1 /**
2 * \file tex2lyx.cpp
3 * This file is part of LyX, the document processor.
4 * Licence details can be found in the file COPYING.
5 *
6 * \author André Pönitz
7 *
8 * Full author contact details are available in file CREDITS.
9 */
10
11 // {[(
12
13 #include <config.h>
14 #include <version.h>
15
16 #include "tex2lyx.h"
17
18 #include "Context.h"
19 #include "Encoding.h"
20 #include "Layout.h"
21 #include "LayoutFile.h"
22 #include "LayoutModuleList.h"
23 #include "ModuleList.h"
24 #include "Preamble.h"
25 #include "TextClass.h"
26
27 #include "support/ConsoleApplication.h"
28 #include "support/convert.h"
29 #include "support/ExceptionMessage.h"
30 #include "support/filetools.h"
31 #include "support/lassert.h"
32 #include "support/lstrings.h"
33 #include "support/os.h"
34 #include "support/Package.h"
35 #include "support/Systemcall.h"
36
37 #include <cstdlib>
38 #include <algorithm>
39 #include <exception>
40 #include <iostream>
41 #include <string>
42 #include <sstream>
43 #include <vector>
44 #include <map>
45
46 using namespace std;
47 using namespace lyx::support;
48 using namespace lyx::support::os;
49
50 namespace lyx {
51
trimSpaceAndEol(string const & a)52 string const trimSpaceAndEol(string const & a)
53 {
54 return trim(a, " \t\n\r");
55 }
56
57
split(string const & s,vector<string> & result,char delim)58 void split(string const & s, vector<string> & result, char delim)
59 {
60 //cerr << "split 1: '" << s << "'\n";
61 istringstream is(s);
62 string t;
63 while (getline(is, t, delim))
64 result.push_back(t);
65 //cerr << "split 2\n";
66 }
67
68
join(vector<string> const & input,char const * delim)69 string join(vector<string> const & input, char const * delim)
70 {
71 ostringstream os;
72 for (size_t i = 0; i != input.size(); ++i) {
73 if (i)
74 os << delim;
75 os << input[i];
76 }
77 return os.str();
78 }
79
80
is_known(string const & str,char const * const * what)81 char const * const * is_known(string const & str, char const * const * what)
82 {
83 for ( ; *what; ++what)
84 if (str == *what)
85 return what;
86 return 0;
87 }
88
89
90
91 // current stack of nested environments
92 vector<string> active_environments;
93
94
active_environment()95 string active_environment()
96 {
97 return active_environments.empty() ? string() : active_environments.back();
98 }
99
100
101 TeX2LyXDocClass textclass;
102 CommandMap known_commands;
103 CommandMap known_environments;
104 CommandMap known_math_environments;
105 FullCommandMap possible_textclass_commands;
106 FullEnvironmentMap possible_textclass_environments;
107 FullCommandMap possible_textclass_theorems;
108 int const LYX_FORMAT = LYX_FORMAT_TEX2LYX;
109
110 /// used modules
111 LayoutModuleList used_modules;
112 vector<string> preloaded_modules;
113
114
convertArgs(string const & o1,bool o2,vector<ArgumentType> & arguments)115 void convertArgs(string const & o1, bool o2, vector<ArgumentType> & arguments)
116 {
117 // We have to handle the following cases:
118 // definition o1 o2 invocation result
119 // \newcommand{\foo}{bar} "" false \foo bar
120 // \newcommand{\foo}[1]{bar #1} "[1]" false \foo{x} bar x
121 // \newcommand{\foo}[1][]{bar #1} "[1]" true \foo bar
122 // \newcommand{\foo}[1][]{bar #1} "[1]" true \foo[x] bar x
123 // \newcommand{\foo}[1][x]{bar #1} "[1]" true \foo[x] bar x
124 unsigned int nargs = 0;
125 string const opt1 = rtrim(ltrim(o1, "["), "]");
126 if (isStrUnsignedInt(opt1)) {
127 // The command has arguments
128 nargs = convert<unsigned int>(opt1);
129 if (nargs > 0 && o2) {
130 // The first argument is optional
131 arguments.push_back(optional);
132 --nargs;
133 }
134 }
135 for (unsigned int i = 0; i < nargs; ++i)
136 arguments.push_back(required);
137 }
138
139
add_known_command(string const & command,string const & o1,bool o2,docstring const & definition)140 void add_known_command(string const & command, string const & o1,
141 bool o2, docstring const & definition)
142 {
143 vector<ArgumentType> arguments;
144 convertArgs(o1, o2, arguments);
145 known_commands[command] = arguments;
146 if (!definition.empty())
147 possible_textclass_commands[command] =
148 FullCommand(arguments, definition);
149 }
150
151
add_known_environment(string const & environment,string const & o1,bool o2,docstring const & beg,docstring const & end)152 void add_known_environment(string const & environment, string const & o1,
153 bool o2, docstring const & beg, docstring const &end)
154 {
155 vector<ArgumentType> arguments;
156 convertArgs(o1, o2, arguments);
157 known_environments[environment] = arguments;
158 if (!beg.empty() || ! end.empty())
159 possible_textclass_environments[environment] =
160 FullEnvironment(arguments, beg, end);
161 }
162
163
add_known_theorem(string const & theorem,string const & o1,bool o2,docstring const & definition)164 void add_known_theorem(string const & theorem, string const & o1,
165 bool o2, docstring const & definition)
166 {
167 vector<ArgumentType> arguments;
168 convertArgs(o1, o2, arguments);
169 if (!definition.empty())
170 possible_textclass_theorems[theorem] =
171 FullCommand(arguments, definition);
172 }
173
174
findLayoutWithoutModule(TextClass const & tc,string const & name,bool command,string const & latexparam)175 Layout const * findLayoutWithoutModule(TextClass const & tc,
176 string const & name, bool command,
177 string const & latexparam)
178 {
179 for (auto const & lay : tc) {
180 if (lay.latexname() == name &&
181 (latexparam.empty() ||
182 (!lay.latexparam().empty() && suffixIs(latexparam, lay.latexparam()))) &&
183 ((command && lay.isCommand()) || (!command && lay.isEnvironment())))
184 return &lay;
185 }
186 return 0;
187 }
188
189
findInsetLayoutWithoutModule(TextClass const & tc,string const & name,bool command,string const & latexparam)190 InsetLayout const * findInsetLayoutWithoutModule(TextClass const & tc,
191 string const & name, bool command,
192 string const & latexparam)
193 {
194 for (auto const & ilay : tc.insetLayouts()) {
195 if (ilay.second.latexname() == name &&
196 (latexparam.empty() ||
197 (!ilay.second.latexparam().empty() && suffixIs(latexparam, ilay.second.latexparam()))) &&
198 ((command && ilay.second.latextype() == InsetLayout::COMMAND) ||
199 (!command && ilay.second.latextype() == InsetLayout::ENVIRONMENT)))
200 return &(ilay.second);
201 }
202 return 0;
203 }
204
205
206 namespace {
207
208 typedef map<string, DocumentClassPtr> ModuleMap;
209 ModuleMap modules;
210
211
addModule(string const & module,LayoutFile const & baseClass,LayoutModuleList & m,vector<string> & visited)212 bool addModule(string const & module, LayoutFile const & baseClass, LayoutModuleList & m, vector<string> & visited)
213 {
214 // avoid endless loop for circular dependency
215 vector<string>::const_iterator const vb = visited.begin();
216 vector<string>::const_iterator const ve = visited.end();
217 if (find(vb, ve, module) != ve) {
218 cerr << "Circular dependency detected for module " << module << '\n';
219 return false;
220 }
221 LyXModule const * const lm = theModuleList[module];
222 if (!lm) {
223 cerr << "Could not find module " << module << " in module list.\n";
224 return false;
225 }
226 bool foundone = false;
227 LayoutModuleList::const_iterator const exclmodstart = baseClass.excludedModules().begin();
228 LayoutModuleList::const_iterator const exclmodend = baseClass.excludedModules().end();
229 LayoutModuleList::const_iterator const provmodstart = baseClass.providedModules().begin();
230 LayoutModuleList::const_iterator const provmodend = baseClass.providedModules().end();
231 vector<string> const reqs = lm->getRequiredModules();
232 if (reqs.empty())
233 foundone = true;
234 else {
235 LayoutModuleList::const_iterator mit = m.begin();
236 LayoutModuleList::const_iterator men = m.end();
237 vector<string>::const_iterator rit = reqs.begin();
238 vector<string>::const_iterator ren = reqs.end();
239 for (; rit != ren; ++rit) {
240 if (find(mit, men, *rit) != men) {
241 foundone = true;
242 break;
243 }
244 if (find(provmodstart, provmodend, *rit) != provmodend) {
245 foundone = true;
246 break;
247 }
248 }
249 if (!foundone) {
250 visited.push_back(module);
251 for (rit = reqs.begin(); rit != ren; ++rit) {
252 if (find(exclmodstart, exclmodend, *rit) == exclmodend) {
253 if (addModule(*rit, baseClass, m, visited)) {
254 foundone = true;
255 break;
256 }
257 }
258 }
259 visited.pop_back();
260 }
261 }
262 if (!foundone) {
263 cerr << "Could not add required modules for " << module << ".\n";
264 return false;
265 }
266 if (!m.moduleCanBeAdded(module, &baseClass))
267 return false;
268 m.push_back(module);
269 return true;
270 }
271
272
initModules()273 void initModules()
274 {
275 // Create list of dummy document classes if not already done.
276 // This is needed since a module cannot be read on its own, only as
277 // part of a document class.
278 LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
279 static bool init = true;
280 if (init) {
281 baseClass.load();
282 LyXModuleList::const_iterator const end = theModuleList.end();
283 LyXModuleList::const_iterator it = theModuleList.begin();
284 for (; it != end; ++it) {
285 string const module = it->getID();
286 LayoutModuleList m;
287 LayoutModuleList c;
288 vector<string> v;
289 if (!addModule(module, baseClass, m, v))
290 continue;
291 modules[module] = getDocumentClass(baseClass, m, c);
292 }
293 init = false;
294 }
295 }
296
297
addModule(string const & module)298 bool addModule(string const & module)
299 {
300 initModules();
301 LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
302 if (!used_modules.moduleCanBeAdded(module, &baseClass))
303 return false;
304 FileName layout_file = libFileSearch("layouts", module, "module");
305 if (textclass.read(layout_file, TextClass::MODULE)) {
306 used_modules.push_back(module);
307 // speed up further searches:
308 // the module does not need to be checked anymore.
309 ModuleMap::iterator const it = modules.find(module);
310 if (it != modules.end())
311 modules.erase(it);
312 return true;
313 }
314 return false;
315 }
316
317 } // namespace
318
319
checkModule(string const & name,bool command)320 bool checkModule(string const & name, bool command)
321 {
322 // Cache to avoid slowdown by repated searches
323 static set<string> failed[2];
324
325 // Record whether the command was actually defined in the LyX preamble
326 bool theorem = false;
327 bool preamble_def = true;
328 if (command) {
329 if (possible_textclass_commands.find('\\' + name) == possible_textclass_commands.end())
330 preamble_def = false;
331 } else {
332 if (possible_textclass_environments.find(name) == possible_textclass_environments.end()) {
333 if (possible_textclass_theorems.find(name) != possible_textclass_theorems.end())
334 theorem = true;
335 else
336 preamble_def = false;
337 }
338 }
339 if (failed[command].find(name) != failed[command].end())
340 return false;
341
342 initModules();
343 LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
344
345 // Try to find a module that defines the command.
346 // For commands with preamble definitions we prefer modules where the definition
347 // can be found in the preamble of the style that corresponds to the command.
348 // For others we check whether the command or module requires a package that is loaded
349 // in the tex file and use a style with the respective command.
350 // This is a heuristic and different from the way how we parse the builtin
351 // commands of the text class (in that case we only compare the name),
352 // but it is needed since it is not unlikely that two different modules define a
353 // command with the same name.
354 string found_module;
355 vector<string> potential_modules;
356 ModuleMap::iterator const end = modules.end();
357 for (ModuleMap::iterator it = modules.begin(); it != end; ++it) {
358 string const module = it->first;
359 if (used_modules.moduleConflicts(module, &baseClass))
360 continue;
361 if (findLayoutWithoutModule(textclass, name, command))
362 continue;
363 if (findInsetLayoutWithoutModule(textclass, name, command))
364 continue;
365 DocumentClassConstPtr c = it->second;
366 Layout const * layout = findLayoutWithoutModule(*c, name, command);
367 InsetLayout const * insetlayout = layout ? nullptr :
368 findInsetLayoutWithoutModule(*c, name, command);
369 docstring dpre;
370 std::set<std::string> cmd_reqs;
371 bool found_style = false;
372 if (layout) {
373 found_style = true;
374 dpre = layout->preamble();
375 std::set<std::string> lreqs = layout->requires();
376 if (!lreqs.empty())
377 cmd_reqs.insert(lreqs.begin(), lreqs.end());
378 } else if (insetlayout) {
379 found_style = true;
380 dpre = insetlayout->preamble();
381 std::set<std::string> lreqs = insetlayout->requires();
382 if (!lreqs.empty())
383 cmd_reqs.insert(lreqs.begin(), lreqs.end());
384 }
385 if (dpre.empty() && preamble_def)
386 continue;
387 bool const package_cmd = dpre.empty();
388 bool match_req = false;
389 if (package_cmd) {
390 std::set<std::string> mreqs = it->second->requires();
391 if (!mreqs.empty())
392 cmd_reqs.insert(mreqs.begin(), mreqs.end());
393 for (auto const & pack : cmd_reqs) {
394 // If a requirement of the module matches a used package
395 // we load the module except if we have an auto-loaded package
396 // which is only required generally by the module, and the module
397 // does not provide the [inset]layout we are looking for.
398 // This heuristics should
399 // * load modules if the provide a style we don't have in the class
400 // * load modules that provide a package support generally (such as fixltx2e)
401 // * not unnecessarily load modules just because they require a package which we
402 // load anyway.
403 if (preamble.isPackageUsed(pack)
404 && (found_style || !preamble.isPackageAutoLoaded(pack))) {
405 if (found_style)
406 match_req = true;
407 else
408 potential_modules.push_back(module);
409 break;
410 }
411 }
412 }
413 bool add = match_req;
414 if (preamble_def) {
415 if (command) {
416 FullCommand const & cmd =
417 possible_textclass_commands['\\' + name];
418 if (dpre.find(cmd.def) != docstring::npos)
419 add = true;
420 } else if (theorem) {
421 FullCommand const & thm =
422 possible_textclass_theorems[name];
423 if (dpre.find(thm.def) != docstring::npos)
424 add = true;
425 } else {
426 FullEnvironment const & env =
427 possible_textclass_environments[name];
428 if (dpre.find(env.beg) != docstring::npos &&
429 dpre.find(env.end) != docstring::npos)
430 add = true;
431 }
432 }
433 if (add) {
434 found_module = module;
435 break;
436 }
437 }
438 if (found_module.empty()) {
439 // take one of the second row
440 if (!potential_modules.empty())
441 found_module = potential_modules.front();
442 }
443
444 if (!found_module.empty()) {
445 vector<string> v;
446 LayoutModuleList mods;
447 // addModule is necessary in order to catch required modules
448 // as well (see #11156)
449 if (!addModule(found_module, baseClass, mods, v))
450 return false;
451 for (auto const & mod : mods) {
452 if (!used_modules.moduleCanBeAdded(mod, &baseClass))
453 return false;
454 FileName layout_file = libFileSearch("layouts", mod, "module");
455 if (textclass.read(layout_file, TextClass::MODULE)) {
456 used_modules.push_back(mod);
457 // speed up further searches:
458 // the module does not need to be checked anymore.
459 ModuleMap::iterator const it = modules.find(mod);
460 if (it != modules.end())
461 modules.erase(it);
462 return true;
463 }
464 }
465 }
466 failed[command].insert(name);
467 return false;
468 }
469
470
isProvided(string const & name)471 bool isProvided(string const & name)
472 {
473 // This works only for features that are named like the LaTeX packages
474 return textclass.provides(name) || preamble.isPackageUsed(name);
475 }
476
477
478 bool noweb_mode = false;
479 bool pdflatex = false;
480 bool xetex = false;
481 bool is_nonCJKJapanese = false;
482 bool roundtrip = false;
483
484
485 namespace {
486
487
488 /*!
489 * Read one command definition from the syntax file
490 */
read_command(Parser & p,string command,CommandMap & commands)491 void read_command(Parser & p, string command, CommandMap & commands)
492 {
493 if (p.next_token().asInput() == "*") {
494 p.get_token();
495 command += '*';
496 }
497 vector<ArgumentType> arguments;
498 while (p.next_token().cat() == catBegin ||
499 p.next_token().asInput() == "[") {
500 if (p.next_token().cat() == catBegin) {
501 string const arg = p.getArg('{', '}');
502 if (arg == "translate")
503 arguments.push_back(required);
504 else if (arg == "group")
505 arguments.push_back(req_group);
506 else if (arg == "item")
507 arguments.push_back(item);
508 else if (arg == "displaymath")
509 arguments.push_back(displaymath);
510 else
511 arguments.push_back(verbatim);
512 } else {
513 string const arg = p.getArg('[', ']');
514 if (arg == "group")
515 arguments.push_back(opt_group);
516 else
517 arguments.push_back(optional);
518 }
519 }
520 commands[command] = arguments;
521 }
522
523
524 /*!
525 * Read a class of environments from the syntax file
526 */
read_environment(Parser & p,string const & begin,CommandMap & environments)527 void read_environment(Parser & p, string const & begin,
528 CommandMap & environments)
529 {
530 string environment;
531 while (p.good()) {
532 Token const & t = p.get_token();
533 if (t.cat() == catLetter)
534 environment += t.asInput();
535 else if (!environment.empty()) {
536 p.putback();
537 read_command(p, environment, environments);
538 environment.erase();
539 }
540 if (t.cat() == catEscape && t.asInput() == "\\end") {
541 string const end = p.getArg('{', '}');
542 if (end == begin)
543 return;
544 }
545 }
546 }
547
548
549 /*!
550 * Read a list of TeX commands from a reLyX compatible syntax file.
551 * Since this list is used after all commands that have a LyX counterpart
552 * are handled, it does not matter that the "syntax.default" file
553 * has almost all of them listed. For the same reason the reLyX-specific
554 * reLyXre environment is ignored.
555 */
read_syntaxfile(FileName const & file_name)556 bool read_syntaxfile(FileName const & file_name)
557 {
558 ifdocstream is(file_name.toFilesystemEncoding().c_str());
559 if (!is.good()) {
560 cerr << "Could not open syntax file \"" << file_name
561 << "\" for reading." << endl;
562 return false;
563 }
564 // We can use our TeX parser, since the syntax of the layout file is
565 // modeled after TeX.
566 // Unknown tokens are just silently ignored, this helps us to skip some
567 // reLyX specific things.
568 Parser p(is, string());
569 while (p.good()) {
570 Token const & t = p.get_token();
571 if (t.cat() == catEscape) {
572 string const command = t.asInput();
573 if (command == "\\begin") {
574 string const name = p.getArg('{', '}');
575 if (name == "environments" || name == "reLyXre")
576 // We understand "reLyXre", but it is
577 // not as powerful as "environments".
578 read_environment(p, name,
579 known_environments);
580 else if (name == "mathenvironments")
581 read_environment(p, name,
582 known_math_environments);
583 } else {
584 read_command(p, command, known_commands);
585 }
586 }
587 }
588 return true;
589 }
590
591
592 string documentclass;
593 string default_encoding;
594 bool fixed_encoding = false;
595 string syntaxfile;
596 bool copy_files = false;
597 bool overwrite_files = false;
598 bool skip_children = false;
599 int error_code = 0;
600
601 /// return the number of arguments consumed
602 typedef int (*cmd_helper)(string const &, string const &);
603
604
605 class StopException : public exception
606 {
607 public:
StopException(int status)608 StopException(int status) : status_(status) {}
status() const609 int status() const { return status_; }
610 private:
611 int status_;
612 };
613
614
615 /// The main application class
616 class TeX2LyXApp : public ConsoleApplication
617 {
618 public:
TeX2LyXApp(int & argc,char * argv[])619 TeX2LyXApp(int & argc, char * argv[])
620 : ConsoleApplication("tex2lyx" PROGRAM_SUFFIX, argc, argv),
621 argc_(argc), argv_(argv)
622 {
623 }
doExec()624 void doExec()
625 {
626 try {
627 int const exit_status = run();
628 exit(exit_status);
629 }
630 catch (StopException & e) {
631 exit(e.status());
632 }
633 }
634 private:
635 void easyParse();
636 /// Do the real work
637 int run();
638 int & argc_;
639 char ** argv_;
640 };
641
642
parse_help(string const &,string const &)643 int parse_help(string const &, string const &)
644 {
645 cout << "Usage: tex2lyx [options] infile.tex [outfile.lyx]\n"
646 "Options:\n"
647 "\t-c textclass Declare the textclass.\n"
648 "\t-m mod1[,mod2...] Load the given modules.\n"
649 "\t-copyfiles Copy all included files to the directory of outfile.lyx.\n"
650 "\t-e encoding Set the default encoding (latex name).\n"
651 "\t-fixedenc encoding Like -e, but ignore encoding changing commands while parsing.\n"
652 "\t-f Force overwrite of .lyx files.\n"
653 "\t-help Print this message and quit.\n"
654 "\t-n translate literate programming (noweb, sweave,... ) file.\n"
655 "\t-skipchildren Do not translate included child documents.\n"
656 "\t-roundtrip re-export created .lyx file infile.lyx.lyx to infile.lyx.tex.\n"
657 "\t-s syntaxfile read additional syntax file.\n"
658 "\t-sysdir SYSDIR Set system directory to SYSDIR.\n"
659 "\t Default: " << package().system_support() << "\n"
660 "\t-userdir USERDIR Set user directory to USERDIR.\n"
661 "\t Default: " << package().user_support() << "\n"
662 "\t-version Summarize version and build info.\n"
663 "Paths:\n"
664 "\tThe program searches for the files \"encodings\", \"lyxmodules.lst\",\n"
665 "\t\"textclass.lst\", \"syntax.default\", and \"unicodesymbols\", first in\n"
666 "\t\"USERDIR\", then in \"SYSDIR\". The subdirectories \"USERDIR/layouts\"\n"
667 "\tand \"SYSDIR/layouts\" are searched for layout and module files.\n"
668 "Check the tex2lyx man page for more details."
669 << endl;
670 throw StopException(error_code);
671 }
672
673
parse_version(string const &,string const &)674 int parse_version(string const &, string const &)
675 {
676 cout << "tex2lyx " << lyx_version
677 << " (" << lyx_release_date << ")" << endl;
678
679 cout << lyx_version_info << endl;
680 throw StopException(error_code);
681 }
682
683
error_message(string const & message)684 void error_message(string const & message)
685 {
686 cerr << "tex2lyx: " << message << "\n\n";
687 error_code = EXIT_FAILURE;
688 parse_help(string(), string());
689 }
690
691
parse_class(string const & arg,string const &)692 int parse_class(string const & arg, string const &)
693 {
694 if (arg.empty())
695 error_message("Missing textclass string after -c switch");
696 documentclass = arg;
697 return 1;
698 }
699
700
parse_module(string const & arg,string const &)701 int parse_module(string const & arg, string const &)
702 {
703 if (arg.empty())
704 error_message("Missing modules string after -m switch");
705 split(arg, preloaded_modules, ',');
706 return 1;
707 }
708
709
parse_encoding(string const & arg,string const &)710 int parse_encoding(string const & arg, string const &)
711 {
712 if (arg.empty())
713 error_message("Missing encoding string after -e switch");
714 default_encoding = arg;
715 return 1;
716 }
717
718
parse_fixed_encoding(string const & arg,string const &)719 int parse_fixed_encoding(string const & arg, string const &)
720 {
721 if (arg.empty())
722 error_message("Missing encoding string after -fixedenc switch");
723 default_encoding = arg;
724 fixed_encoding = true;
725 return 1;
726 }
727
728
parse_syntaxfile(string const & arg,string const &)729 int parse_syntaxfile(string const & arg, string const &)
730 {
731 if (arg.empty())
732 error_message("Missing syntaxfile string after -s switch");
733 syntaxfile = internal_path(arg);
734 return 1;
735 }
736
737
738 // Filled with the command line arguments "foo" of "-sysdir foo" or
739 // "-userdir foo".
740 string cl_system_support;
741 string cl_user_support;
742
743
parse_sysdir(string const & arg,string const &)744 int parse_sysdir(string const & arg, string const &)
745 {
746 if (arg.empty())
747 error_message("Missing directory for -sysdir switch");
748 cl_system_support = internal_path(arg);
749 return 1;
750 }
751
752
parse_userdir(string const & arg,string const &)753 int parse_userdir(string const & arg, string const &)
754 {
755 if (arg.empty())
756 error_message("Missing directory for -userdir switch");
757 cl_user_support = internal_path(arg);
758 return 1;
759 }
760
761
parse_force(string const &,string const &)762 int parse_force(string const &, string const &)
763 {
764 overwrite_files = true;
765 return 0;
766 }
767
768
parse_noweb(string const &,string const &)769 int parse_noweb(string const &, string const &)
770 {
771 noweb_mode = true;
772 return 0;
773 }
774
775
parse_skipchildren(string const &,string const &)776 int parse_skipchildren(string const &, string const &)
777 {
778 skip_children = true;
779 return 0;
780 }
781
782
parse_roundtrip(string const &,string const &)783 int parse_roundtrip(string const &, string const &)
784 {
785 roundtrip = true;
786 return 0;
787 }
788
789
parse_copyfiles(string const &,string const &)790 int parse_copyfiles(string const &, string const &)
791 {
792 copy_files = true;
793 return 0;
794 }
795
796
easyParse()797 void TeX2LyXApp::easyParse()
798 {
799 map<string, cmd_helper> cmdmap;
800
801 cmdmap["-h"] = parse_help;
802 cmdmap["-help"] = parse_help;
803 cmdmap["--help"] = parse_help;
804 cmdmap["-v"] = parse_version;
805 cmdmap["-version"] = parse_version;
806 cmdmap["--version"] = parse_version;
807 cmdmap["-c"] = parse_class;
808 cmdmap["-m"] = parse_module;
809 cmdmap["-e"] = parse_encoding;
810 cmdmap["-fixedenc"] = parse_fixed_encoding;
811 cmdmap["-f"] = parse_force;
812 cmdmap["-s"] = parse_syntaxfile;
813 cmdmap["-n"] = parse_noweb;
814 cmdmap["-skipchildren"] = parse_skipchildren;
815 cmdmap["-sysdir"] = parse_sysdir;
816 cmdmap["-userdir"] = parse_userdir;
817 cmdmap["-roundtrip"] = parse_roundtrip;
818 cmdmap["-copyfiles"] = parse_copyfiles;
819
820 for (int i = 1; i < argc_; ++i) {
821 map<string, cmd_helper>::const_iterator it
822 = cmdmap.find(argv_[i]);
823
824 // don't complain if not found - may be parsed later
825 if (it == cmdmap.end()) {
826 if (argv_[i][0] == '-')
827 error_message(string("Unknown option `") + argv_[i] + "'.");
828 else
829 continue;
830 }
831
832 string arg = (i + 1 < argc_) ? os::utf8_argv(i + 1) : string();
833 string arg2 = (i + 2 < argc_) ? os::utf8_argv(i + 2) : string();
834
835 int const remove = 1 + it->second(arg, arg2);
836
837 // Now, remove used arguments by shifting
838 // the following ones remove places down.
839 os::remove_internal_args(i, remove);
840 argc_ -= remove;
841 for (int j = i; j < argc_; ++j)
842 argv_[j] = argv_[j + remove];
843 --i;
844 }
845 }
846
847
848 // path of the first parsed file
849 string masterFilePathLyX;
850 string masterFilePathTeX;
851 // path of the currently parsed file
852 string parentFilePathTeX;
853
854 } // anonymous namespace
855
856
getMasterFilePath(bool input)857 string getMasterFilePath(bool input)
858 {
859 return input ? masterFilePathTeX : masterFilePathLyX;
860 }
861
getParentFilePath(bool input)862 string getParentFilePath(bool input)
863 {
864 if (input)
865 return parentFilePathTeX;
866 string const rel = to_utf8(makeRelPath(from_utf8(masterFilePathTeX),
867 from_utf8(parentFilePathTeX)));
868 if (rel.substr(0, 3) == "../") {
869 // The parent is not below the master - keep the path
870 return parentFilePathTeX;
871 }
872 return makeAbsPath(rel, masterFilePathLyX).absFileName();
873 }
874
875
copyFiles()876 bool copyFiles()
877 {
878 return copy_files;
879 }
880
881
overwriteFiles()882 bool overwriteFiles()
883 {
884 return overwrite_files;
885 }
886
887
skipChildren()888 bool skipChildren()
889 {
890 return skip_children;
891 }
892
893
roundtripMode()894 bool roundtripMode()
895 {
896 return roundtrip;
897 }
898
899
900 namespace {
901
902 /*!
903 * Reads tex input from \a is and writes lyx output to \a os.
904 * Uses some common settings for the preamble, so this should only
905 * be used more than once for included documents.
906 * Caution: Overwrites the existing preamble settings if the new document
907 * contains a preamble.
908 * You must ensure that \p parentFilePathTeX is properly set before calling
909 * this function!
910 */
tex2lyx(idocstream & is,ostream & os,string const & encoding,string const & outfiledir)911 bool tex2lyx(idocstream & is, ostream & os, string const & encoding,
912 string const & outfiledir)
913 {
914 Parser p(is, fixed_encoding ? default_encoding : string());
915 p.setEncoding(encoding);
916 //p.dump();
917
918 preamble.parse(p, documentclass, textclass);
919 list<string> removed_modules;
920 LayoutFile const & baseClass = LayoutFileList::get()[textclass.name()];
921 if (!used_modules.adaptToBaseClass(&baseClass, removed_modules)) {
922 cerr << "Could not load default modules for text class." << endl;
923 return false;
924 }
925
926 // Load preloaded modules.
927 // This needs to be done after the preamble is parsed, since the text
928 // class may not be known before. It neds to be done before parsing
929 // body, since otherwise the commands/environments provided by the
930 // modules would be parsed as ERT.
931 for (size_t i = 0; i < preloaded_modules.size(); ++i) {
932 if (!addModule(preloaded_modules[i])) {
933 cerr << "Error: Could not load module \""
934 << preloaded_modules[i] << "\"." << endl;
935 return false;
936 }
937 }
938 // Ensure that the modules are not loaded again for included files
939 preloaded_modules.clear();
940
941 active_environments.push_back("document");
942 Context context(true, textclass);
943 stringstream ss;
944 // store the document language in the context to be able to handle the
945 // commands like \foreignlanguage and \textenglish etc.
946 context.font.language = preamble.defaultLanguage();
947 // parse the main text
948 parse_text(p, ss, FLAG_END, true, context);
949 // check if we need a commented bibtex inset (biblatex)
950 check_comment_bib(ss, context);
951 if (Context::empty)
952 // Empty document body. LyX needs at least one paragraph.
953 context.check_layout(ss);
954 context.check_end_layout(ss);
955 ss << "\n\\end_body\n\\end_document\n";
956 active_environments.pop_back();
957
958 // We know the used modules only after parsing the full text
959 if (!used_modules.empty()) {
960 LayoutModuleList::const_iterator const end = used_modules.end();
961 LayoutModuleList::const_iterator it = used_modules.begin();
962 for (; it != end; ++it)
963 preamble.addModule(*it);
964 }
965 if (!preamble.writeLyXHeader(os, !active_environments.empty(), outfiledir)) {
966 cerr << "Could not write LyX file header." << endl;
967 return false;
968 }
969
970 ss.seekg(0);
971 os << ss.str();
972 #ifdef TEST_PARSER
973 p.reset();
974 ofdocstream parsertest("parsertest.tex");
975 while (p.good())
976 parsertest << p.get_token().asInput();
977 // <origfile> and parsertest.tex should now have identical content
978 #endif
979 return true;
980 }
981
982
983 /// convert TeX from \p infilename to LyX and write it to \p os
tex2lyx(FileName const & infilename,ostream & os,string encoding,string const & outfiledir)984 bool tex2lyx(FileName const & infilename, ostream & os, string encoding,
985 string const & outfiledir)
986 {
987 // Set a sensible default encoding.
988 // This is used until an encoding command is found.
989 // For child documents use the encoding of the master, else try to
990 // detect it from the preamble, since setting an encoding of an open
991 // fstream does currently not work on OS X.
992 // Always start with ISO-8859-1, (formerly known by its latex name
993 // latin1), since ISO-8859-1 does not cause an iconv error if the
994 // actual encoding is different (bug 7509).
995 if (encoding.empty()) {
996 Encoding const * enc = 0;
997 if (preamble.inputencoding() == "auto") {
998 ifdocstream is(setEncoding("ISO-8859-1"));
999 // forbid buffering on this stream
1000 is.rdbuf()->pubsetbuf(0, 0);
1001 is.open(infilename.toFilesystemEncoding().c_str());
1002 if (is.good()) {
1003 Parser ep(is, string());
1004 ep.setEncoding("ISO-8859-1");
1005 Preamble encodingpreamble;
1006 string const e = encodingpreamble
1007 .parseEncoding(ep, documentclass);
1008 if (!e.empty())
1009 enc = encodings.fromLyXName(e, true);
1010 }
1011 } else
1012 enc = encodings.fromLyXName(
1013 preamble.inputencoding(), true);
1014 if (enc)
1015 encoding = enc->iconvName();
1016 else
1017 encoding = "ISO-8859-1";
1018 }
1019
1020 ifdocstream is(setEncoding(encoding));
1021 // forbid buffering on this stream
1022 is.rdbuf()->pubsetbuf(0, 0);
1023 is.open(infilename.toFilesystemEncoding().c_str());
1024 if (!is.good()) {
1025 cerr << "Could not open input file \"" << infilename
1026 << "\" for reading." << endl;
1027 return false;
1028 }
1029 string const oldParentFilePath = parentFilePathTeX;
1030 parentFilePathTeX = onlyPath(infilename.absFileName());
1031 bool retval = tex2lyx(is, os, encoding, outfiledir);
1032 parentFilePathTeX = oldParentFilePath;
1033 return retval;
1034 }
1035
1036 } // anonymous namespace
1037
1038
tex2lyx(string const & infilename,FileName const & outfilename,string const & encoding)1039 bool tex2lyx(string const & infilename, FileName const & outfilename,
1040 string const & encoding)
1041 {
1042 if (outfilename.isReadableFile()) {
1043 if (overwrite_files) {
1044 cerr << "Overwriting existing file "
1045 << outfilename << endl;
1046 } else {
1047 cerr << "Not overwriting existing file "
1048 << outfilename << endl;
1049 return false;
1050 }
1051 } else {
1052 cerr << "Creating file " << outfilename << endl;
1053 }
1054 ofstream os(outfilename.toFilesystemEncoding().c_str());
1055 if (!os.good()) {
1056 cerr << "Could not open output file \"" << outfilename
1057 << "\" for writing." << endl;
1058 return false;
1059 }
1060 #ifdef FILEDEBUG
1061 cerr << "Input file: " << infilename << "\n";
1062 cerr << "Output file: " << outfilename << "\n";
1063 #endif
1064 return tex2lyx(FileName(infilename), os, encoding,
1065 outfilename.onlyPath().absFileName() + '/');
1066 }
1067
1068
tex2tex(string const & infilename,FileName const & outfilename,string const & encoding)1069 bool tex2tex(string const & infilename, FileName const & outfilename,
1070 string const & encoding)
1071 {
1072 if (!tex2lyx(infilename, outfilename, encoding))
1073 return false;
1074 string command = quoteName(package().lyx_binary().toFilesystemEncoding());
1075 if (overwrite_files)
1076 command += " -f main";
1077 else
1078 command += " -f none";
1079 if (pdflatex)
1080 command += " -e pdflatex ";
1081 else if (xetex)
1082 command += " -e xetex ";
1083 else
1084 command += " -e latex ";
1085 command += quoteName(outfilename.toFilesystemEncoding());
1086 Systemcall one;
1087 if (one.startscript(Systemcall::Wait, command) == 0)
1088 return true;
1089 cerr << "Error: Running '" << command << "' failed." << endl;
1090 return false;
1091 }
1092
1093
1094 namespace {
1095
run()1096 int TeX2LyXApp::run()
1097 {
1098 // qt changes this, and our numeric conversions require the C locale
1099 setlocale(LC_NUMERIC, "C");
1100
1101 try {
1102 init_package(internal_path(os::utf8_argv(0)), string(), string());
1103 } catch (ExceptionMessage const & message) {
1104 cerr << to_utf8(message.title_) << ":\n"
1105 << to_utf8(message.details_) << endl;
1106 if (message.type_ == ErrorException)
1107 return EXIT_FAILURE;
1108 }
1109
1110 easyParse();
1111
1112 if (argc_ <= 1)
1113 error_message("Not enough arguments.");
1114
1115 try {
1116 init_package(internal_path(os::utf8_argv(0)),
1117 cl_system_support, cl_user_support);
1118 } catch (ExceptionMessage const & message) {
1119 cerr << to_utf8(message.title_) << ":\n"
1120 << to_utf8(message.details_) << endl;
1121 if (message.type_ == ErrorException)
1122 return EXIT_FAILURE;
1123 }
1124
1125 // Check that user LyX directory is ok.
1126 FileName const sup = package().user_support();
1127 if (sup.exists() && sup.isDirectory()) {
1128 string const lock_file = package().getConfigureLockName();
1129 int fd = fileLock(lock_file.c_str());
1130 if (configFileNeedsUpdate("lyxrc.defaults") ||
1131 configFileNeedsUpdate("lyxmodules.lst") ||
1132 configFileNeedsUpdate("textclass.lst") ||
1133 configFileNeedsUpdate("packages.lst") ||
1134 configFileNeedsUpdate("lyxciteengines.lst") ||
1135 configFileNeedsUpdate("xtemplates.lst"))
1136 package().reconfigureUserLyXDir("");
1137 fileUnlock(fd, lock_file.c_str());
1138 } else
1139 error_message("User directory does not exist.");
1140
1141 // Now every known option is parsed. Look for input and output
1142 // file name (the latter is optional).
1143 string infilename = internal_path(os::utf8_argv(1));
1144 infilename = makeAbsPath(infilename).absFileName();
1145
1146 string outfilename;
1147 if (argc_ > 2) {
1148 outfilename = internal_path(os::utf8_argv(2));
1149 if (outfilename != "-")
1150 outfilename = makeAbsPath(outfilename).absFileName();
1151 if (roundtrip) {
1152 if (outfilename == "-") {
1153 cerr << "Error: Writing to standard output is "
1154 "not supported in roundtrip mode."
1155 << endl;
1156 return EXIT_FAILURE;
1157 }
1158 string texfilename = changeExtension(outfilename, ".tex");
1159 if (equivalent(FileName(infilename), FileName(texfilename))) {
1160 cerr << "Error: The input file `" << infilename
1161 << "´ would be overwritten by the TeX file exported from `"
1162 << outfilename << "´ in roundtrip mode." << endl;
1163 return EXIT_FAILURE;
1164 }
1165 }
1166 } else if (roundtrip) {
1167 // avoid overwriting the input file
1168 outfilename = changeExtension(infilename, ".lyx.lyx");
1169 } else
1170 outfilename = changeExtension(infilename, ".lyx");
1171
1172 // Read the syntax tables
1173 FileName const system_syntaxfile = libFileSearch("", "syntax.default");
1174 if (system_syntaxfile.empty()) {
1175 cerr << "Error: Could not find syntax file \"syntax.default\"." << endl;
1176 return EXIT_FAILURE;
1177 }
1178 if (!read_syntaxfile(system_syntaxfile))
1179 return 2;
1180 if (!syntaxfile.empty())
1181 if (!read_syntaxfile(makeAbsPath(syntaxfile)))
1182 return 2;
1183
1184 // Read the encodings table.
1185 FileName const symbols_path = libFileSearch(string(), "unicodesymbols");
1186 if (symbols_path.empty()) {
1187 cerr << "Error: Could not find file \"unicodesymbols\"."
1188 << endl;
1189 return EXIT_FAILURE;
1190 }
1191 FileName const enc_path = libFileSearch(string(), "encodings");
1192 if (enc_path.empty()) {
1193 cerr << "Error: Could not find file \"encodings\"."
1194 << endl;
1195 return EXIT_FAILURE;
1196 }
1197 encodings.read(enc_path, symbols_path);
1198 if (!default_encoding.empty()) {
1199 Encoding const * const enc = encodings.fromLaTeXName(
1200 default_encoding, Encoding::any, true);
1201 if (!enc)
1202 error_message("Unknown LaTeX encoding `" + default_encoding + "'");
1203 default_encoding = enc->iconvName();
1204 if (fixed_encoding)
1205 preamble.setInputencoding(enc->name());
1206 }
1207
1208 // Load the layouts
1209 LayoutFileList::get().read();
1210 //...and the modules
1211 theModuleList.read();
1212
1213 // The real work now.
1214 masterFilePathTeX = onlyPath(infilename);
1215 parentFilePathTeX = masterFilePathTeX;
1216 if (outfilename == "-") {
1217 // assume same directory as input file
1218 masterFilePathLyX = masterFilePathTeX;
1219 if (tex2lyx(FileName(infilename), cout, default_encoding, masterFilePathLyX))
1220 return EXIT_SUCCESS;
1221 } else {
1222 masterFilePathLyX = onlyPath(outfilename);
1223 if (copy_files) {
1224 FileName const path(masterFilePathLyX);
1225 if (!path.isDirectory()) {
1226 if (!path.createPath()) {
1227 cerr << "Warning: Could not create directory for file `"
1228 << masterFilePathLyX << "´." << endl;
1229 return EXIT_FAILURE;
1230 }
1231 }
1232 }
1233 if (roundtrip) {
1234 if (tex2tex(infilename, FileName(outfilename), default_encoding))
1235 return EXIT_SUCCESS;
1236 } else {
1237 if (lyx::tex2lyx(infilename, FileName(outfilename), default_encoding))
1238 return EXIT_SUCCESS;
1239 }
1240 }
1241 return EXIT_FAILURE;
1242 }
1243
1244 } // anonymous namespace
1245 } // namespace lyx
1246
1247
main(int argc,char * argv[])1248 int main(int argc, char * argv[])
1249 {
1250 //setlocale(LC_CTYPE, "");
1251
1252 lyx::lyxerr.setStream(cerr);
1253
1254 os::init(argc, &argv);
1255
1256 lyx::TeX2LyXApp app(argc, argv);
1257 return app.exec();
1258 }
1259
1260 // }])
1261