]> git.lyx.org Git - lyx.git/blobdiff - src/tex2lyx/preamble.cpp
fix problem with non-recognized utf8 encoding that lead to broken .lyx
[lyx.git] / src / tex2lyx / preamble.cpp
index 0a8f577b36bfb8c9c39c1cbcffe5c2c72a75855e..62dd2abb491dee139071574ec1488661d72227bb 100644 (file)
@@ -25,6 +25,8 @@
 #include "support/filetools.h"
 #include "support/lstrings.h"
 
+#include <boost/regex.hpp>
+
 #include <algorithm>
 #include <iostream>
 #include <sstream>
@@ -34,6 +36,8 @@
 
 using namespace std;
 using namespace lyx::support;
+using boost::regex;
+using boost::smatch;
 
 namespace lyx {
 
@@ -83,7 +87,21 @@ const char * const known_typewriter_fonts[] = { "beramono", "cmtl", "cmtt",
 "courier", "lmtt", "luximono", "fourier", "lmodern", "mathpazo", "mathptmx",
 "newcent", 0};
 
-// some ugly stuff
+const char * const known_paper_sizes[] = { "a3paper", "b3paper", "a4paper",
+"b4paper", "a5paper", "b5paper", "executivepaper", "legalpaper",
+"letterpaper", 0};
+
+const char * const known_class_paper_sizes[] = { "a4paper", "a5paper",
+"executivepaper", "legalpaper", "letterpaper", 0};
+
+const char * const known_paper_margins[] = { "lmargin", "tmargin", "rmargin", 
+"bmargin", "headheight", "headsep", "footskip", "columnsep", 0};
+
+const char * const known_coded_paper_margins[] = { "leftmargin", "topmargin",
+"rightmargin", "bottommargin", "headheight", "headsep", "footskip",
+"columnsep", 0};
+
+// default settings
 ostringstream h_preamble;
 string h_textclass               = "article";
 string h_options                 = string();
@@ -116,6 +134,7 @@ string h_papersides              = string();
 string h_paperpagestyle          = "default";
 string h_tracking_changes        = "false";
 string h_output_changes          = "false";
+string h_margins                 = "";
 
 
 void handle_opt(vector<string> & opts, char const * const * what, string & target)
@@ -229,7 +248,8 @@ string const scale_as_percentage(string const & scale)
 }
 
 
-void handle_package(string const & name, string const & opts)
+void handle_package(string const & name, string const & opts,
+                   bool in_lyx_preamble)
 {
        vector<string> options = split_options(opts);
        add_package(name, options);
@@ -309,15 +329,17 @@ void handle_package(string const & name, string const & opts)
                // only set when there is not more than one inputenc option
                // therefore check for the "," character
                // also only set when there is not more then one babel language option
-               if (opts.find(",") == string::npos && one_language == true)
+               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
+                       else if (!opts.empty())
                                h_inputencoding = opts;
+               }
                options.clear();
        }
+
        else if (name == "makeidx")
                ; // ignore this
 
@@ -335,6 +357,10 @@ void handle_package(string const & name, string const & opts)
        else if (name == "setspace")
                ; // ignore this
 
+       else if (name == "geometry")
+               ; // Ignore this, the geometry settings are made by the \geometry
+                 // command. This command is handled below.
+
        else if (is_known(name, known_languages)) {
                if (is_known(name, known_french_languages))
                        h_language = "french";
@@ -366,13 +392,16 @@ void handle_package(string const & name, string const & opts)
        }
        else if (name == "jurabib")
                h_cite_engine = "jurabib";
-
-       else if (options.empty())
-               h_preamble << "\\usepackage{" << name << "}\n";
-       else {
-               h_preamble << "\\usepackage[" << opts << "]{" << name << "}\n";
-               options.clear();
+       else if (!in_lyx_preamble) {
+               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, ",")
@@ -411,6 +440,7 @@ void end_preamble(ostream & os, TextClass const & /*textclass*/)
           << "\\cite_engine " << h_cite_engine << "\n"
           << "\\use_bibtopic " << h_use_bibtopic << "\n"
           << "\\paperorientation " << h_paperorientation << "\n"
+          << h_margins
           << "\\secnumdepth " << h_secnumdepth << "\n"
           << "\\tocdepth " << h_tocdepth << "\n"
           << "\\paragraph_separation " << h_paragraph_separation << "\n"
@@ -435,6 +465,8 @@ void parse_preamble(Parser & p, ostream & os,
        // initialize fixed types
        special_columns['D'] = 3;
        bool is_full_document = false;
+       bool is_lyx_file = false;
+       bool in_lyx_preamble = true;
 
        // determine whether this is a full document or a fragment for inclusion
        while (p.good()) {
@@ -457,39 +489,61 @@ void parse_preamble(Parser & p, ostream & os,
                //
                // cat codes
                //
-               if (t.cat() == catLetter ||
-                         t.cat() == catSuper ||
-                         t.cat() == catSub ||
-                         t.cat() == catOther ||
-                         t.cat() == catMath ||
-                         t.cat() == catActive ||
-                         t.cat() == catBegin ||
-                         t.cat() == catEnd ||
-                         t.cat() == catAlign ||
-                         t.cat() == catParameter)
-               h_preamble << t.character();
-
-               else if (t.cat() == catSpace || t.cat() == catNewline)
+               if (!in_lyx_preamble &&
+                   (t.cat() == catLetter ||
+                    t.cat() == catSuper ||
+                    t.cat() == catSub ||
+                    t.cat() == catOther ||
+                    t.cat() == catMath ||
+                    t.cat() == catActive ||
+                    t.cat() == catBegin ||
+                    t.cat() == catEnd ||
+                    t.cat() == catAlign ||
+                    t.cat() == catParameter))
+                       h_preamble << t.character();
+
+               else if (!in_lyx_preamble && 
+                        (t.cat() == catSpace || t.cat() == catNewline))
                        h_preamble << t.asInput();
 
-               else if (t.cat() == catComment)
-                       h_preamble << t.asInput();
+               else if (t.cat() == catComment) {
+                       // regex to parse comments
+                       static regex const islyxfile("%% LyX .* created this file");
+                       static regex const usercommands("User specified LaTeX commands");
+                       
+                       string const comment = t.asInput();
+                       cerr << "Seen comment: " << comment << std::endl;
+                       smatch sub;
+                       if (regex_search(comment, sub, islyxfile))
+                               is_lyx_file = true;
+                       else if (is_lyx_file
+                                && regex_search(comment, sub, usercommands))
+                               in_lyx_preamble = false;
+                       else if (!in_lyx_preamble)
+                               h_preamble << t.asInput();
+                       cerr << "lyx_file: " << is_lyx_file << ", lyx_preamble "
+                            << in_lyx_preamble << std::endl;
+               }
 
                else if (t.cs() == "pagestyle")
                        h_paperpagestyle = p.verbatim_item();
 
                else if (t.cs() == "makeatletter") {
+                       if (!is_lyx_file || !in_lyx_preamble
+                           || p.getCatCode('@') != catLetter)
+                               h_preamble << "\\makeatletter";
                        p.setCatCode('@', catLetter);
                }
 
                else if (t.cs() == "makeatother") {
+                       if (!is_lyx_file || !in_lyx_preamble
+                           || p.getCatCode('@') != catOther)
+                               h_preamble << "\\makeatother";
                        p.setCatCode('@', catOther);
                }
 
-               else if (t.cs() == "newcommand" 
-                        || t.cs() == "renewcommand"
-                        || t.cs() == "providecommand"
-                        || t.cs() == "newlyxcommand") {
+               else if (t.cs() == "newcommand" || t.cs() == "renewcommand"
+                           || t.cs() == "providecommand") {
                        bool star = false;
                        if (p.next_token().character() == '*') {
                                p.get_token();
@@ -518,19 +572,7 @@ void parse_preamble(Parser & p, ostream & os,
                                h_font_default_family = family.erase(0,1);
                        }
                        // only non-lyxspecific stuff
-                       if (   name != "\\noun"
-                           && name != "\\tabularnewline"
-                           && name != "\\LyX"
-                           && name != "\\lyxline"
-                           && name != "\\lyxaddress"
-                           && name != "\\lyxrightaddress"
-                           && name != "\\lyxdot"
-                           && name != "\\boldsymbol"
-                           && name != "\\lyxarrow"
-                           && name != "\\rmdefault"
-                           && name != "\\sfdefault"
-                           && name != "\\ttdefault"
-                           && name != "\\familydefault") {
+                       if (!in_lyx_preamble) {
                                ostringstream ss;
                                ss << '\\' << t.cs();
                                if (star)
@@ -550,6 +592,7 @@ void parse_preamble(Parser & p, ostream & os,
                }
 
                else if (t.cs() == "documentclass") {
+                       vector<string>::iterator it;
                        vector<string> opts = split_options(p.getArg('[', ']'));
                        handle_opt(opts, known_fontsizes, h_paperfontsize);
                        delete_opt(opts, known_fontsizes);
@@ -572,6 +615,39 @@ void parse_preamble(Parser & p, ostream & os,
                        else if (is_known(h_language, known_ukrainian_languages))
                                h_language = "ukrainian";
                        h_quotes_language = h_language;
+                       // paper orientation
+                       if ((it = find(opts.begin(), opts.end(), "landscape")) != opts.end()) {
+                               h_paperorientation = "landscape";
+                               opts.erase(it);
+                       }
+                       // paper sides
+                       if ((it = find(opts.begin(), opts.end(), "oneside"))
+                                != opts.end()) {
+                               h_papersides = "1";
+                               opts.erase(it);
+                       }
+                       if ((it = find(opts.begin(), opts.end(), "twoside"))
+                                != opts.end()) {
+                               h_papersides = "2";
+                               opts.erase(it);
+                       }
+                       // paper columns
+                       if ((it = find(opts.begin(), opts.end(), "onecolumn"))
+                                != opts.end()) {
+                               h_papercolumns = "1";
+                               opts.erase(it);
+                       }
+                       if ((it = find(opts.begin(), opts.end(), "twocolumn"))
+                                != opts.end()) {
+                               h_papercolumns = "2";
+                               opts.erase(it);
+                       }
+                       // paper sizes
+                       // some size options are know to any document classes, other sizes
+                       // are handled by the \geometry command of the geometry package
+                       handle_opt(opts, known_class_paper_sizes, h_papersize);
+                       delete_opt(opts, known_class_paper_sizes);
+                       // the remaining options
                        h_options = join(opts, ",");
                        h_textclass = p.getArg('{', '}');
                }
@@ -585,12 +661,17 @@ void parse_preamble(Parser & p, ostream & os,
                                vector<string>::const_iterator it  = vecnames.begin();
                                vector<string>::const_iterator end = vecnames.end();
                                for (; it != end; ++it)
-                                       handle_package(trim(*it), string());
+                                       handle_package(trim(*it), string(), 
+                                                      in_lyx_preamble);
                        } else {
-                               handle_package(name, options);
+                               handle_package(name, options, in_lyx_preamble);
                        }
                }
 
+               else if (t.cs() == "inputencoding") {
+                       h_inputencoding = p.getArg('{','}');
+               }
+
                else if (t.cs() == "newenvironment") {
                        string const name = p.getArg('{', '}');
                        ostringstream ss;
@@ -599,9 +680,7 @@ void parse_preamble(Parser & p, ostream & os,
                        ss << p.getOpt();
                        ss << '{' << p.verbatim_item() << '}';
                        ss << '{' << p.verbatim_item() << '}';
-                       if (name != "lyxcode" && name != "lyxlist" &&
-                           name != "lyxrightadress" &&
-                           name != "lyxaddress" && name != "lyxgreyedout")
+                       if (!in_lyx_preamble)
                                h_preamble << ss.str();
                }
 
@@ -609,8 +688,9 @@ void parse_preamble(Parser & p, ostream & os,
                        string name = p.get_token().cs();
                        while (p.next_token().cat() != catBegin)
                                name += p.get_token().asString();
-                       h_preamble << "\\def\\" << name << '{'
-                                  << p.verbatim_item() << "}";
+                       if (!in_lyx_preamble)
+                               h_preamble << "\\def\\" << name << '{'
+                                          << p.verbatim_item() << "}";
                }
 
                else if (t.cs() == "newcolumntype") {
@@ -660,13 +740,13 @@ void parse_preamble(Parser & p, ostream & os,
                                h_preamble << "\\setlength{" << name << "}{" << content << "}";
                }
 
-               else if (t.cs() =="onehalfspacing")
+               else if (t.cs() == "onehalfspacing")
                        h_spacing = "onehalf";
 
-               else if (t.cs() =="doublespacing")
+               else if (t.cs() == "doublespacing")
                        h_spacing = "double";
 
-               else if (t.cs() =="setstretch")
+               else if (t.cs() == "setstretch")
                        h_spacing = "other " + p.verbatim_item();
 
                else if (t.cs() == "begin") {
@@ -676,6 +756,35 @@ void parse_preamble(Parser & p, ostream & os,
                        h_preamble << "\\begin{" << name << "}";
                }
 
+               else if (t.cs() == "geometry") {
+                       h_use_geometry = "true";
+                       vector<string> opts = split_options(p.getArg('{', '}'));
+                       vector<string>::iterator it;
+                       // paper orientation
+                       if ((it = find(opts.begin(), opts.end(), "landscape")) != opts.end()) {
+                               h_paperorientation = "landscape";
+                               opts.erase(it);
+                       }
+                       // paper size
+                       handle_opt(opts, known_paper_sizes, h_papersize);
+                       delete_opt(opts, known_paper_sizes);
+                       // page margins
+                       char const * const * margin = known_paper_margins;
+                       int k = -1;
+                       for (; *margin; ++margin) {
+                               k += 1;
+                               // search for the "=" in e.g. "lmargin=2cm" to get the value
+                               for(size_t i = 0; i != opts.size(); i++) {
+                                       if (opts.at(i).find(*margin) != string::npos) {
+                                               string::size_type pos = opts.at(i).find("=");
+                                               string value = opts.at(i).substr(pos + 1);
+                                               string name = known_coded_paper_margins[k];
+                                               h_margins += "\\" + name + " " + value + "\n";
+                                       }
+                               }
+                       }
+               }
+
                else if (t.cs() == "jurabibsetup") {
                        vector<string> jurabibsetup =
                                split_options(p.getArg('{', '}'));
@@ -687,7 +796,7 @@ void parse_preamble(Parser & p, ostream & os,
                        }
                }
 
-               else if (!t.cs().empty())
+               else if (!t.cs().empty() && !in_lyx_preamble)
                        h_preamble << '\\' << t.cs();
        }
        p.skip_spaces();