]> git.lyx.org Git - lyx.git/blobdiff - src/tex2lyx/preamble.cpp
tex2lyx/text.cpp: take care of the brace pair after comment environments that are...
[lyx.git] / src / tex2lyx / preamble.cpp
index 77f72cdb315eaac7172cebebd3aef71b63094e3d..0a8f577b36bfb8c9c39c1cbcffe5c2c72a75855e 100644 (file)
 
 #include "tex2lyx.h"
 
+#include "LayoutFile.h"
 #include "Layout.h"
 #include "Lexer.h"
 #include "TextClass.h"
+
+#include "support/convert.h"
+#include "support/FileName.h"
 #include "support/filetools.h"
 #include "support/lstrings.h"
 
 #include <vector>
 #include <map>
 
+using namespace std;
+using namespace lyx::support;
 
 namespace lyx {
 
-using std::find;
-using std::istringstream;
-using std::ostream;
-using std::ostringstream;
-using std::string;
-using std::vector;
-using std::cerr;
-using std::endl;
-using std::find;
+// special columntypes
+extern map<char, int> special_columns;
 
-using support::FileName;
-using support::libFileSearch;
+map<string, vector<string> > used_packages;
 
-// special columntypes
-extern std::map<char, int> special_columns;
+// needed to handle encodings with babel
+bool one_language = true;
 
-std::map<string, vector<string> > used_packages;
+// to avoid that the babel options overwrite the documentclass options
+bool documentclass_language;
 
 namespace {
 
-const char * const known_languages[] = { "afrikaans", "albanian", "american",
-"arabic_arabtex", "arabic_arabi", "armenian", "austrian", "bahasa", "basque",
-"brazilian", "breton", "british", "bulgarian", "canadian", "canadien",
-"catalan", "chinese-simplified", "chinese-traditional", "croatian", "czech",
-"danish", "dutch", "english", "esperanto", "estonian", "farsi", "finnish",
-"francais", "french", "galician", "german", "greek", "hebrew", "icelandic",
-"irish", "italian", "japanese", "japanese-plain", "kazakh", "korean", "latin",
-"latvian", "lithuanian", "lowersorbian", "magyar", "naustrian", "ngerman",
-"norsk", "nynorsk ", "polish", "portuges", "romanian", "russian", "samin",
-"scottish", "serbian", "slovak", "slovene", "spanish", "swedish", "thai",
-"turkish", "ukrainian", "uppersorbian", "vietnamese", "welsh", 0};
-
-const char * const known_fontsizes[] = { "10pt", "11pt", "12pt", 0 };
-
-const char * const known_encodings[] = { "auto", "latin1", "latin2", "latin3",
-"latin4", "latin5", "latin9", "latin10", "iso88595", "8859-6", "iso-8859-7",
-"8859-8", "l7xen", "cp437", "cp437de", "cp850", "cp852", "cp855", "cp858",
-"cp862", "cp865", "cp866", "cp1250", "cp1251", "cp1252", "cp1255", "cp1256",
-"cp1257", "koi8-r", "koi8-u", "pt154", "pt254", "utf8", 0};
+const char * const known_languages[] = { "afrikaans", "american", "arabic",
+"austrian", "bahasa", "basque", "belarusian", "brazil", "breton", "british",
+"bulgarian", "canadian", "canadien", "catalan", "croatian", "czech", "danish",
+"dutch", "english", "esperanto", "estonian", "finnish", "francais", "french",
+"frenchb", "frenchle", "frenchpro", "galician", "german", "germanb", "greek",
+"hebrew", "icelandic", "irish", "italian", "lsorbian", "magyar", "naustrian",
+"ngerman", "ngermanb", "norsk", "nynorsk", "polish", "portuges", "romanian",
+"russian", "russianb", "scottish", "serbian", "slovak", "slovene", "spanish",
+"swedish", "thai", "turkish", "ukraineb", "ukrainian", "usorbian", "welsh", 0};
+
+//note this when updating to lyxformat 305:
+//bahasai, indonesian, and indon = equal to bahasa
+//malay, and meyalu = equal to bahasam
+
+const char * const known_french_languages[] = {"french", "frenchb", "francais",
+                                               "frenchle", "frenchpro", 0};
+const char * const known_german_languages[] = {"german", "germanb", 0};
+const char * const known_ngerman_languages[] = {"ngerman", "ngermanb", 0};
+const char * const known_russian_languages[] = {"russian", "russianb", 0};
+const char * const known_ukrainian_languages[] = {"ukrainian", "ukraineb", 0};
+
+char const * const known_fontsizes[] = { "10pt", "11pt", "12pt", 0 };
 
 const char * const known_roman_fonts[] = { "ae", "bookman", "charter",
 "cmr", "fourier", "lmodern", "mathpazo", "mathptmx", "newcent", 0};
@@ -83,39 +85,37 @@ const char * const known_typewriter_fonts[] = { "beramono", "cmtl", "cmtt",
 
 // some ugly stuff
 ostringstream h_preamble;
-string h_textclass                             = "article";
-string h_options                               = string();
-string h_language                              = "english";
-string h_inputencoding                 = "auto";
-string h_font_roman                            = "default";
-string h_font_sans                             = "default";
-string h_font_typewriter               = "default";
-string h_font_default_family   = "default";
-string h_font_sc                               = "false";
-string h_font_osf                              = "false";
-string h_font_sf_scale                 = "100";
-string h_font_tt_scale                 = "100";
-string h_graphics                              = "default";
-string h_paperfontsize                 = "default";
-string h_spacing                               = "single";
-string h_papersize                             = "default";
-string h_use_geometry                  = "false";
-string h_use_amsmath                   = "1";
-string h_use_esint                             = "1";
-string h_cite_engine                   = "basic";
-string h_use_bibtopic                  = "false";
-string h_paperorientation              = "portrait";
-string h_secnumdepth                   = "3";
-string h_tocdepth                              = "3";
-string h_paragraph_separation  = "indent";
-string h_defskip                               = "medskip";
-string h_quotes_language               = "english";
-string h_papercolumns                  = "1";
-string h_papersides                            = string();
-string h_paperpagestyle                        = "default";
-string h_tracking_changes              = "false";
-string h_output_changes                        = "false";
-string h_use_hyperref                  = "false";
+string h_textclass               = "article";
+string h_options                 = string();
+string h_language                = "english";
+string h_inputencoding           = "auto";
+string h_font_roman              = "default";
+string h_font_sans               = "default";
+string h_font_typewriter         = "default";
+string h_font_default_family     = "default";
+string h_font_sc                 = "false";
+string h_font_osf                = "false";
+string h_font_sf_scale           = "100";
+string h_font_tt_scale           = "100";
+string h_graphics                = "default";
+string h_paperfontsize           = "default";
+string h_spacing                 = "single";
+string h_papersize               = "default";
+string h_use_geometry            = "false";
+string h_use_amsmath             = "0";
+string h_cite_engine             = "basic";
+string h_use_bibtopic            = "false";
+string h_paperorientation        = "portrait";
+string h_secnumdepth             = "3";
+string h_tocdepth                = "3";
+string h_paragraph_separation    = "indent";
+string h_defskip                 = "medskip";
+string h_quotes_language         = "english";
+string h_papercolumns            = "1";
+string h_papersides              = string();
+string h_paperpagestyle          = "default";
+string h_tracking_changes        = "false";
+string h_output_changes          = "false";
 
 
 void handle_opt(vector<string> & opts, char const * const * what, string & target)
@@ -123,18 +123,39 @@ void handle_opt(vector<string> & opts, char const * const * what, string & targe
        if (opts.empty())
                return;
 
+       // the last language option is the document language (for babel and LyX)
+       // the last size option is the document font size
+       vector<string>::iterator it;
+       vector<string>::iterator position = opts.begin();
        for (; *what; ++what) {
-               vector<string>::iterator it = find(opts.begin(), opts.end(), *what);
-               // the last language option is the document language
+               it = find(opts.begin(), opts.end(), *what);
                if (it != opts.end()) {
-                       //cerr << "### found option '" << *what << "'\n";
-                       target = *what;
-                       opts.erase(it);
+                       documentclass_language = true;
+                       if (it >= position) {
+                               target = *what;
+                               position = it;
+                       }
                }
        }
 }
 
 
+void delete_opt(vector<string> & opts, char const * const * what)
+{
+       if (opts.empty())
+               return;
+
+       // remove found options from the list
+       // do this after handle_opt to avoid potential memory leaks and to be able
+       // to find in every case the last language option
+       vector<string>::iterator it;
+       for (; *what; ++what) {
+               it = find(opts.begin(), opts.end(), *what);
+               if (it != opts.end())
+                       opts.erase(it);
+       }
+}
+
 
 /*!
  * Split a package options string (keyval format) into a vector.
@@ -152,7 +173,7 @@ vector<string> split_options(string const & input)
        while (p.good()) {
                Token const & t = p.get_token();
                if (t.asInput() == ",") {
-                       options.push_back(option);
+                       options.push_back(trim(option));
                        option.erase();
                } else if (t.asInput() == "=") {
                        option += '=';
@@ -164,7 +185,7 @@ vector<string> split_options(string const & input)
        }
 
        if (!option.empty())
-               options.push_back(option);
+               options.push_back(trim(option));
 
        return options;
 }
@@ -193,31 +214,49 @@ void add_package(string const & name, vector<string> & options)
 }
 
 
+// Given is a string like "scaled=0.9", return 0.9 * 100
+string const scale_as_percentage(string const & scale)
+{
+       string::size_type pos = scale.find('=');
+       if (pos != string::npos) {
+               string value = scale.substr(pos + 1);
+               if (isStrDbl(value))
+                       return convert<string>(100 * convert<double>(value));
+       }
+       // If the input string didn't match our expectations.
+       // return the default value "100"
+       return "100";
+}
+
+
 void handle_package(string const & name, string const & opts)
 {
        vector<string> options = split_options(opts);
        add_package(name, options);
-       size_t pos;
        string scale;
 
-       // cerr << "handle_package: '" << name << "'\n";
-
        // roman fonts
        if (is_known(name, known_roman_fonts))
                h_font_roman = name;
-       if (name == "fourier")
+
+       if (name == "fourier") {
                h_font_roman = "utopia";
+               // when font uses real small capitals
+               if (opts == "expert")
+                       h_font_sc = "true";
+       }
        if (name == "mathpazo")
                h_font_roman = "palatino";
+
        if (name == "mathptmx")
                h_font_roman = "times";
+
        // sansserif fonts
        if (is_known(name, known_sans_fonts)) {
                h_font_sans = name;
                if (!opts.empty()) {
                        scale = opts;
-                       pos = scale.find(".", 0);
-                       h_font_sf_scale = scale.erase(0, pos + 1);
+                       h_font_sf_scale = scale_as_percentage(scale);
                }
        }
        // typewriter fonts
@@ -225,35 +264,93 @@ void handle_package(string const & name, string const & opts)
                h_font_typewriter = name;
                if (!opts.empty()) {
                        scale = opts;
-                       pos = scale.find(".", 0);
-                       h_font_tt_scale = scale.erase(0, pos + 1);
+                       h_font_tt_scale = scale_as_percentage(scale);
+               }
+       }
+       // font uses old-style figure
+       if (name == "eco")
+               h_font_osf = "true";
+
+       else if (name == "amsmath" || name == "amssymb")
+               h_use_amsmath = "1";
+
+       else if (name == "babel" && !opts.empty()) {
+               // check if more than one option was used - used later for inputenc
+               // in case inputenc is parsed before babel, set the encoding to auto
+               if (options.begin() != options.end() - 1) {
+                       one_language = false;
+                       h_inputencoding = "auto";
+               }
+               // only set the document language when there was not already one set
+               // via the documentclass options
+               // babel takes the the last language given in the documentclass options
+               // as document language. If there is no such language option, the last
+               // option of its \usepackage call is used.
+               if (documentclass_language == false) {
+                       handle_opt(options, known_languages, h_language);
+                       delete_opt(options, known_languages);
+                       if (is_known(h_language, known_french_languages))
+                               h_language = "french";
+                       else if (is_known(h_language, known_german_languages))
+                               h_language = "german";
+                       else if (is_known(h_language, known_ngerman_languages))
+                               h_language = "ngerman";
+                       else if (is_known(h_language, known_russian_languages))
+                               h_language = "russian";
+                       else if (is_known(h_language, known_ukrainian_languages))
+                               h_language = "ukrainian";
+                       h_quotes_language = h_language;
                }
        }
-
-       else if (name == "amssymb")
-               h_use_amsmath = "2";
-       else if (name == "esint")
-               h_use_esint = "2";
-       else if (name == "babel")
-               ; // ignore this
        else if (name == "fontenc")
                ; // ignore this
+
        else if (name == "inputenc") {
-               // only set when there are not more than one inputenc option
+               // only set when there is not more than one inputenc option
                // therefore check for the "," character
-               if ((pos = opts.find(",", 0)) == string::npos)
-                       h_inputencoding = opts;
+               // also only set when there is not more then one babel language option
+               if (opts.find(",") == string::npos && one_language == true)
+                       if (opts == "ascii")
+                               //change ascii to auto to be in the unicode range, see
+                               //http://bugzilla.lyx.org/show_bug.cgi?id=4719
+                               h_inputencoding = "auto";
+                       else
+                               h_inputencoding = opts;
                options.clear();
-       } else if (name == "makeidx")
+       }
+       else if (name == "makeidx")
                ; // ignore this
+
        else if (name == "verbatim")
                ; // ignore this
+
+       else if (name == "color")
+               // with the following command this package is only loaded when needed for
+               // undefined colors, since we only support the predefined colors
+               h_preamble << "\\@ifundefined{definecolor}\n {\\usepackage{color}}{}\n";
+
        else if (name == "graphicx")
                ; // ignore this
+
+       else if (name == "setspace")
+               ; // ignore this
+
        else if (is_known(name, known_languages)) {
-               h_language = name;
-               h_quotes_language = name;
-       } else if (name == "natbib") {
+               if (is_known(name, known_french_languages))
+                       h_language = "french";
+               else if (is_known(name, known_german_languages))
+                       h_language = "german";
+               else if (is_known(name, known_ngerman_languages))
+                       h_language = "ngerman";
+               else if (is_known(name, known_russian_languages))
+                       h_language = "russian";
+               else if (is_known(name, known_ukrainian_languages))
+                       h_language = "ukrainian";
+               else
+                       h_language = name;
+               h_quotes_language = h_language;
+       }
+       else if (name == "natbib") {
                h_cite_engine = "natbib_authoryear";
                vector<string>::iterator it =
                        find(options.begin(), options.end(), "authoryear");
@@ -266,15 +363,16 @@ void handle_package(string const & name, string const & opts)
                                options.erase(it);
                        }
                }
-       } else if (name == "jurabib") {
+       }
+       else if (name == "jurabib")
                h_cite_engine = "jurabib";
-       } else if (options.empty())
+
+       else if (options.empty())
                h_preamble << "\\usepackage{" << name << "}\n";
        else {
                h_preamble << "\\usepackage[" << opts << "]{" << name << "}\n";
                options.clear();
        }
-
        // We need to do something with the options...
        if (!options.empty())
                cerr << "Ignoring options '" << join(options, ",")
@@ -282,10 +380,11 @@ void handle_package(string const & name, string const & opts)
 }
 
 
+
 void end_preamble(ostream & os, TextClass const & /*textclass*/)
 {
-       os << "#LyX file created by  tex2lyx 0.1.5\n"
-          << "\\lyxformat 245\n"
+       os << "#LyX file created by tex2lyx " << PACKAGE_VERSION << "\n"
+          << "\\lyxformat 247\n"
           << "\\begin_document\n"
           << "\\begin_header\n"
           << "\\textclass " << h_textclass << "\n";
@@ -306,11 +405,9 @@ void end_preamble(ostream & os, TextClass const & /*textclass*/)
           << "\\graphics " << h_graphics << "\n"
           << "\\paperfontsize " << h_paperfontsize << "\n"
           << "\\spacing " << h_spacing << "\n"
-          << "\\use_hyperref " << h_use_hyperref << "\n"
           << "\\papersize " << h_papersize << "\n"
           << "\\use_geometry " << h_use_geometry << "\n"
           << "\\use_amsmath " << h_use_amsmath << "\n"
-          << "\\use_esint " << h_use_esint << "\n"
           << "\\cite_engine " << h_cite_engine << "\n"
           << "\\use_bibtopic " << h_use_bibtopic << "\n"
           << "\\paperorientation " << h_paperorientation << "\n"
@@ -332,8 +429,8 @@ void end_preamble(ostream & os, TextClass const & /*textclass*/)
 
 } // anonymous namespace
 
-
-TextClass const parse_preamble(Parser & p, ostream & os, string const & forceclass)
+void parse_preamble(Parser & p, ostream & os, 
+       string const & forceclass, TeX2LyXDocClass & tc)
 {
        // initialize fixed types
        special_columns['D'] = 3;
@@ -370,25 +467,29 @@ TextClass const parse_preamble(Parser & p, ostream & os, string const & forcecla
                          t.cat() == catEnd ||
                          t.cat() == catAlign ||
                          t.cat() == catParameter)
-               ;//h_preamble << t.character();
+               h_preamble << t.character();
 
                else if (t.cat() == catSpace || t.cat() == catNewline)
-                       ;//h_preamble << t.asInput();
+                       h_preamble << t.asInput();
 
                else if (t.cat() == catComment)
-                       ;//h_preamble << t.asInput();
+                       h_preamble << t.asInput();
 
                else if (t.cs() == "pagestyle")
                        h_paperpagestyle = p.verbatim_item();
 
-               else if (t.cs() == "makeatletter")
+               else if (t.cs() == "makeatletter") {
                        p.setCatCode('@', catLetter);
+               }
 
-               else if (t.cs() == "makeatother")
+               else if (t.cs() == "makeatother") {
                        p.setCatCode('@', catOther);
+               }
 
-               else if (t.cs() == "newcommand" || t.cs() == "renewcommand"
-                           || t.cs() == "providecommand") {
+               else if (t.cs() == "newcommand" 
+                        || t.cs() == "renewcommand"
+                        || t.cs() == "providecommand"
+                        || t.cs() == "newlyxcommand") {
                        bool star = false;
                        if (p.next_token().character() == '*') {
                                p.get_token();
@@ -420,16 +521,16 @@ TextClass const parse_preamble(Parser & p, ostream & os, string const & forcecla
                        if (   name != "\\noun"
                            && name != "\\tabularnewline"
                            && name != "\\LyX"
-                           && name != "\\lyxline"       
+                           && name != "\\lyxline"
                            && name != "\\lyxaddress"
                            && name != "\\lyxrightaddress"
                            && name != "\\lyxdot"
                            && name != "\\boldsymbol"
                            && name != "\\lyxarrow"
-                               && name != "\\rmdefault"
-                               && name != "\\sfdefault"
-                               && name != "\\ttdefault"
-                               && name != "\\familydefault") {
+                           && name != "\\rmdefault"
+                           && name != "\\sfdefault"
+                           && name != "\\ttdefault"
+                           && name != "\\familydefault") {
                                ostringstream ss;
                                ss << '\\' << t.cs();
                                if (star)
@@ -446,18 +547,30 @@ TextClass const parse_preamble(Parser & p, ostream & os, string const & forcecla
                                    << opts << "{" << body << "}";
 */
                        }
-
                }
 
                else if (t.cs() == "documentclass") {
-                       vector<string> opts;
-                       split(p.getArg('[', ']'), opts, ',');
+                       vector<string> opts = split_options(p.getArg('[', ']'));
                        handle_opt(opts, known_fontsizes, h_paperfontsize);
-                       handle_opt(opts, known_languages, h_language);
+                       delete_opt(opts, known_fontsizes);
                        // delete "pt" at the end
                        string::size_type i = h_paperfontsize.find("pt");
                        if (i != string::npos)
                                h_paperfontsize.erase(i);
+                       // to avoid that the babel options overwrite the documentclass options
+                       documentclass_language = false;
+                       handle_opt(opts, known_languages, h_language);
+                       delete_opt(opts, known_languages);
+                       if (is_known(h_language, known_french_languages))
+                               h_language = "french";
+                       else if (is_known(h_language, known_german_languages))
+                               h_language = "german";
+                       else if (is_known(h_language, known_ngerman_languages))
+                               h_language = "ngerman";
+                       else if (is_known(h_language, known_russian_languages))
+                               h_language = "russian";
+                       else if (is_known(h_language, known_ukrainian_languages))
+                               h_language = "ukrainian";
                        h_quotes_language = h_language;
                        h_options = join(opts, ",");
                        h_textclass = p.getArg('{', '}');
@@ -507,7 +620,6 @@ TextClass const parse_preamble(Parser & p, ostream & os, string const & forcecla
                        string opts = p.getOpt();
                        if (!opts.empty()) {
                                istringstream is(string(opts, 1));
-                               //cerr << "opt: " << is.str() << "\n";
                                is >> nargs;
                        }
                        special_columns[name[0]] = nargs;
@@ -531,15 +643,32 @@ TextClass const parse_preamble(Parser & p, ostream & os, string const & forcecla
                else if (t.cs() == "setlength") {
                        string const name = p.verbatim_item();
                        string const content = p.verbatim_item();
-                       // Is this correct?
-                       if (name == "parskip")
-                               h_paragraph_separation = "skip";
-                       else if (name == "parindent")
-                               h_paragraph_separation = "skip";
-                       else
+                       // the paragraphs are only not indented when \parindent is set to zero
+                       if (name == "\\parindent" && content != "") {
+                               if (content[0] == '0')
+                                       h_paragraph_separation = "skip";
+                       } else if (name == "\\parskip") {
+                               if (content == "\\smallskipamount")
+                                       h_defskip = "smallskip";
+                               else if (content == "\\medskipamount")
+                                       h_defskip = "medskip";
+                               else if (content == "\\bigskipamount")
+                                       h_defskip = "bigskip";
+                               else
+                                       h_defskip = content;
+                       } else
                                h_preamble << "\\setlength{" << name << "}{" << content << "}";
                }
 
+               else if (t.cs() =="onehalfspacing")
+                       h_spacing = "onehalf";
+
+               else if (t.cs() =="doublespacing")
+                       h_spacing = "double";
+
+               else if (t.cs() =="setstretch")
+                       h_spacing = "other " + p.verbatim_item();
+
                else if (t.cs() == "begin") {
                        string const name = p.getArg('{', '}');
                        if (name == "document")
@@ -566,25 +695,22 @@ TextClass const parse_preamble(Parser & p, ostream & os, string const & forcecla
        // Force textclass if the user wanted it
        if (!forceclass.empty())
                h_textclass = forceclass;
-       if (noweb_mode && !lyx::support::prefixIs(h_textclass, "literate-"))
+       if (noweb_mode && !prefixIs(h_textclass, "literate-"))
                h_textclass.insert(0, "literate-");
        FileName layoutfilename = libFileSearch("layouts", h_textclass, "layout");
        if (layoutfilename.empty()) {
                cerr << "Error: Could not find layout file for textclass \"" << h_textclass << "\"." << endl;
                exit(1);
        }
-       TextClass textclass;
-       textclass.read(layoutfilename);
+       tc.read(layoutfilename);
        if (h_papersides.empty()) {
                ostringstream ss;
-               ss << textclass.sides();
+               ss << tc.sides();
                h_papersides = ss.str();
        }
-       end_preamble(os, textclass);
-       return textclass;
+       end_preamble(os, tc);
 }
 
-
 // }])