catBegin) {
+ string const arg = p.getArg('{', '}');
+ if (arg == "translate")
+ arguments.push_back(required);
+ else
+ arguments.push_back(verbatim);
} else {
- // ordinary line
- //cerr << "ordinary line\n";
- if (eating) {
- eating = false;
- os << last << "\n\n";
+ p.getArg('[', ']');
+ arguments.push_back(optional);
+ }
+ }
+ commands[command] = arguments;
+}
+
+
+/*!
+ * Read a class of environments from the syntax file
+ */
+void read_environment(Parser & p, string const & begin,
+ CommandMap & environments)
+{
+ string environment;
+ while (p.good()) {
+ Token const & t = p.get_token();
+ if (t.cat() == catLetter)
+ environment += t.asInput();
+ else if (!environment.empty()) {
+ p.putback();
+ read_command(p, environment, environments);
+ environment.erase();
+ }
+ if (t.cat() == catEscape && t.asInput() == "\\end") {
+ string const end = p.getArg('{', '}');
+ if (end == begin)
+ return;
+ }
+ }
+}
+
+
+/*!
+ * Read a list of TeX commands from a reLyX compatible syntax file.
+ * Since this list is used after all commands that have a LyX counterpart
+ * are handled, it does not matter that the "syntax.default" file
+ * has almost all of them listed. For the same reason the reLyX-specific
+ * reLyXre environment is ignored.
+ */
+void read_syntaxfile(string const & file_name)
+{
+ ifstream is(file_name.c_str());
+ if (!is.good()) {
+ cerr << "Could not open syntax file \"" << file_name
+ << "\" for reading." << endl;
+ exit(2);
+ }
+ // We can use our TeX parser, since the syntax of the layout file is
+ // modeled after TeX.
+ // Unknown tokens are just silently ignored, this helps us to skip some
+ // reLyX specific things.
+ Parser p(is);
+ while (p.good()) {
+ Token const & t = p.get_token();
+ if (t.cat() == catEscape) {
+ string const command = t.asInput();
+ if (command == "\\begin") {
+ string const name = p.getArg('{', '}');
+ if (name == "environments" || name == "reLyXre")
+ // We understand "reLyXre", but it is
+ // not as powerful as "environments".
+ read_environment(p, name,
+ known_environments);
+ else if (name == "mathenvironments")
+ read_environment(p, name,
+ known_math_environments);
+ } else {
+ read_command(p, command, known_commands);
}
- os << line << "\n";
}
}
}
-int main(int argc, char * argv[])
+string documentclass;
+string syntaxfile;
+bool overwrite_files = false;
+
+
+/// return the number of arguments consumed
+typedef boost::function<int(string const &, string const &)> cmd_helper;
+
+
+int parse_help(string const &, string const &)
{
- if (argc <= 1) {
- cerr << "Usage: " << argv[0] << " <infile.tex>" << endl;
- return 2;
+ cerr << "Usage: tex2lyx [ command line switches ] <infile.tex> [<outfile.lyx>]\n"
+ "Command line switches (case sensitive):\n"
+ "\t-help summarize tex2lyx usage\n"
+ "\t-f Force creation of .lyx files even if they exist already\n"
+ "\t-userdir dir try to set user directory to dir\n"
+ "\t-sysdir dir try to set system directory to dir\n"
+ "\t-c textclass declare the textclass\n"
+ "\t-s syntaxfile read additional syntax file" << endl;
+ exit(0);
+}
+
+
+int parse_class(string const & arg, string const &)
+{
+ if (arg.empty()) {
+ cerr << "Missing textclass string after -c switch" << endl;
+ exit(1);
+ }
+ documentclass = arg;
+ return 1;
+}
+
+
+int parse_syntaxfile(string const & arg, string const &)
+{
+ if (arg.empty()) {
+ cerr << "Missing syntaxfile string after -s switch" << endl;
+ exit(1);
+ }
+ syntaxfile = arg;
+ return 1;
+}
+
+
+// Filled with the command line arguments "foo" of "-sysdir foo" or
+// "-userdir foo".
+string cl_system_support;
+string cl_user_support;
+
+
+int parse_sysdir(string const & arg, string const &)
+{
+ if (arg.empty()) {
+ cerr << "Missing directory for -sysdir switch" << endl;
+ exit(1);
+ }
+ cl_system_support = arg;
+ return 1;
+}
+
+
+int parse_userdir(string const & arg, string const &)
+{
+ if (arg.empty()) {
+ cerr << "Missing directory for -userdir switch" << endl;
+ exit(1);
+ }
+ cl_user_support = arg;
+ return 1;
+}
+
+
+int parse_force(string const &, string const &)
+{
+ overwrite_files = true;
+ return 0;
+}
+
+
+void easyParse(int & argc, char * argv[])
+{
+ map<string, cmd_helper> cmdmap;
+
+ cmdmap["-c"] = parse_class;
+ cmdmap["-f"] = parse_force;
+ cmdmap["-s"] = parse_syntaxfile;
+ cmdmap["-help"] = parse_help;
+ cmdmap["--help"] = parse_help;
+ cmdmap["-sysdir"] = parse_sysdir;
+ cmdmap["-userdir"] = parse_userdir;
+
+ for (int i = 1; i < argc; ++i) {
+ std::map<string, cmd_helper>::const_iterator it
+ = cmdmap.find(argv[i]);
+
+ // don't complain if not found - may be parsed later
+ if (it == cmdmap.end())
+ continue;
+
+ string arg((i + 1 < argc) ? argv[i + 1] : "");
+ string arg2((i + 2 < argc) ? argv[i + 2] : "");
+
+ int const remove = 1 + it->second(arg, arg2);
+
+ // Now, remove used arguments by shifting
+ // the following ones remove places down.
+ argc -= remove;
+ for (int j = i; j < argc; ++j)
+ argv[j] = argv[j + remove];
+ --i;
}
+}
+
+
+// path of the first parsed file
+string masterFilePath;
+// path of the currently parsed file
+string parentFilePath;
+
+} // anonymous namespace
+
+
+string getMasterFilePath()
+{
+ return masterFilePath;
+}
+
+string getParentFilePath()
+{
+ return parentFilePath;
+}
+
- lyx::support::os::init(&argc, &argv);
- lyx::support::setLyxPaths();
+namespace {
- ifstream is(argv[1]);
+/*!
+ * Reads tex input from \a is and writes lyx output to \a os.
+ * Uses some common settings for the preamble, so this should only
+ * be used more than once for included documents.
+ * Caution: Overwrites the existing preamble settings if the new document
+ * contains a preamble.
+ * You must ensure that \p parentFilePath is properly set before calling
+ * this function!
+ */
+void tex2lyx(std::istream &is, std::ostream &os)
+{
Parser p(is);
//p.dump();
stringstream ss;
- LyXTextClass textclass = parse_preamble(p, ss);
- active_environments.push_back("document");
- parse_text(p, ss, FLAG_END, true, textclass);
- ss << "\n\\the_end\n";
+ LyXTextClass textclass = parse_preamble(p, ss, documentclass);
+ active_environments.push_back("document");
+ Context context(true, textclass);
+ parse_text(p, ss, FLAG_END, true, context);
+ if (Context::empty)
+ // Empty document body. LyX needs at least one paragraph.
+ context.check_layout(ss);
+ context.check_end_layout(ss);
+ ss << "\n\\end_body\n\\end_document\n";
+ active_environments.pop_back();
ss.seekg(0);
- clean_layouts(ss, cout);
+ os << ss.str();
+#ifdef TEST_PARSER
+ p.reset();
+ ofstream parsertest("parsertest.tex");
+ while (p.good())
+ parsertest << p.get_token().asInput();
+ // <origfile> and parsertest.tex should now have identical content
+#endif
+}
- return 0;
+
+/// convert TeX from \p infilename to LyX and write it to \p os
+bool tex2lyx(string const &infilename, std::ostream &os)
+{
+ BOOST_ASSERT(lyx::support::absolutePath(infilename));
+ ifstream is(infilename.c_str());
+ if (!is.good()) {
+ cerr << "Could not open input file \"" << infilename
+ << "\" for reading." << endl;
+ return false;
+ }
+ string const oldParentFilePath = parentFilePath;
+ parentFilePath = onlyPath(infilename);
+ tex2lyx(is, os);
+ parentFilePath = oldParentFilePath;
+ return true;
+}
+
+} // anonymous namespace
+
+
+bool tex2lyx(string const &infilename, string const &outfilename)
+{
+ if (isFileReadable(outfilename)) {
+ if (overwrite_files) {
+ cerr << "Overwriting existing file "
+ << outfilename << endl;
+ } else {
+ cerr << "Not overwriting existing file "
+ << outfilename << endl;
+ return false;
+ }
+ } else {
+ cerr << "Creating file " << outfilename << endl;
+ }
+ ofstream os(outfilename.c_str());
+ if (!os.good()) {
+ cerr << "Could not open output file \"" << outfilename
+ << "\" for writing." << endl;
+ return false;
+ }
+#ifdef FILEDEBUG
+ cerr << "Input file: " << infilename << "\n";
+ cerr << "Output file: " << outfilename << "\n";
+#endif
+ return tex2lyx(infilename, os);
+}
+
+
+int main(int argc, char * argv[])
+{
+ fs::path::default_name_check(fs::no_check);
+
+ easyParse(argc, argv);
+
+ if (argc <= 1) {
+ cerr << "Usage: tex2lyx [ command line switches ] <infile.tex> [<outfile.lyx>]\n"
+ "See tex2lyx -help." << endl;
+ return 2;
+ }
+
+ lyx::support::os::init(argc, argv);
+ lyx::support::init_package(argv[0], cl_system_support, cl_user_support,
+ lyx::support::top_build_dir_is_two_levels_up);
+
+ // Now every known option is parsed. Look for input and output
+ // file name (the latter is optional).
+ string const infilename = makeAbsPath(argv[1]);
+ string outfilename;
+ if (argc > 2) {
+ outfilename = argv[2];
+ if (outfilename != "-")
+ outfilename = makeAbsPath(argv[2]);
+ } else
+ outfilename = changeExtension(infilename, ".lyx");
+
+ string const system_syntaxfile = lyx::support::libFileSearch("", "syntax.default");
+ if (system_syntaxfile.empty()) {
+ cerr << "Error: Could not find syntax file \"syntax.default\"." << endl;
+ exit(1);
+ }
+ read_syntaxfile(system_syntaxfile);
+ if (!syntaxfile.empty())
+ read_syntaxfile(syntaxfile);
+
+ masterFilePath = onlyPath(infilename);
+ parentFilePath = masterFilePath;
+
+ if (outfilename == "-") {
+ if (tex2lyx(infilename, cout))
+ return EXIT_SUCCESS;
+ else
+ return EXIT_FAILURE;
+ } else {
+ if (tex2lyx(infilename, outfilename))
+ return EXIT_SUCCESS;
+ else
+ return EXIT_FAILURE;
+ }
}
// }])
projects / lyx.git / blobdiff