]> git.lyx.org Git - lyx.git/blobdiff - src/lyxlex.C
Added html export for LinuxDoc and DocBook. LinuxDoc import now available in file...
[lyx.git] / src / lyxlex.C
index 213abecfbfb038cce4b6eb97363f3e545f1a68ff..c35246550517c7163083d821cdefe58523871399 100644 (file)
 //  It can be used for simple syntax parsers, like lyxrc,
 //  texclass and others to come.   [asierra30/03/96]
 //
-//   (C) 1996 Lyx Team.
+//   Copyright 1996 Lyx Team.
 
 #include <config.h>
-//#include "definitions.h"
 
-#include <stdlib.h>
-#include <string.h>
+#include <algorithm>
+#include <cstdlib>
 
 #ifdef __GNUG__
 #pragma implementation "lyxlex.h"
 #endif
 
 #include "lyxlex.h"
-#include "error.h"
-#include "filetools.h"
+#include "debug.h"
+#include "support/filetools.h"
+#include "support/lyxalgo.h"
 
-//     $Id: lyxlex.C,v 1.1 1999/09/27 18:44:37 larsbj Exp $    
+using std::ios;
+using std::lower_bound;
+using std::sort;
 
-#if !defined(lint) && !defined(WITH_WARNINGS)
-static char vcid[] = "$Id: lyxlex.C,v 1.1 1999/09/27 18:44:37 larsbj Exp $";
-#endif /* lint */
+
+// namespace {
+struct compare_tags {
+       // used by lower_bound
+       inline
+       int operator()(keyword_item const & a, char const * const tag) const {
+               return compare_no_case(a.tag, tag) < 0;
+       }
+       // used by sorted and sort
+       inline
+       int operator()(keyword_item const & a, keyword_item const & b) const {
+               return compare_no_case(a.tag, b.tag) < 0;
+       }
+};
+// } // end of anon namespace
 
 
-LyXLex::LyXLex(keyword_item* tab, int num)
-       : table(tab), no_items(num)
+LyXLex::LyXLex(keyword_item * tab, int num)
+       : is(&fb__), table(tab), no_items(num)
 {
-       file = NULL;
-       owns_file = false;
-       status = 0;
-       pushed = NULL;
+       status = 0; 
+       pushed = 0;
+       // Check if the table is sorted and if not, sort it.
+       if (table && !sorted(table, table + no_items, compare_tags())) {
+               lyxerr << "The table passed to LyXLex is not sorted!!\n"
+                      << "Tell the developers to fix it!" << endl;
+               // We sort it anyway to avoid problems.
+               lyxerr << "\nUnsorted:\n";
+               printTable(lyxerr);
+               
+               sort(table, table + no_items,
+                    compare_tags());
+               lyxerr << "\nSorted:\n";
+               printTable(lyxerr);
+       }
 }
 
 
-void LyXLex::pushTable(keyword_item* tab, int num)
+void LyXLex::pushTable(keyword_item * tab, int num)
 {
-       pushed_table *tmppu = new pushed_table;
+       pushed_table * tmppu = new pushed_table;
        tmppu->next = pushed;
        tmppu->table_elem = table;
        tmppu->table_siz = no_items;
        pushed = tmppu;
        table = tab;
        no_items = num;
+       // Check if the table is sorted and if not, sort it.
+       if (table && !sorted(table, table + no_items, compare_tags())) {
+               lyxerr << "The table passed to LyXLex is not sorted!!\n"
+                      << "Tell the developers to fix it!" << endl;
+               // We sort it anyway to avoid problems.
+               lyxerr << "\nUnsorted:\n";
+               printTable(lyxerr);
+               
+               sort(table, table + no_items, compare_tags());
+               lyxerr << "\nSorted:\n";
+               printTable(lyxerr);
+       }
 }
 
 
 void LyXLex::popTable()
 {
-       if (pushed == NULL)
-               lyxerr.print("LyXLex error: nothing to pop!");
+       if (pushed == 0)
+               lyxerr << "LyXLex error: nothing to pop!" << endl;
 
-       pushed_table *tmp;
+       pushed_table * tmp;
        tmp = pushed;
        table = tmp->table_elem;
        no_items = tmp->table_siz;
-       tmp->table_elem = NULL;
+       tmp->table_elem = 0;
        pushed = tmp->next;
        delete tmp;
 }
 
 
-void LyXLex::printTable()
+void LyXLex::printTable(ostream & os)
 {
-       lyxerr.print(LString("\nNumber of tags: ")+no_items);
-       for(int i=0; i<no_items; i++)
-               lyxerr.print(LString("table[")+i+
-                              "]:  tag: `"+table[i].tag+
-                              "'  code:"+(long)table[i].code);
-       lyxerr.print(LString());
+       os << "\nNumber of tags: " << no_items << '\n';
+       for(int i= 0; i < no_items; ++i)
+               os << "table[" << i
+                  << "]:  tag: `" << table[i].tag
+                  << "'  code:" << table[i].code << '\n';
+       os.flush();
 }
 
 
-void LyXLex::printError(LString const & message)
+void LyXLex::printError(string const & message) const
 {
-       LString tmpmsg = message;
-       tmpmsg.subst("$$Token",GetString());
-       lyxerr.print("LyX: "+tmpmsg+" [around line "+lineno+" of file "
-                     +MakeDisplayPath(name)+']');
+       string tmpmsg = subst(message, "$$Token", GetString());
+       lyxerr << "LyX: " << tmpmsg << " [around line " << lineno
+              << " of file " << MakeDisplayPath(name) << ']' << endl;
 }
 
 
-bool LyXLex::setFile(LString const & filename)
+bool LyXLex::setFile(string const & filename)
 {
-        if (file) 
-               lyxerr.print("Error in LyXLex::setFile: file already set.");
-       file = fopen(filename.c_str(), "r");
+       if (fb__.is_open())
+               lyxerr << "Error in LyXLex::setFile: "
+                       "file or stream already set." << endl;
+       fb__.open(filename.c_str(), ios::in);
+       is.rdbuf(&fb__);
        name = filename;
-       owns_file = true;
        lineno = 0;
-       return (file ? true : false);
+       return fb__.is_open() && is.good();
 }
 
 
-void LyXLex::setFile(FILE *f)
+void LyXLex::setStream(istream & i)
 {
-        if (file) 
-               lyxerr.print("Error in LyXLex::setFile: file already set.");
-       file = f;
-       owns_file = false;
-       lineno = 0; // this is bogus if the file already has been read from
+       if (fb__.is_open() || is.rdbuf()->in_avail())
+               lyxerr << "Error in LyXLex::setStream: "
+                       "file or stream already set." << endl;
+       is.rdbuf(i.rdbuf());
+       lineno = 0;
 }
 
 
 int LyXLex::lex()
 {
        //NOTE: possible bug.
-   if (next() && status==LEX_TOKEN)
-       return search_kw(buff);
-   else
-       return status;
+       if (next() && status == LEX_TOKEN)
+               return search_kw(buff);
+       else
+               return status;
 }
 
 
-int LyXLex::GetInteger()
+int LyXLex::GetInteger() const
 {
-   if (buff[0]>' ')   
-       return atoi(buff);
-   else {
-       printError("Bad integer `$$Token'");
-       return -1;
-   }
+       if (buff[0] > ' ')   
+               return atoi(buff);
+       else {
+               printError("Bad integer `$$Token'");
+               return -1;
+       }
 }
 
 
-float LyXLex::GetFloat()
+float LyXLex::GetFloat() const
 {
-   if (buff[0]>' ')   
-       return (float)strtod(buff, (char**)NULL);
+   if (buff[0] > ' ')   
+       return atof(buff);
    else {
        printError("Bad float `$$Token'");
        return -1;
@@ -136,18 +173,18 @@ float LyXLex::GetFloat()
 }
 
 
-LString LyXLex::GetString() const
+string LyXLex::GetString() const
 {
-       return LString(buff);
+       return string(buff);
 }
 
 
 // I would prefer to give a tag number instead of an explicit token
 // here, but it is not possible because Buffer::readLyXformat2 uses
 // explicit tokens (JMarc) 
-LString LyXLex::getLongString(LString const &endtoken)
+string LyXLex::getLongString(string const & endtoken)
 {
-       LString str, prefix;
+       string str, prefix;
        bool firstline = true;
 
        while (IsOK()) {
@@ -155,30 +192,30 @@ LString LyXLex::getLongString(LString const &endtoken)
                        // blank line in the file being read
                        continue;
                
-               LString const token = GetString().strip().frontStrip();
+               string const token = frontStrip(strip(GetString()), " \t");
                
-               lyxerr.debug("LongString: `"+GetString()+'\'', Error::LEX_PARSER);
+               lyxerr[Debug::PARSER] << "LongString: `"
+                                     << GetString() << '\'' << endl;
 
                // We do a case independent comparison, like search_kw
                // does.
-                if (strcasecmp(token.c_str(), endtoken.c_str()) != 0) {
-                       LString tmpstr = GetString();
+                if (compare_no_case(token, endtoken) != 0) {
+                       string tmpstr = GetString();
                        if (firstline) {
-                               int i = 0;
+                               unsigned int i = 0;
                                while(i < tmpstr.length()
                                      && tmpstr[i] == ' ') {
-                                       i++;
+                                       ++i;
                                        prefix += ' ';
                                }
                                firstline = false;
-                               lyxerr.debug("Prefix = `"+prefix+'\'',
-                                             Error::LEX_PARSER); 
+                               lyxerr[Debug::PARSER] << "Prefix = `" << prefix
+                                                     << '\'' << endl;
                        } 
 
                        if (!prefix.empty() 
-                           && tmpstr.prefixIs(prefix.c_str())) {
-                               tmpstr.substring(prefix.length(), 
-                                                tmpstr.length()-1);
+                           && prefixIs(tmpstr, prefix.c_str())) {
+                               tmpstr.erase(0, prefix.length() - 1);
                        }
                        str += tmpstr + '\n';
                 }
@@ -192,34 +229,37 @@ LString LyXLex::getLongString(LString const &endtoken)
 }
 
 
-bool LyXLex::GetBool()
+bool LyXLex::GetBool() const
 {
-   if (strcmp(buff, "true") == 0)
-       return true;
-   else if (strcmp(buff, "false") != 0)
-       printError("Bad boolean `$$Token'. Use \"false\" or \"true\"");
-   return false;
+       if (compare(buff, "true") == 0)
+               return true;
+       else if (compare(buff, "false") != 0)
+               printError("Bad boolean `$$Token'. Use \"false\" or \"true\"");
+       return false;
 }
 
 
 bool LyXLex::EatLine()
 {
-       int i=0;
-       int c = '\0'; // getc() returns an int
-
-       while (!feof(file) && c!='\n' && i!=(LEX_MAX_BUFF-1)) {
-               c = getc(file);
+       int i = 0;
+       unsigned char c = '\0';
+       char cc = 0;
+       while(is && c != '\n' && i != (LEX_MAX_BUFF - 1)) {
+               is.get(cc);
+               c = cc;
+               lyxerr[Debug::LYXLEX] << "LyXLex::EatLine read char: `"
+                                     << c << "'" << endl;
                if (c != '\r')
                        buff[i++] = c;
        }
-       if (i==(LEX_MAX_BUFF-1) && c !='\n') {
-               printError("Line too long");
+       if (i == (LEX_MAX_BUFF - 1) && c != '\n') {
+               printError("Line too long");
                c = '\n'; // Pretend we had an end of line
-               lineno--; // but don't increase line counter (netto effect)
-               i++; // and preserve last character read.
+               --lineno; // but don't increase line counter (netto effect)
+               ++i; // and preserve last character read.
        }
-       if (c=='\n') {
-               lineno++;
+       if (c == '\n') {
+               ++lineno;
                buff[--i] = '\0'; // i can never be 0 here, so no danger
                status = LEX_DATA;
                return true;
@@ -232,70 +272,52 @@ bool LyXLex::EatLine()
 
 int LyXLex::search_kw(char const * const tag) const
 {
-       int m, k=0 , l= 0, r=no_items;
-
-       while (l < r) {
-               m = (l+r)/2;
-
-               if (lyxerr.debugging(Error::LEX_PARSER)) {
-                       LString my_l;
-                       my_l+="LyXLex::search_kw: elem " ;
-                       my_l+= m; 
-                       my_l+=" tag "; 
-                       my_l+=table[m].tag;
-                       my_l+=" search tag ";
-                       my_l+= tag;
-                       lyxerr.print(my_l);
-               }
-
-               if (table[m].tag)
-                       k = strcasecmp(table[m].tag, tag);
-               if (k==0)
-                       return table[m].code;
-               else
-                       if (k<0) l = m+1; else r = m;
-       }
-       return -1;
+       keyword_item * res =
+               lower_bound(table, table + no_items, tag, compare_tags());
+       if (res != table + no_items && !compare_no_case(res->tag, tag))
+               return res->code;
+       return LEX_UNDEF;
 }
 
 
 bool LyXLex::next(bool esc)
 {
-
        if (!esc) {
-               int c; // getc() returns an int
-               int i;
-               
-               
+               unsigned char c = 0; // getc() returns an int
+               char cc = 0;
                status = 0;
-               while (!feof(file) && !status) { 
-                       c = getc(file);
-                       if (c=='#') {
+               while (is && !status) {
+                       is.get(cc);
+                       c = cc;
+                       if (c == '#') {
                                // Read rest of line (fast :-)
-                               fgets(buff, sizeof(buff), file);
-                               lineno++;
+                               is.getline(buff, sizeof(buff));
+                               lyxerr[Debug::LYXLEX] << "Comment read: `" << c
+                                                     << buff << "'" << endl;
+                               ++lineno;
                                continue;
                        }
                        
-                       if (c=='\"') {
-                               i = -1;
+                       if (c == '\"') {
+                               int i = -1;
                                do {
-                                       c = getc(file);
+                                       is.get(cc);
+                                       c = cc;
                                        if (c != '\r')
                                                buff[++i] = c;
-                               } while (c!='\"' && c!='\n' && !feof(file) &&
-                                        i!=(LEX_MAX_BUFF-2));
+                               } while (c != '\"' && c != '\n' && is &&
+                                        i != (LEX_MAX_BUFF - 2));
                                
-                               if (i==(LEX_MAX_BUFF-2)) {
+                               if (i == (LEX_MAX_BUFF - 2)) {
                                        printError("Line too long");
                                        c = '\"'; // Pretend we got a "
-                                       i++;
+                                       ++i;
                                }
                                
-                               if (c!='\"') {
+                               if (c != '\"') {
                                        printError("Missing quote");
-                                       if (c=='\n')
-                                               lineno++;
+                                       if (c == '\n')
+                                               ++lineno;
                                }
                                
                                buff[i] = '\0';
@@ -303,65 +325,73 @@ bool LyXLex::next(bool esc)
                                break; 
                        }
                        
-                       if (c==',')
+                       if (c == ',')
                                continue;              /* Skip ','s */
-                       
-                       if (c > ' ' && !feof(file))  {
-                               i = 0;
+
+                       // using relational operators with chars other
+                       // than == and != is not safe. And if it is done
+                       // the type _have_ to be unsigned. It usually a
+                       // lot better to use the functions from cctype
+                       if (c > ' ' && is)  {
+                               int i = 0;
                                do {
                                        buff[i++] = c;
-                                       c = getc(file);
-                               } while (c > ' ' && c != ',' && !feof(file) &&
-                                        (i != LEX_MAX_BUFF-1) );
-                               if (i == LEX_MAX_BUFF-1) {
+                                       is.get(cc);
+                                       c = cc;
+                               } while (c > ' ' && c != ',' && is
+                                        && (i != LEX_MAX_BUFF - 1) );
+                               if (i == LEX_MAX_BUFF - 1) {
                                        printError("Line too long");
                                }
                                buff[i] = '\0';
                                status = LEX_TOKEN;
                        }
                        
-                       if (c=='\r' && !feof(file)) {
+                       if (c == '\r' && is) {
                                // The Windows support has lead to the
                                // possibility of "\r\n" at the end of
                                // a line.  This will stop LyX choking
                                // when it expected to find a '\n'
-                               c = getc(file);
+                               is.get(cc);
+                               c = cc;
                        }
 
-                       if (c=='\n')
-                               lineno++;
+                       if (c == '\n')
+                               ++lineno;
                        
                }
                if (status) return true;
                
-               status = (feof(file)) ? LEX_FEOF: LEX_UNDEF;
+               status = is.eof() ? LEX_FEOF: LEX_UNDEF;
                buff[0] = '\0';
                return false;
        } else {
-               int c; // getc() returns an int
-               int i;
-               
+               unsigned char c = 0; // getc() returns an int
+               char cc = 0;
                
                status = 0;
-               while (!feof(file) && !status) { 
-                       c = getc(file);
+               while (is && !status) {
+                       is.get(cc);
+                       c = cc;
 
                        // skip ','s
-                       if (c==',') continue;
+                       if (c == ',') continue;
                        
-                       if (c=='\\') {
+                       if (c == '\\') {
                                // escape
-                               i = 0;
+                               int i = 0;
                                do {
                                        if (c == '\\') {
                                                // escape the next char
-                                               c = getc(file);
+                                               is.get(cc);
+                                               c = cc;
                                        }
                                        buff[i++] = c;
-                                       c = getc(file);
-                               } while (c > ' ' && c != ',' && !feof(file) &&
-                                        (i != LEX_MAX_BUFF-1) );
-                               if (i == LEX_MAX_BUFF-1) {
+                                       is.get(cc);
+                                       c = cc;
+                               } while (c > ' ' && c != ',' && is
+                                        && (i != LEX_MAX_BUFF - 1) );
+                               if (i == LEX_MAX_BUFF - 1) {
                                        printError("Line too long");
                                }
                                buff[i] = '\0';
@@ -369,42 +399,46 @@ bool LyXLex::next(bool esc)
                                continue;
                        }
                        
-                       if (c=='#') {
+                       if (c == '#') {
                                // Read rest of line (fast :-)
-                               fgets(buff, sizeof(buff), file);
-                               lineno++;
+                               is.getline(buff, sizeof(buff));
+                               lyxerr[Debug::LYXLEX] << "Comment read: `" << c
+                                                     << buff << "'" << endl;
+                               ++lineno;
                                continue;
                        }
 
                        // string
-                       if (c=='\"') {
-                               i = -1;
+                       if (c == '\"') {
+                               int i = -1;
                                bool escaped = false;
                                do {
                                        escaped = false;
-                                       c = getc(file);
+                                       is.get(cc);
+                                       c = cc;
                                        if (c == '\r') continue;
                                        if (c == '\\') {
                                                // escape the next char
-                                               c = getc(file);
+                                               is.get(cc);
+                                               c = cc;
                                                escaped = true;
                                        }
                                        buff[++i] = c;
                                
                                        if (!escaped && c == '\"') break;
-                               } while (c!='\n' && !feof(file) &&
-                                        i!=(LEX_MAX_BUFF-2));
+                               } while (c != '\n' && is &&
+                                        i != (LEX_MAX_BUFF - 2));
                                
-                               if (i==(LEX_MAX_BUFF-2)) {
+                               if (i == (LEX_MAX_BUFF - 2)) {
                                        printError("Line too long");
                                        c = '\"'; // Pretend we got a "
-                                       i++;
+                                       ++i;
                                }
                                
-                               if (c!='\"') {
+                               if (c != '\"') {
                                        printError("Missing quote");
-                                       if (c=='\n')
-                                               lineno++;
+                                       if (c == '\n')
+                                               ++lineno;
                                }
                                
                                buff[i] = '\0';
@@ -412,93 +446,95 @@ bool LyXLex::next(bool esc)
                                break; 
                        }
                        
-                       if (c > ' ' && !feof(file))  {
-                               i = 0;
+                       if (c > ' ' && is) {
+                               int i = 0;
                                do {
                                        if (c == '\\') {
                                                // escape the next char
-                                               c = getc(file);
+                                               is.get(cc);
+                                               c = cc;
                                                //escaped = true;
                                        }
                                        buff[i++] = c;
-                                       c = getc(file);
-                               } while (c > ' ' && c != ',' && !feof(file) &&
-                                        (i != LEX_MAX_BUFF-1) );
+                                       is.get(cc);
+                                       c = cc;
+                               } while (c > ' ' && c != ',' && is
+                                        && (i != LEX_MAX_BUFF-1) );
                                if (i == LEX_MAX_BUFF-1) {
                                        printError("Line too long");
                                }
                                buff[i] = '\0';
                                status = LEX_TOKEN;
                        }
-
                        // new line
-                       if (c=='\n')
-                               lineno++;
+                       if (c == '\n')
+                               ++lineno;
                }
                
                if (status) return true;
                
-               status = (feof(file)) ? LEX_FEOF: LEX_UNDEF;
+               status = is.eof() ? LEX_FEOF : LEX_UNDEF;
                buff[0] = '\0';
-               return false;   
+               return false;
        }
 }
 
 
 bool LyXLex::nextToken()
 {
-       int c; // getc() returns an int
-       int i;
-        
         status = 0;
-       while (!feof(file) && !status) { 
-               c = getc(file);
-          
-               if (c >= ' ' && !feof(file))  {
-                       i = 0;
+       while (is && !status) {
+               unsigned char c = 0;
+               char cc = 0;
+               is.get(cc);
+               c = cc;
+               if (c >= ' ' && is) {
+                       int i = 0;
                        if (c == '\\') { // first char == '\\'
                                do {
                                        buff[i++] = c;
-                                       c = getc(file);
-                               } while (c > ' ' && c != '\\' && !feof(file) &&
-                                        i != (LEX_MAX_BUFF-1));
+                                       is.get(cc);
+                                       c = cc;
+                               } while (c > ' ' && c != '\\' && is
+                                        && i != (LEX_MAX_BUFF-1));
                        } else {
                                do {
                                        buff[i++] = c;
-                                       c = getc(file);
-                               } while (c >= ' ' && c != '\\' && !feof(file)
+                                       is.get(cc);
+                                       c = cc;
+                               } while (c >= ' ' && c != '\\' && is
                                         && i != (LEX_MAX_BUFF-1));
                        }
 
-                       if (i == (LEX_MAX_BUFF-1)) {
+                       if (i == (LEX_MAX_BUFF - 1)) {
                                printError("Line too long");
                        }
 
-                       if (c == '\\') ungetc(c,file); // put it back
+                       if (c == '\\') is.putback(c); // put it back
                        buff[i] = '\0';
                        status = LEX_TOKEN;
                }
                  
-               if (c=='\n')
-                       lineno++;
+               if (c == '\n')
+                       ++lineno;
        
        }
         if (status)  return true;
         
-        status = (feof(file)) ? LEX_FEOF: LEX_UNDEF;
+        status = is.eof() ? LEX_FEOF: LEX_UNDEF;
         buff[0] = '\0';
         return false;
 }
 
 
-int LyXLex::FindToken(char const* string[])
+int LyXLex::FindToken(char const * str[])
 {  
    int i = -1;
    
    if (next()) {
-      if (strcmp(buff, "default")) {
-        for (i=0; string[i][0] && strcmp(string[i], buff); i++);
-        if (!string[i][0]) {
+      if (compare(buff, "default")) {
+        for (i = 0; str[i][0] && compare(str[i], buff); ++i);
+        if (!str[i][0]) {
            printError("Unknown argument `$$Token'");
            i = -1;
         }
@@ -509,13 +545,13 @@ int LyXLex::FindToken(char const* string[])
 }
 
 
-int LyXLex::CheckToken(char const* string[], int print_error)
+int LyXLex::CheckToken(char const * str[], int print_error)
 {  
    int i = -1;
    
-   if (strcmp(buff, "default")) {
-       for (i=0; string[i][0] && strcmp(string[i], buff); i++);
-       if (!string[i][0]) {
+   if (compare(buff, "default")) {
+       for (i = 0; str[i][0] && compare(str[i], buff); ++i);
+       if (!str[i][0]) {
            if (print_error)
                printError("Unknown argument `$$Token'");
            i = -1;