* This file is part of LyX, the document processor.
* Licence details can be found in the file COPYING.
*
- * \author André Pönitz
+ * \author André Pönitz
*
* Full author contact details are available in file CREDITS.
*/
#include <config.h>
#include "tex2lyx.h"
-#include "Context.h"
-#include "TextClass.h"
+#include "Context.h"
+#include "Encoding.h"
#include "Layout.h"
+#include "LayoutFile.h"
+#include "LayoutModuleList.h"
+#include "ModuleList.h"
+#include "TextClass.h"
#include "support/convert.h"
#include "support/debug.h"
#include "support/ExceptionMessage.h"
#include "support/filetools.h"
+#include "support/lassert.h"
#include "support/lstrings.h"
+#include "support/Messages.h"
#include "support/os.h"
#include "support/Package.h"
+#include "support/Systemcall.h"
-#include <fstream>
+#include <cstdlib>
#include <iostream>
#include <string>
#include <sstream>
namespace lyx {
-// Hacks to allow the thing to link in the lyxlayout stuff
-LayoutPtr captionlayout;
+namespace frontend {
+namespace Alert {
+ void warning(docstring const & title, docstring const & message,
+ bool const &)
+ {
+ LYXERR0(title);
+ LYXERR0(message);
+ }
+}
+}
+
+
+// Dummy texrow support
+void TexRow::newline()
+{}
+
+
+void TexRow::newlines(int)
+{}
+
+
+// Dummy LyXRC support
+class LyXRC {
+ string icon_set;
+} lyxrc;
+
+
+// Dummy translation support
+Messages messages_;
+Messages const & getMessages(std::string const &)
+{
+ return messages_;
+}
+
+
+Messages const & getGuiMessages()
+{
+ return messages_;
+}
+
+
+// Keep the linker happy on Windows
+void lyx_exit(int)
+{}
string const trim(string const & a, char const * p)
{
- // BOOST_ASSERT(p);
+ // LASSERT(p, /**/);
if (a.empty() || !*p)
return a;
}
+TeX2LyXDocClass textclass;
CommandMap known_commands;
CommandMap known_environments;
CommandMap known_math_environments;
+FullCommandMap possible_textclass_commands;
+FullEnvironmentMap possible_textclass_environments;
+/// used modules
+LayoutModuleList used_modules;
-void add_known_command(string const & command, string const & o1,
- unsigned optionalsNum)
+
+void convertArgs(string const & o1, bool o2, vector<ArgumentType> & arguments)
{
// We have to handle the following cases:
// definition o1 o2 invocation result
// \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
- // and the same with \newlyxcommand
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);
- for (unsigned int i = 0; i < optionalsNum; ++i) {
+ 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);
+}
+
+
+void add_known_command(string const & command, string const & o1,
+ bool o2, docstring const & definition)
+{
+ vector<ArgumentType> arguments;
+ convertArgs(o1, o2, arguments);
known_commands[command] = arguments;
+ if (!definition.empty())
+ possible_textclass_commands[command] =
+ FullCommand(arguments, definition);
+}
+
+
+void add_known_environment(string const & environment, string const & o1,
+ bool o2, docstring const & beg, docstring const &end)
+{
+ vector<ArgumentType> arguments;
+ convertArgs(o1, o2, arguments);
+ known_environments[environment] = arguments;
+ if (!beg.empty() || ! end.empty())
+ possible_textclass_environments[environment] =
+ FullEnvironment(arguments, beg, end);
+}
+
+
+Layout const * findLayoutWithoutModule(TextClass const & textclass,
+ string const & name, bool command)
+{
+ DocumentClass::const_iterator it = textclass.begin();
+ DocumentClass::const_iterator en = textclass.end();
+ for (; it != en; ++it) {
+ if (it->latexname() == name &&
+ ((command && it->isCommand()) || (!command && it->isEnvironment())))
+ return &*it;
+ }
+ return 0;
+}
+
+
+InsetLayout const * findInsetLayoutWithoutModule(TextClass const & textclass,
+ string const & name, bool command)
+{
+ DocumentClass::InsetLayouts::const_iterator it = textclass.insetLayouts().begin();
+ DocumentClass::InsetLayouts::const_iterator en = textclass.insetLayouts().end();
+ for (; it != en; ++it) {
+ if (it->second.latexname() == name &&
+ ((command && it->second.latextype() == InsetLayout::COMMAND) ||
+ (!command && it->second.latextype() == InsetLayout::ENVIRONMENT)))
+ return &(it->second);
+ }
+ return 0;
+}
+
+
+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
+ 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;
+ }
+ if (failed[command].find(name) != failed[command].end())
+ return false;
+
+ // 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++) {
+ string const module = it->getID();
+ LayoutModuleList m;
+ // FIXME this excludes all modules that depend on another one
+ if (!m.moduleCanBeAdded(module, &baseClass))
+ continue;
+ m.push_back(module);
+ modules[module] = &bundle.makeDocumentClass(baseClass, m);
+ }
+ init = false;
+ }
+
+ // Try to find a module that defines the command.
+ // Only add it if the definition can be found in the preamble of the
+ // style that corresponds to the command. This is a heuristic and
+ // different from the way how we parse the builtin commands of the
+ // text class (in that case we only compare the name), but it is
+ // 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++) {
+ string const module = it->first;
+ if (!used_modules.moduleCanBeAdded(module, &baseClass))
+ continue;
+ if (findLayoutWithoutModule(textclass, name, command))
+ continue;
+ if (findInsetLayoutWithoutModule(textclass, name, command))
+ continue;
+ DocumentClass const * c = it->second;
+ Layout const * layout = findLayoutWithoutModule(*c, name, command);
+ InsetLayout const * insetlayout = layout ? 0 :
+ findInsetLayoutWithoutModule(*c, name, command);
+ docstring preamble;
+ if (layout)
+ preamble = layout->preamble();
+ else if (insetlayout)
+ preamble = insetlayout->preamble();
+ if (preamble.empty())
+ continue;
+ bool add = false;
+ if (command) {
+ FullCommand const & cmd =
+ possible_textclass_commands['\\' + name];
+ if (preamble.find(cmd.def) != docstring::npos)
+ add = true;
+ } else {
+ FullEnvironment const & env =
+ possible_textclass_environments[name];
+ if (preamble.find(env.beg) != docstring::npos &&
+ preamble.find(env.end) != docstring::npos)
+ add = true;
+ }
+ if (add) {
+ 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.
+ modules.erase(it);
+ return true;
+ }
+ }
+ }
+ failed[command].insert(name);
+ return false;
}
bool noweb_mode = false;
+bool pdflatex = false;
+bool roundtrip = false;
namespace {
string const arg = p.getArg('{', '}');
if (arg == "translate")
arguments.push_back(required);
+ else if (arg == "item")
+ arguments.push_back(item);
else
arguments.push_back(verbatim);
} else {
*/
void read_syntaxfile(FileName const & file_name)
{
- ifstream is(file_name.toFilesystemEncoding().c_str());
+ ifdocstream is(file_name.toFilesystemEncoding().c_str());
if (!is.good()) {
cerr << "Could not open syntax file \"" << file_name
<< "\" for reading." << endl;
string documentclass;
+string default_encoding;
string syntaxfile;
bool overwrite_files = false;
-
+int error_code = 0;
/// return the number of arguments consumed
typedef int (*cmd_helper)(string const &, string const &);
int parse_help(string const &, string const &)
{
- 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"
+ cerr << "Usage: tex2lyx [options] infile.tex [outfile.lyx]\n"
+ "Options:\n"
+ "\t-c textclass Declare the textclass.\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-s syntaxfile read additional syntax file" << endl;
- exit(0);
+ "\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 dir Set system directory to DIR.\n"
+ "\t-userdir DIR Set user directory to DIR."
+ << endl;
+ exit(error_code);
+}
+
+
+void error_message(string const & message)
+{
+ cerr << "tex2lyx: " << message << "\n\n";
+ error_code = 1;
+ parse_help(string(), string());
}
int parse_class(string const & arg, string const &)
{
- if (arg.empty()) {
- cerr << "Missing textclass string after -c switch" << endl;
- exit(1);
- }
+ if (arg.empty())
+ error_message("Missing textclass string after -c switch");
documentclass = arg;
return 1;
}
+int parse_encoding(string const & arg, string const &)
+{
+ if (arg.empty())
+ error_message("Missing encoding string after -e switch");
+ default_encoding = arg;
+ return 1;
+}
+
+
int parse_syntaxfile(string const & arg, string const &)
{
- if (arg.empty()) {
- cerr << "Missing syntaxfile string after -s switch" << endl;
- exit(1);
- }
+ if (arg.empty())
+ error_message("Missing syntaxfile string after -s switch");
syntaxfile = internal_path(arg);
return 1;
}
int parse_sysdir(string const & arg, string const &)
{
- if (arg.empty()) {
- cerr << "Missing directory for -sysdir switch" << endl;
- exit(1);
- }
+ if (arg.empty())
+ error_message("Missing directory for -sysdir switch");
cl_system_support = internal_path(arg);
return 1;
}
int parse_userdir(string const & arg, string const &)
{
- if (arg.empty()) {
- cerr << "Missing directory for -userdir switch" << endl;
- exit(1);
- }
+ if (arg.empty())
+ error_message("Missing directory for -userdir switch");
cl_user_support = internal_path(arg);
return 1;
}
}
+int parse_roundtrip(string const &, string const &)
+{
+ roundtrip = true;
+ return 0;
+}
+
+
void easyParse(int & argc, char * argv[])
{
map<string, cmd_helper> cmdmap;
cmdmap["-c"] = parse_class;
+ cmdmap["-e"] = parse_encoding;
cmdmap["-f"] = parse_force;
cmdmap["-s"] = parse_syntaxfile;
cmdmap["-help"] = parse_help;
cmdmap["-n"] = parse_noweb;
cmdmap["-sysdir"] = parse_sysdir;
cmdmap["-userdir"] = parse_userdir;
+ cmdmap["-roundtrip"] = parse_roundtrip;
for (int i = 1; i < argc; ++i) {
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;
+ if (it == cmdmap.end()) {
+ if (argv[i][0] == '-')
+ error_message(string("Unknown option `") + argv[i] + "'.");
+ else
+ continue;
+ }
- string arg(to_utf8(from_local8bit((i + 1 < argc) ? argv[i + 1] : "")));
- string arg2(to_utf8(from_local8bit((i + 2 < argc) ? argv[i + 2] : "")));
+ string arg = (i + 1 < argc) ? os::utf8_argv(i + 1) : string();
+ string arg2 = (i + 2 < argc) ? os::utf8_argv(i + 2) : string();
int const remove = 1 + it->second(arg, arg2);
// Now, remove used arguments by shifting
// the following ones remove places down.
+ os::remove_internal_args(i, remove);
argc -= remove;
for (int j = i; j < argc; ++j)
argv[j] = argv[j + remove];
* You must ensure that \p parentFilePath is properly set before calling
* this function!
*/
-void tex2lyx(istream & is, ostream & os)
+void tex2lyx(idocstream & is, ostream & os, string const & encoding)
{
Parser p(is);
+ if (!encoding.empty())
+ p.setEncoding(encoding);
//p.dump();
- stringstream ss;
- TextClass textclass = parse_preamble(p, ss, documentclass);
- captionlayout = LayoutPtr(Layout::forCaption());
+ ostringstream ps;
+ parse_preamble(p, ps, documentclass, textclass);
active_environments.push_back("document");
Context context(true, textclass);
+ stringstream ss;
parse_text(p, ss, FLAG_END, true, context);
if (Context::empty)
// Empty document body. LyX needs at least one paragraph.
context.check_end_layout(ss);
ss << "\n\\end_body\n\\end_document\n";
active_environments.pop_back();
+
+ // We know the used modules only after parsing the full text
+ ostringstream ms;
+ if (!used_modules.empty()) {
+ ms << "\\begin_modules\n";
+ LayoutModuleList::const_iterator const end = used_modules.end();
+ LayoutModuleList::const_iterator it = used_modules.begin();
+ for (; it != end; it++)
+ ms << *it << '\n';
+ ms << "\\end_modules\n";
+ }
+ os << subst(ps.str(), modules_placeholder, ms.str());
+
ss.seekg(0);
os << ss.str();
#ifdef TEST_PARSER
p.reset();
- ofstream parsertest("parsertest.tex");
+ ofdocstream parsertest("parsertest.tex");
while (p.good())
parsertest << p.get_token().asInput();
// <origfile> and parsertest.tex should now have identical content
/// convert TeX from \p infilename to LyX and write it to \p os
-bool tex2lyx(FileName const & infilename, ostream & os)
+bool tex2lyx(FileName const & infilename, ostream & os, string const & encoding)
{
- ifstream is(infilename.toFilesystemEncoding().c_str());
+ ifdocstream is;
+ // forbid buffering on this stream
+ is.rdbuf()->pubsetbuf(0,0);
+ is.open(infilename.toFilesystemEncoding().c_str());
if (!is.good()) {
cerr << "Could not open input file \"" << infilename
<< "\" for reading." << endl;
return false;
}
string const oldParentFilePath = parentFilePath;
- parentFilePath = onlyPath(infilename.absFilename());
- tex2lyx(is, os);
+ parentFilePath = onlyPath(infilename.absFileName());
+ tex2lyx(is, os, encoding);
parentFilePath = oldParentFilePath;
return true;
}
} // anonymous namespace
-bool tex2lyx(string const & infilename, FileName const & outfilename)
+bool tex2lyx(string const & infilename, FileName const & outfilename,
+ string const & encoding)
{
if (outfilename.isReadableFile()) {
if (overwrite_files) {
cerr << "Input file: " << infilename << "\n";
cerr << "Output file: " << outfilename << "\n";
#endif
- return tex2lyx(FileName(infilename), os);
+ return tex2lyx(FileName(infilename), os, encoding);
+}
+
+
+bool tex2tex(string const & infilename, FileName const & outfilename,
+ string const & encoding)
+{
+ if (!tex2lyx(infilename, outfilename, encoding))
+ return false;
+ string command = quoteName(package().lyx_binary().toFilesystemEncoding());
+ if (overwrite_files)
+ command += " -f main";
+ else
+ command += " -f none";
+ if (pdflatex)
+ command += " -e pdflatex ";
+ else
+ command += " -e latex ";
+ command += quoteName(outfilename.toFilesystemEncoding());
+ Systemcall one;
+ if (one.startscript(Systemcall::Wait, command) == 0)
+ return true;
+ cerr << "Error: Running '" << command << "' failed." << endl;
+ return false;
}
} // namespace lyx
{
using namespace lyx;
+ //setlocale(LC_CTYPE, "");
+
lyxerr.setStream(cerr);
- easyParse(argc, argv);
+ os::init(argc, argv);
- if (argc <= 1) {
- cerr << "Usage: tex2lyx [ command line switches ] <infile.tex> [<outfile.lyx>]\n"
- "See tex2lyx -help." << endl;
- return 2;
- }
+ easyParse(argc, argv);
- os::init(argc, argv);
+ if (argc <= 1)
+ error_message("Not enough arguments.");
- try { init_package(internal_path(to_utf8(from_local8bit(argv[0]))),
- cl_system_support, cl_user_support,
- top_build_dir_is_two_levels_up);
+ try {
+ init_package(internal_path(os::utf8_argv(0)),
+ cl_system_support, cl_user_support,
+ top_build_dir_is_two_levels_up);
} catch (ExceptionMessage const & message) {
cerr << to_utf8(message.title_) << ":\n"
- << to_utf8(message.details_) << endl;
+ << to_utf8(message.details_) << endl;
if (message.type_ == ErrorException)
- exit(1);
+ return EXIT_FAILURE;
}
-
+
// Now every known option is parsed. Look for input and output
// file name (the latter is optional).
- string infilename = internal_path(to_utf8(from_local8bit(argv[1])));
- infilename = makeAbsPath(infilename).absFilename();
-
+ string infilename = internal_path(os::utf8_argv(1));
+ infilename = makeAbsPath(infilename).absFileName();
+
string outfilename;
- if (argc > 2) {
- outfilename = internal_path(to_utf8(from_local8bit(argv[2])));
+ 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) {
+ outfilename = internal_path(os::utf8_argv(2));
if (outfilename != "-")
- outfilename = makeAbsPath(outfilename).absFilename();
+ outfilename = makeAbsPath(outfilename).absFileName();
} else
outfilename = changeExtension(infilename, ".lyx");
+ // Read the syntax tables
FileName const system_syntaxfile = libFileSearch("", "syntax.default");
if (system_syntaxfile.empty()) {
cerr << "Error: Could not find syntax file \"syntax.default\"." << endl;
- exit(1);
+ return EXIT_FAILURE;
}
read_syntaxfile(system_syntaxfile);
if (!syntaxfile.empty())
read_syntaxfile(makeAbsPath(syntaxfile));
+ // Read the encodings table.
+ FileName const symbols_path = libFileSearch(string(), "unicodesymbols");
+ if (symbols_path.empty()) {
+ 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\"."
+ << 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 + "'");
+
+ // Load the layouts
+ LayoutFileList::get().read();
+ //...and the modules
+ theModuleList.read();
+
+ // The real work now.
masterFilePath = onlyPath(infilename);
parentFilePath = masterFilePath;
-
if (outfilename == "-") {
- if (tex2lyx(FileName(infilename), cout))
+ if (tex2lyx(FileName(infilename), cout, default_encoding))
+ return EXIT_SUCCESS;
+ } else if (roundtrip) {
+ if (tex2tex(infilename, FileName(outfilename), default_encoding))
return EXIT_SUCCESS;
- else
- return EXIT_FAILURE;
} else {
- if (tex2lyx(infilename, FileName(outfilename)))
+ if (tex2lyx(infilename, FileName(outfilename), default_encoding))
return EXIT_SUCCESS;
- else
- return EXIT_FAILURE;
}
+ return EXIT_FAILURE;
}
// }])