]> git.lyx.org Git - lyx.git/blobdiff - src/tex2lyx/tex2lyx.cpp
Natbib authoryear uses (Ref1; Ref2) by default.
[lyx.git] / src / tex2lyx / tex2lyx.cpp
index 5a2d299844d336bff8c25fe3379a6c4fc02f5a7e..9239d30984be78fba2a1f7626bdc5848790e05cc 100644 (file)
@@ -36,6 +36,7 @@
 #include "support/Systemcall.h"
 
 #include <cstdlib>
+#include <algorithm>
 #include <iostream>
 #include <string>
 #include <sstream>
@@ -150,10 +151,12 @@ CommandMap known_environments;
 CommandMap known_math_environments;
 FullCommandMap possible_textclass_commands;
 FullEnvironmentMap possible_textclass_environments;
+FullCommandMap possible_textclass_theorems;
 int const LYX_FORMAT = LYX_FORMAT_TEX2LYX;
 
 /// used modules
 LayoutModuleList used_modules;
+vector<string> preloaded_modules;
 
 
 void convertArgs(string const & o1, bool o2, vector<ArgumentType> & arguments)
@@ -205,6 +208,17 @@ void add_known_environment(string const & environment, string const & o1,
 }
 
 
+void add_known_theorem(string const & theorem, string const & o1,
+                       bool o2, docstring const & definition)
+{
+       vector<ArgumentType> arguments;
+       convertArgs(o1, o2, arguments);
+       if (!definition.empty())
+               possible_textclass_theorems[theorem] =
+                       FullCommand(arguments, definition);
+}
+
+
 Layout const * findLayoutWithoutModule(TextClass const & textclass,
                                        string const & name, bool command)
 {
@@ -234,45 +248,142 @@ InsetLayout const * findInsetLayoutWithoutModule(TextClass const & textclass,
 }
 
 
-bool checkModule(string const & name, bool command)
-{
-       // Cache to avoid slowdown by repated searches
-       static set<string> failed[2];
+namespace {
 
-       // Only add the module if the command was actually defined in the LyX preamble
-       if (command) {
-               if (possible_textclass_commands.find('\\' + name) == possible_textclass_commands.end())
-                       return false;
-       } else {
-               if (possible_textclass_environments.find(name) == possible_textclass_environments.end())
-                       return false;
+typedef map<string, DocumentClassPtr> ModuleMap;
+ModuleMap modules;
+
+
+bool addModule(string const module, LayoutFile const & baseClass, LayoutModuleList & m, vector<string> & visited)
+{
+       // avoid endless loop for circular dependency
+       vector<string>::const_iterator const vb = visited.begin();
+       vector<string>::const_iterator const ve = visited.end();
+       if (find(vb, ve, module) != ve) {
+               cerr << "Circular dependency detected for module " << module << '\n';
+               return false;
        }
-       if (failed[command].find(name) != failed[command].end())
+       LyXModule const * const lm = theModuleList[module];
+       if (!lm) {
+               cerr << "Could not find module " << module << " in module list.\n";
                return false;
+       }
+       bool foundone = false;
+       LayoutModuleList::const_iterator const exclmodstart = baseClass.excludedModules().begin();
+       LayoutModuleList::const_iterator const exclmodend = baseClass.excludedModules().end();
+       LayoutModuleList::const_iterator const provmodstart = baseClass.providedModules().begin();
+       LayoutModuleList::const_iterator const provmodend = baseClass.providedModules().end();
+       vector<string> const reqs = lm->getRequiredModules();
+       if (reqs.empty())
+               foundone = true;
+       else {
+               LayoutModuleList::const_iterator mit = m.begin();
+               LayoutModuleList::const_iterator men = m.end();
+               vector<string>::const_iterator rit = reqs.begin();
+               vector<string>::const_iterator ren = reqs.end();
+               for (; rit != ren; ++rit) {
+                       if (find(mit, men, *rit) != men) {
+                               foundone = true;
+                               break;
+                       }
+                       if (find(provmodstart, provmodend, *rit) != provmodend) {
+                               foundone = true;
+                               break;
+                       }
+               }
+               if (!foundone) {
+                       visited.push_back(module);
+                       for (rit = reqs.begin(); rit != ren; ++rit) {
+                               if (find(exclmodstart, exclmodend, *rit) == exclmodend) {
+                                       if (addModule(*rit, baseClass, m, visited)) {
+                                               foundone = true;
+                                               break;
+                                       }
+                               }
+                       }
+                       visited.pop_back();
+               }
+       }
+       if (!foundone) {
+               cerr << "Could not add required modules for " << module << ".\n";
+               return false;
+       }
+       if (!m.moduleCanBeAdded(module, &baseClass))
+               return false;
+       m.push_back(module);
+       return true;
+}
 
+
+void initModules()
+{
        // Create list of dummy document classes if not already done.
        // This is needed since a module cannot be read on its own, only as
        // part of a document class.
        LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
-       typedef map<string, DocumentClass *> ModuleMap;
-       static ModuleMap modules;
        static bool init = true;
        if (init) {
                baseClass.load();
-               DocumentClassBundle & bundle = DocumentClassBundle::get();
                LyXModuleList::const_iterator const end = theModuleList.end();
                LyXModuleList::const_iterator it = theModuleList.begin();
-               for (; it != end; it++) {
+               for (; it != end; ++it) {
                        string const module = it->getID();
                        LayoutModuleList m;
-                       // FIXME this excludes all modules that depend on another one
-                       if (!m.moduleCanBeAdded(module, &baseClass))
+                       vector<string> v;
+                       if (!addModule(module, baseClass, m, v))
                                continue;
-                       m.push_back(module);
-                       modules[module] = &bundle.makeDocumentClass(baseClass, m);
+                       modules[module] = getDocumentClass(baseClass, m);
                }
                init = false;
        }
+}
+
+
+bool addModule(string const & module)
+{
+       initModules();
+       LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
+       if (!used_modules.moduleCanBeAdded(module, &baseClass))
+               return false;
+       FileName layout_file = libFileSearch("layouts", module, "module");
+       if (textclass.read(layout_file, TextClass::MODULE)) {
+               used_modules.push_back(module);
+               // speed up further searches:
+               // the module does not need to be checked anymore.
+               ModuleMap::iterator const it = modules.find(module);
+               if (it != modules.end())
+                       modules.erase(it);
+               return true;
+       }
+       return false;
+}
+
+}
+
+
+bool checkModule(string const & name, bool command)
+{
+       // Cache to avoid slowdown by repated searches
+       static set<string> failed[2];
+
+       // Only add the module if the command was actually defined in the LyX preamble
+       bool theorem = false;
+       if (command) {
+               if (possible_textclass_commands.find('\\' + name) == possible_textclass_commands.end())
+                       return false;
+       } else {
+               if (possible_textclass_environments.find(name) == possible_textclass_environments.end()) {
+                       if (possible_textclass_theorems.find(name) != possible_textclass_theorems.end())
+                               theorem = true;
+                       else
+                               return false;
+               }
+       }
+       if (failed[command].find(name) != failed[command].end())
+               return false;
+
+       initModules();
+       LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
 
        // Try to find a module that defines the command.
        // Only add it if the definition can be found in the preamble of the
@@ -282,15 +393,15 @@ bool checkModule(string const & name, bool command)
        // needed since it is not unlikely that two different modules define a
        // command with the same name.
        ModuleMap::iterator const end = modules.end();
-       for (ModuleMap::iterator it = modules.begin(); it != end; it++) {
+       for (ModuleMap::iterator it = modules.begin(); it != end; ++it) {
                string const module = it->first;
-               if (!used_modules.moduleCanBeAdded(module, &baseClass))
+               if (used_modules.moduleConflicts(module, &baseClass))
                        continue;
                if (findLayoutWithoutModule(textclass, name, command))
                        continue;
                if (findInsetLayoutWithoutModule(textclass, name, command))
                        continue;
-               DocumentClass const * c = it->second;
+               DocumentClassConstPtr c = it->second;
                Layout const * layout = findLayoutWithoutModule(*c, name, command);
                InsetLayout const * insetlayout = layout ? 0 :
                        findInsetLayoutWithoutModule(*c, name, command);
@@ -307,6 +418,11 @@ bool checkModule(string const & name, bool command)
                                possible_textclass_commands['\\' + name];
                        if (preamble.find(cmd.def) != docstring::npos)
                                add = true;
+               } else if (theorem) {
+                       FullCommand const & thm =
+                               possible_textclass_theorems[name];
+                       if (preamble.find(thm.def) != docstring::npos)
+                               add = true;
                } else {
                        FullEnvironment const & env =
                                possible_textclass_environments[name];
@@ -330,8 +446,17 @@ bool checkModule(string const & name, bool command)
 }
 
 
+bool isProvided(string const & name)
+{
+       // This works only for features that are named like the LaTeX packages
+       return textclass.provides(name) || preamble.isPackageUsed(name);
+}
+
+
 bool noweb_mode = false;
 bool pdflatex = false;
+bool xetex = false;
+bool is_nonCJKJapanese = false;
 bool roundtrip = false;
 
 
@@ -354,13 +479,20 @@ void read_command(Parser & p, string command, CommandMap & commands)
                        string const arg = p.getArg('{', '}');
                        if (arg == "translate")
                                arguments.push_back(required);
+                       else if (arg == "group")
+                               arguments.push_back(req_group);
                        else if (arg == "item")
                                arguments.push_back(item);
+                       else if (arg == "displaymath")
+                               arguments.push_back(displaymath);
                        else
                                arguments.push_back(verbatim);
                } else {
-                       p.getArg('[', ']');
-                       arguments.push_back(optional);
+                       string const arg = p.getArg('[', ']');
+                       if (arg == "group")
+                               arguments.push_back(opt_group);
+                       else
+                               arguments.push_back(optional);
                }
        }
        commands[command] = arguments;
@@ -437,7 +569,9 @@ void read_syntaxfile(FileName const & file_name)
 string documentclass;
 string default_encoding;
 string syntaxfile;
+bool copy_files = false;
 bool overwrite_files = false;
+bool skip_children = false;
 int error_code = 0;
 
 /// return the number of arguments consumed
@@ -449,10 +583,13 @@ int parse_help(string const &, string const &)
        cerr << "Usage: tex2lyx [options] infile.tex [outfile.lyx]\n"
                "Options:\n"
                "\t-c textclass       Declare the textclass.\n"
+               "\t-m mod1[,mod2...]  Load the given modules.\n"
+               "\t-copyfiles         Copy all included files to the directory of outfile.lyx.\n"
                "\t-e encoding        Set the default encoding (latex name).\n"
                "\t-f                 Force overwrite of .lyx files.\n"
                "\t-help              Print this message and quit.\n"
-               "\t-n                 translate a noweb (aka literate programming) file.\n"
+               "\t-n                 translate literate programming (noweb, sweave,... ) file.\n"
+               "\t-skipchildren      Do not translate included child documents.\n"
                "\t-roundtrip         re-export created .lyx file infile.lyx.lyx to infile.lyx.tex.\n"
                "\t-s syntaxfile      read additional syntax file.\n"
                "\t-sysdir SYSDIR     Set system directory to SYSDIR.\n"
@@ -499,6 +636,15 @@ int parse_class(string const & arg, string const &)
 }
 
 
+int parse_module(string const & arg, string const &)
+{
+       if (arg.empty())
+               error_message("Missing modules string after -m switch");
+       split(arg, preloaded_modules, ',');
+       return 1;
+}
+
+
 int parse_encoding(string const & arg, string const &)
 {
        if (arg.empty())
@@ -555,6 +701,13 @@ int parse_noweb(string const &, string const &)
 }
 
 
+int parse_skipchildren(string const &, string const &)
+{
+       skip_children = true;
+       return 0;
+}
+
+
 int parse_roundtrip(string const &, string const &)
 {
        roundtrip = true;
@@ -562,6 +715,13 @@ int parse_roundtrip(string const &, string const &)
 }
 
 
+int parse_copyfiles(string const &, string const &)
+{
+       copy_files = true;
+       return 0;
+}
+
+
 void easyParse(int & argc, char * argv[])
 {
        map<string, cmd_helper> cmdmap;
@@ -573,13 +733,16 @@ void easyParse(int & argc, char * argv[])
        cmdmap["-version"] = parse_version;
        cmdmap["--version"] = parse_version;
        cmdmap["-c"] = parse_class;
+       cmdmap["-m"] = parse_module;
        cmdmap["-e"] = parse_encoding;
        cmdmap["-f"] = parse_force;
        cmdmap["-s"] = parse_syntaxfile;
        cmdmap["-n"] = parse_noweb;
+       cmdmap["-skipchildren"] = parse_skipchildren;
        cmdmap["-sysdir"] = parse_sysdir;
        cmdmap["-userdir"] = parse_userdir;
        cmdmap["-roundtrip"] = parse_roundtrip;
+       cmdmap["-copyfiles"] = parse_copyfiles;
 
        for (int i = 1; i < argc; ++i) {
                map<string, cmd_helper>::const_iterator it
@@ -610,21 +773,48 @@ void easyParse(int & argc, char * argv[])
 
 
 // path of the first parsed file
-string masterFilePath;
+string masterFilePathLyX;
+string masterFilePathTeX;
 // path of the currently parsed file
-string parentFilePath;
+string parentFilePathTeX;
 
 } // anonymous namespace
 
 
-string getMasterFilePath()
+string getMasterFilePath(bool input)
 {
-       return masterFilePath;
+       return input ? masterFilePathTeX : masterFilePathLyX;
 }
 
-string getParentFilePath()
+string getParentFilePath(bool input)
 {
-       return parentFilePath;
+       if (input)
+               return parentFilePathTeX;
+       string const rel = to_utf8(makeRelPath(from_utf8(masterFilePathTeX),
+                                              from_utf8(parentFilePathTeX)));
+       if (rel.substr(0, 3) == "../") {
+               // The parent is not below the master - keep the path
+               return parentFilePathTeX;
+       }
+       return makeAbsPath(rel, masterFilePathLyX).absFileName();
+}
+
+
+bool copyFiles()
+{
+       return copy_files;
+}
+
+
+bool overwriteFiles()
+{
+       return overwrite_files;
+}
+
+
+bool skipChildren()
+{
+       return skip_children;
 }
 
 
@@ -636,21 +826,24 @@ namespace {
  *  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
+ *  You must ensure that \p parentFilePathTeX is properly set before calling
  *  this function!
  */
-void tex2lyx(idocstream & is, ostream & os, string encoding)
+bool tex2lyx(idocstream & is, ostream & os, string encoding)
 {
        // Set a sensible default encoding.
        // This is used until an encoding command is found.
-       // For child documents use the encoding of the master, else latin1,
-       // since latin1 does not cause an iconv error if the actual encoding
-       // is different (bug 7509).
+       // For child documents use the encoding of the master, else ISO8859-1,
+       // (formerly known by its latex name latin1), since ISO8859-1 does not
+       // cause an iconv error if the actual encoding is different (bug 7509).
        if (encoding.empty()) {
                if (preamble.inputencoding() == "auto")
-                       encoding = "latin1";
-               else
-                       encoding = preamble.inputencoding();
+                       encoding = "ISO8859-1";
+               else {
+                       Encoding const * const enc = encodings.fromLyXName(
+                               preamble.inputencoding(), true);
+                       encoding = enc->iconvName();
+               }
        }
 
        Parser p(is);
@@ -658,10 +851,35 @@ void tex2lyx(idocstream & is, ostream & os, string encoding)
        //p.dump();
 
        preamble.parse(p, documentclass, textclass);
+       list<string> removed_modules;
+       LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
+       if (!used_modules.adaptToBaseClass(&baseClass, removed_modules)) {
+               cerr << "Could not load default modules for text class." << endl;
+               return false;
+       }
+
+       // Load preloaded modules.
+       // This needs to be done after the preamble is parsed, since the text
+       // class may not be known before. It neds to be done before parsing
+       // body, since otherwise the commands/environments provided by the
+       // modules would be parsed as ERT.
+       for (size_t i = 0; i < preloaded_modules.size(); ++i) {
+               if (!addModule(preloaded_modules[i])) {
+                       cerr << "Error: Could not load module \""
+                            << preloaded_modules[i] << "\"." << endl;
+                       return false;
+               }
+       }
+       // Ensure that the modules are not loaded again for included files
+       preloaded_modules.clear();
 
        active_environments.push_back("document");
        Context context(true, textclass);
        stringstream ss;
+       // store the document language in the context to be able to handle the
+       // commands like \foreignlanguage and \textenglish etc.
+       context.font.language = preamble.defaultLanguage();
+       // parse the main text
        parse_text(p, ss, FLAG_END, true, context);
        if (Context::empty)
                // Empty document body. LyX needs at least one paragraph.
@@ -674,10 +892,13 @@ void tex2lyx(idocstream & is, ostream & os, string encoding)
        if (!used_modules.empty()) {
                LayoutModuleList::const_iterator const end = used_modules.end();
                LayoutModuleList::const_iterator it = used_modules.begin();
-               for (; it != end; it++)
+               for (; it != end; ++it)
                        preamble.addModule(*it);
        }
-       preamble.writeLyXHeader(os);
+       if (!preamble.writeLyXHeader(os, !active_environments.empty())) {
+               cerr << "Could not write LyX file header." << endl;
+               return false;
+       }
 
        ss.seekg(0);
        os << ss.str();
@@ -688,6 +909,7 @@ void tex2lyx(idocstream & is, ostream & os, string encoding)
                parsertest << p.get_token().asInput();
        // <origfile> and parsertest.tex should now have identical content
 #endif
+       return true;
 }
 
 
@@ -703,17 +925,17 @@ bool tex2lyx(FileName const & infilename, ostream & os, string const & encoding)
                     << "\" for reading." << endl;
                return false;
        }
-       string const oldParentFilePath = parentFilePath;
-       parentFilePath = onlyPath(infilename.absFileName());
-       tex2lyx(is, os, encoding);
-       parentFilePath = oldParentFilePath;
-       return true;
+       string const oldParentFilePath = parentFilePathTeX;
+       parentFilePathTeX = onlyPath(infilename.absFileName());
+       bool retval = tex2lyx(is, os, encoding);
+       parentFilePathTeX = oldParentFilePath;
+       return retval;
 }
 
 } // anonymous namespace
 
 
-bool tex2lyx(string const & infilename, FileName const & outfilename, 
+bool tex2lyx(string const & infilename, FileName const & outfilename,
             string const & encoding)
 {
        if (outfilename.isReadableFile()) {
@@ -754,6 +976,8 @@ bool tex2tex(string const & infilename, FileName const & outfilename,
                command += " -f none";
        if (pdflatex)
                command += " -e pdflatex ";
+       else if (xetex)
+               command += " -e xetex ";
        else
                command += " -e latex ";
        command += quoteName(outfilename.toFilesystemEncoding());
@@ -788,7 +1012,7 @@ int main(int argc, char * argv[])
 
        easyParse(argc, argv);
 
-       if (argc <= 1) 
+       if (argc <= 1)
                error_message("Not enough arguments.");
 
        try {
@@ -807,20 +1031,28 @@ int main(int argc, char * argv[])
        infilename = makeAbsPath(infilename).absFileName();
 
        string outfilename;
-       if (roundtrip) {
-               if (argc > 2) {
-                       // Do not allow a user supplied output filename
-                       // (otherwise it could easily happen that LyX would
-                       // overwrite the original .tex file)
-                       cerr << "Error: output filename must not be given in roundtrip mode."
-                            << endl;
-                       return EXIT_FAILURE;
-               }
-               outfilename = changeExtension(infilename, ".lyx.lyx");
-       } else if (argc > 2) {
+       if (argc > 2) {
                outfilename = internal_path(os::utf8_argv(2));
                if (outfilename != "-")
                        outfilename = makeAbsPath(outfilename).absFileName();
+               if (roundtrip) {
+                       if (outfilename == "-") {
+                               cerr << "Error: Writing to standard output is "
+                                       "not supported in roundtrip mode."
+                                    << endl;
+                               return EXIT_FAILURE;
+                       }
+                       string texfilename = changeExtension(outfilename, ".tex");
+                       if (equivalent(FileName(infilename), FileName(texfilename))) {
+                               cerr << "Error: The input file `" << infilename
+                                    << "´ would be overwritten by the TeX file exported from `"
+                                    << outfilename << "´ in roundtrip mode." << endl;
+                               return EXIT_FAILURE;
+                       }
+               }
+       } else if (roundtrip) {
+               // avoid overwriting the input file
+               outfilename = changeExtension(infilename, ".lyx.lyx");
        } else
                outfilename = changeExtension(infilename, ".lyx");
 
@@ -837,19 +1069,24 @@ int main(int argc, char * argv[])
        // Read the encodings table.
        FileName const symbols_path = libFileSearch(string(), "unicodesymbols");
        if (symbols_path.empty()) {
-               cerr << "Error: Could not find file \"unicodesymbols\"." 
+               cerr << "Error: Could not find file \"unicodesymbols\"."
                     << endl;
                return EXIT_FAILURE;
        }
        FileName const enc_path = libFileSearch(string(), "encodings");
        if (enc_path.empty()) {
-               cerr << "Error: Could not find file \"encodings\"." 
+               cerr << "Error: Could not find file \"encodings\"."
                     << endl;
                return EXIT_FAILURE;
        }
        encodings.read(enc_path, symbols_path);
-       if (!default_encoding.empty() && !encodings.fromLaTeXName(default_encoding))
-               error_message("Unknown LaTeX encoding `" + default_encoding + "'");
+       if (!default_encoding.empty()) {
+               Encoding const * const enc = encodings.fromLaTeXName(
+                       default_encoding, Encoding::any, true);
+               if (!enc)
+                       error_message("Unknown LaTeX encoding `" + default_encoding + "'");
+               default_encoding = enc->iconvName();
+       }
 
        // Load the layouts
        LayoutFileList::get().read();
@@ -857,17 +1094,32 @@ int main(int argc, char * argv[])
        theModuleList.read();
 
        // The real work now.
-       masterFilePath = onlyPath(infilename);
-       parentFilePath = masterFilePath;
+       masterFilePathTeX = onlyPath(infilename);
+       parentFilePathTeX = masterFilePathTeX;
        if (outfilename == "-") {
+               // assume same directory as input file
+               masterFilePathLyX = masterFilePathTeX;
                if (tex2lyx(FileName(infilename), cout, default_encoding))
                        return EXIT_SUCCESS;
-       } else if (roundtrip) {
-               if (tex2tex(infilename, FileName(outfilename), default_encoding))
-                       return EXIT_SUCCESS;
        } else {
-               if (tex2lyx(infilename, FileName(outfilename), default_encoding))
-                       return EXIT_SUCCESS;
+               masterFilePathLyX = onlyPath(outfilename);
+               if (copy_files) {
+                       FileName const path(masterFilePathLyX);
+                       if (!path.isDirectory()) {
+                               if (!path.createPath()) {
+                                       cerr << "Warning: Could not create directory for file `"
+                                            << masterFilePathLyX << "´." << endl;
+                                       return EXIT_FAILURE;
+                               }
+                       }
+               }
+               if (roundtrip) {
+                       if (tex2tex(infilename, FileName(outfilename), default_encoding))
+                               return EXIT_SUCCESS;
+               } else {
+                       if (tex2lyx(infilename, FileName(outfilename), default_encoding))
+                               return EXIT_SUCCESS;
+               }
        }
        return EXIT_FAILURE;
 }