Update to latest lemon.

1 files changed, 262 insertions, 116 deletions

diff --git a/libraries/lemon/lemon.c b/libraries/lemon/lemon.c
index c3e612b..8dca20c 100644
--- a/libraries/lemon/lemon.c
+++ b/libraries/lemon/lemon.c
@@ -15,7 +15,7 @@
 
 #ifndef __WIN32__
 #   if defined(_WIN32) || defined(WIN32)
-#       define __WIN32__
+#       define __WIN32__
 #   endif
 #endif
 
@@ -50,6 +50,107 @@ static char *msort(char*,char**,int(*)(const char*,const char*));
 */
 #define lemonStrlen(X)   ((int)strlen(X))
 
+/*
+** Compilers are starting to complain about the use of sprintf() and strcpy(),
+** saying they are unsafe.  So we define our own versions of those routines too.
+**
+** There are three routines here:  lemon_sprintf(), lemon_vsprintf(), and
+** lemon_addtext().  The first two are replacements for sprintf() and vsprintf().
+** The third is a helper routine for vsnprintf() that adds texts to the end of a
+** buffer, making sure the buffer is always zero-terminated.
+**
+** The string formatter is a minimal subset of stdlib sprintf() supporting only
+** a few simply conversions:
+**
+**   %d
+**   %s
+**   %.*s
+**
+*/
+static void lemon_addtext(
+  char *zBuf,           /* The buffer to which text is added */
+  int *pnUsed,          /* Slots of the buffer used so far */
+  const char *zIn,      /* Text to add */
+  int nIn,              /* Bytes of text to add.  -1 to use strlen() */
+  int iWidth            /* Field width.  Negative to left justify */
+){
+  if( nIn<0 ) for(nIn=0; zIn[nIn]; nIn++){}
+  while( iWidth>nIn ){ zBuf[(*pnUsed)++] = ' '; iWidth--; }
+  if( nIn==0 ) return;
+  memcpy(&zBuf[*pnUsed], zIn, nIn);
+  *pnUsed += nIn;
+  while( (-iWidth)>nIn ){ zBuf[(*pnUsed)++] = ' '; iWidth++; }
+  zBuf[*pnUsed] = 0;
+}
+static int lemon_vsprintf(char *str, const char *zFormat, va_list ap){
+  int i, j, k, c, size;
+  int nUsed = 0;
+  const char *z;
+  char zTemp[50];
+  str[0] = 0;
+  for(i=j=0; (c = zFormat[i])!=0; i++){
+    if( c=='%' ){
+      int iWidth = 0;
+      lemon_addtext(str, &nUsed, &zFormat[j], i-j, 0);
+      c = zFormat[++i];
+      if( isdigit(c) || (c=='-' && isdigit(zFormat[i+1])) ){
+        if( c=='-' ) i++;
+        while( isdigit(zFormat[i]) ) iWidth = iWidth*10 + zFormat[i++] - '0';
+        if( c=='-' ) iWidth = -iWidth;
+        c = zFormat[i];
+      }
+      if( c=='d' ){
+        int v = va_arg(ap, int);
+        if( v<0 ){
+          lemon_addtext(str, &nUsed, "-", 1, iWidth);
+          v = -v;
+        }else if( v==0 ){
+          lemon_addtext(str, &nUsed, "0", 1, iWidth);
+        }
+        k = 0;
+        while( v>0 ){
+          k++;
+          zTemp[sizeof(zTemp)-k] = (v%10) + '0';
+          v /= 10;
+        }
+        lemon_addtext(str, &nUsed, &zTemp[sizeof(zTemp)-k], k, iWidth);
+      }else if( c=='s' ){
+        z = va_arg(ap, const char*);
+        lemon_addtext(str, &nUsed, z, -1, iWidth);
+      }else if( c=='.' && memcmp(&zFormat[i], ".*s", 3)==0 ){
+        i += 2;
+        k = va_arg(ap, int);
+        z = va_arg(ap, const char*);
+        lemon_addtext(str, &nUsed, z, k, iWidth);
+      }else if( c=='%' ){
+        lemon_addtext(str, &nUsed, "%", 1, 0);
+      }else{
+        fprintf(stderr, "illegal format\n");
+        exit(1);
+      }
+      j = i+1;
+    }
+  }
+  lemon_addtext(str, &nUsed, &zFormat[j], i-j, 0);
+  return nUsed;
+}
+static int lemon_sprintf(char *str, const char *format, ...){
+  va_list ap;
+  int rc;
+  va_start(ap, format);
+  rc = lemon_vsprintf(str, format, ap);
+  va_end(ap);
+  return rc;
+}
+static void lemon_strcpy(char *dest, const char *src){
+  while( (*(dest++) = *(src++))!=0 ){}
+}
+static void lemon_strcat(char *dest, const char *src){
+  while( *dest ) dest++;
+  lemon_strcpy(dest, src);
+}
+
+
 /* a few forward declarations... */
 struct rule;
 struct lemon;
@@ -653,7 +754,7 @@ void FindRulePrecedences(struct lemon *xp)
           }
         }else if( sp->prec>=0 ){
           rp->precsym = rp->rhs[i];
-        }
+        }
       }
     }
   }
@@ -711,12 +812,12 @@ void FindFirstSets(struct lemon *lemp)
             progress += SetAdd(s1->firstset,s2->subsym[j]->index);
           }
           break;
-        }else if( s1==s2 ){
+        }else if( s1==s2 ){
           if( s1->lambda==LEMON_FALSE ) break;
-        }else{
+        }else{
           progress += SetUnion(s1->firstset,s2->firstset);
           if( s2->lambda==LEMON_FALSE ) break;
-        }
+        }
       }
     }
   }while( progress );
@@ -959,8 +1060,8 @@ void FindFollowSets(struct lemon *lemp)
           if( change ){
             plp->cfp->status = INCOMPLETE;
             progress = 1;
-          }
-        }
+          }
+        }
         cfp->status = COMPLETE;
       }
     }
@@ -993,7 +1094,7 @@ void FindActions(struct lemon *lemp)
             ** rule "cfp->rp" if the lookahead symbol is "lemp->symbols[j]" */
             Action_add(&stp->ap,REDUCE,lemp->symbols[j],(char *)cfp->rp);
           }
-        }
+        }
       }
     }
   }
@@ -1262,11 +1363,11 @@ void Configlist_closure(struct lemon *lemp)
               SetAdd(newcfp->fws, xsp->subsym[k]->index);
             }
             break;
-          }else{
+          }else{
             SetUnion(newcfp->fws,xsp->firstset);
             if( xsp->lambda==LEMON_FALSE ) break;
-          }
-        }
+          }
+        }
         if( i==rp->nrhs ) Plink_add(&cfp->fplp,newcfp);
       }
     }
@@ -1367,7 +1468,7 @@ static void handle_D_option(char *z){
     fprintf(stderr,"out of memory\n");
     exit(1);
   }
-  strcpy(*paz, z);
+  lemon_strcpy(*paz, z);
   for(z=*paz; *z && *z!='='; z++){}
   *z = 0;
 }
@@ -1378,7 +1479,7 @@ static void handle_T_option(char *z){
   if( user_templatename==0 ){
     memory_error();
   }
-  strcpy(user_templatename, z);
+  lemon_strcpy(user_templatename, z);
 }
 
 /* The main program.  Parse the command line and do it... */
@@ -1447,12 +1548,15 @@ int main(int argc, char **argv)
   }
 
   /* Count and index the symbols of the grammar */
-  lem.nsymbol = Symbol_count();
   Symbol_new("{default}");
+  lem.nsymbol = Symbol_count();
   lem.symbols = Symbol_arrayof();
-  for(i=0; i<=lem.nsymbol; i++) lem.symbols[i]->index = i;
-  qsort(lem.symbols,lem.nsymbol+1,sizeof(struct symbol*), Symbolcmpp);
-  for(i=0; i<=lem.nsymbol; i++) lem.symbols[i]->index = i;
+  for(i=0; i<lem.nsymbol; i++) lem.symbols[i]->index = i;
+  qsort(lem.symbols,lem.nsymbol,sizeof(struct symbol*), Symbolcmpp);
+  for(i=0; i<lem.nsymbol; i++) lem.symbols[i]->index = i;
+  while( lem.symbols[i-1]->type==MULTITERMINAL ){ i--; }
+  assert( strcmp(lem.symbols[i-1]->name,"{default}")==0 );
+  lem.nsymbol = i - 1;
   for(i=1; isupper(lem.symbols[i]->name[0]); i++);
   lem.nterminal = i;
 
@@ -1546,7 +1650,7 @@ int main(int argc, char **argv)
 /*
 ** Return a pointer to the next structure in the linked list.
 */
-#define NEXT(A) (*(char**)(((unsigned long)A)+offset))
+#define NEXT(A) (*(char**)(((char*)A)+offset))
 
 /*
 ** Inputs:
@@ -1940,7 +2044,9 @@ enum e_state {
   WAITING_FOR_DESTRUCTOR_SYMBOL,
   WAITING_FOR_DATATYPE_SYMBOL,
   WAITING_FOR_FALLBACK_ID,
-  WAITING_FOR_WILDCARD_ID
+  WAITING_FOR_WILDCARD_ID,
+  WAITING_FOR_CLASS_ID,
+  WAITING_FOR_CLASS_TOKEN
 };
 struct pstate {
   char *filename;       /* Name of the input file */
@@ -1950,6 +2056,7 @@ struct pstate {
   struct lemon *gp;     /* Global state vector */
   enum e_state state;        /* The state of the parser */
   struct symbol *fallback;   /* The fallback token */
+  struct symbol *tkclass;    /* Token class symbol */
   struct symbol *lhs;        /* Left-hand side of current rule */
   const char *lhsalias;      /* Alias for the LHS */
   int nrhs;                  /* Number of right-hand side symbols seen */
@@ -1996,7 +2103,7 @@ static void parseonetoken(struct pstate *psp)
 "There is no prior rule upon which to attach the code \
 fragment which begins on this line.");
           psp->errorcnt++;
-        }else if( psp->prevrule->code!=0 ){
+        }else if( psp->prevrule->code!=0 ){
           ErrorMsg(psp->filename,psp->tokenlineno,
 "Code fragment beginning on this line is not the first \
 to follow the previous rule.");
@@ -2004,7 +2111,7 @@ to follow the previous rule.");
         }else{
           psp->prevrule->line = psp->tokenlineno;
           psp->prevrule->code = &x[1];
-        }
+        }
       }else if( x[0]=='[' ){
         psp->state = PRECEDENCE_MARK_1;
       }else{
@@ -2097,7 +2204,7 @@ to follow the previous rule.");
             "Can't allocate enough memory for this rule.");
           psp->errorcnt++;
           psp->prevrule = 0;
-        }else{
+        }else{
           int i;
           rp->ruleline = psp->tokenlineno;
           rp->rhs = (struct symbol**)&rp[1];
@@ -2105,7 +2212,7 @@ to follow the previous rule.");
           for(i=0; i<psp->nrhs; i++){
             rp->rhs[i] = psp->rhs[i];
             rp->rhsalias[i] = psp->alias[i];
-          }
+          }
           rp->lhs = psp->lhs;
           rp->lhsalias = psp->lhsalias;
           rp->nrhs = psp->nrhs;
@@ -2117,12 +2224,12 @@ to follow the previous rule.");
           rp->next = 0;
           if( psp->firstrule==0 ){
             psp->firstrule = psp->lastrule = rp;
-          }else{
+          }else{
             psp->lastrule->next = rp;
             psp->lastrule = rp;
-          }
+          }
           psp->prevrule = rp;
-        }
+        }
         psp->state = WAITING_FOR_DECL_OR_RULE;
       }else if( isalpha(x[0]) ){
         if( psp->nrhs>=MAXRHS ){
@@ -2131,11 +2238,11 @@ to follow the previous rule.");
             x);
           psp->errorcnt++;
           psp->state = RESYNC_AFTER_RULE_ERROR;
-        }else{
+        }else{
           psp->rhs[psp->nrhs] = Symbol_new(x);
           psp->alias[psp->nrhs] = 0;
           psp->nrhs++;
-        }
+        }
       }else if( (x[0]=='|' || x[0]=='/') && psp->nrhs>0 ){
         struct symbol *msp = psp->rhs[psp->nrhs-1];
         if( msp->type!=MULTITERMINAL ){
@@ -2199,24 +2306,24 @@ to follow the previous rule.");
         if( strcmp(x,"name")==0 ){
           psp->declargslot = &(psp->gp->name);
           psp->insertLineMacro = 0;
-        }else if( strcmp(x,"include")==0 ){
+        }else if( strcmp(x,"include")==0 ){
           psp->declargslot = &(psp->gp->include);
-        }else if( strcmp(x,"code")==0 ){
+        }else if( strcmp(x,"code")==0 ){
           psp->declargslot = &(psp->gp->extracode);
-        }else if( strcmp(x,"token_destructor")==0 ){
+        }else if( strcmp(x,"token_destructor")==0 ){
           psp->declargslot = &psp->gp->tokendest;
-        }else if( strcmp(x,"default_destructor")==0 ){
+        }else if( strcmp(x,"default_destructor")==0 ){
           psp->declargslot = &psp->gp->vardest;
-        }else if( strcmp(x,"token_prefix")==0 ){
+        }else if( strcmp(x,"token_prefix")==0 ){
           psp->declargslot = &psp->gp->tokenprefix;
           psp->insertLineMacro = 0;
-        }else if( strcmp(x,"syntax_error")==0 ){
+        }else if( strcmp(x,"syntax_error")==0 ){
           psp->declargslot = &(psp->gp->error);
-        }else if( strcmp(x,"parse_accept")==0 ){
+        }else if( strcmp(x,"parse_accept")==0 ){
           psp->declargslot = &(psp->gp->accept);
-        }else if( strcmp(x,"parse_failure")==0 ){
+        }else if( strcmp(x,"parse_failure")==0 ){
           psp->declargslot = &(psp->gp->failure);
-        }else if( strcmp(x,"stack_overflow")==0 ){
+        }else if( strcmp(x,"stack_overflow")==0 ){
           psp->declargslot = &(psp->gp->overflow);
         }else if( strcmp(x,"extra_argument")==0 ){
           psp->declargslot = &(psp->gp->arg);
@@ -2245,21 +2352,23 @@ to follow the previous rule.");
           psp->preccounter++;
           psp->declassoc = NONE;
           psp->state = WAITING_FOR_PRECEDENCE_SYMBOL;
-        }else if( strcmp(x,"destructor")==0 ){
+        }else if( strcmp(x,"destructor")==0 ){
           psp->state = WAITING_FOR_DESTRUCTOR_SYMBOL;
-        }else if( strcmp(x,"type")==0 ){
+        }else if( strcmp(x,"type")==0 ){
           psp->state = WAITING_FOR_DATATYPE_SYMBOL;
         }else if( strcmp(x,"fallback")==0 ){
           psp->fallback = 0;
           psp->state = WAITING_FOR_FALLBACK_ID;
         }else if( strcmp(x,"wildcard")==0 ){
           psp->state = WAITING_FOR_WILDCARD_ID;
+        }else if( strcmp(x,"token_class")==0 ){
+          psp->state = WAITING_FOR_CLASS_ID;
         }else{
           ErrorMsg(psp->filename,psp->tokenlineno,
             "Unknown declaration keyword: \"%%%s\".",x);
           psp->errorcnt++;
           psp->state = RESYNC_AFTER_DECL_ERROR;
-        }
+        }
       }else{
         ErrorMsg(psp->filename,psp->tokenlineno,
           "Illegal declaration keyword: \"%s\".",x);
@@ -2314,10 +2423,10 @@ to follow the previous rule.");
           ErrorMsg(psp->filename,psp->tokenlineno,
             "Symbol \"%s\" has already be given a precedence.",x);
           psp->errorcnt++;
-        }else{
+        }else{
           sp->prec = psp->preccounter;
           sp->assoc = psp->declassoc;
-        }
+        }
       }else{
         ErrorMsg(psp->filename,psp->tokenlineno,
           "Can't assign a precedence to \"%s\".",x);
@@ -2347,7 +2456,7 @@ to follow the previous rule.");
           for(z=psp->filename, nBack=0; *z; z++){
             if( *z=='\\' ) nBack++;
           }
-          sprintf(zLine, "#line %d ", psp->tokenlineno);
+          lemon_sprintf(zLine, "#line %d ", psp->tokenlineno);
           nLine = lemonStrlen(zLine);
           n += nLine + lemonStrlen(psp->filename) + nBack;
         }
@@ -2422,6 +2531,40 @@ to follow the previous rule.");
         }
       }
       break;
+    case WAITING_FOR_CLASS_ID:
+      if( !islower(x[0]) ){
+        ErrorMsg(psp->filename, psp->tokenlineno,
+          "%%token_class must be followed by an identifier: ", x);
+        psp->errorcnt++;
+        psp->state = RESYNC_AFTER_DECL_ERROR;
+     }else if( Symbol_find(x) ){
+        ErrorMsg(psp->filename, psp->tokenlineno,
+          "Symbol \"%s\" already used", x);
+        psp->errorcnt++;
+        psp->state = RESYNC_AFTER_DECL_ERROR;
+      }else{
+        psp->tkclass = Symbol_new(x);
+        psp->tkclass->type = MULTITERMINAL;
+        psp->state = WAITING_FOR_CLASS_TOKEN;
+      }
+      break;
+    case WAITING_FOR_CLASS_TOKEN:
+      if( x[0]=='.' ){
+        psp->state = WAITING_FOR_DECL_OR_RULE;
+      }else if( isupper(x[0]) || ((x[0]=='|' || x[0]=='/') && isupper(x[1])) ){
+        struct symbol *msp = psp->tkclass;
+        msp->nsubsym++;
+        msp->subsym = (struct symbol **) realloc(msp->subsym,
+          sizeof(struct symbol*)*msp->nsubsym);
+        if( !isupper(x[0]) ) x++;
+        msp->subsym[msp->nsubsym-1] = Symbol_new(x);
+      }else{
+        ErrorMsg(psp->filename, psp->tokenlineno,
+          "%%token_class argument \"%s\" should be a token", x);
+        psp->errorcnt++;
+        psp->state = RESYNC_AFTER_DECL_ERROR;
+      }
+      break;
     case RESYNC_AFTER_RULE_ERROR:
 /*      if( x[0]=='.' ) psp->state = WAITING_FOR_DECL_OR_RULE;
 **      break; */
@@ -2516,9 +2659,8 @@ void Parse(struct lemon *gp)
   filesize = ftell(fp);
   rewind(fp);
   filebuf = (char *)malloc( filesize+1 );
-  if( filebuf==0 ){
-    ErrorMsg(ps.filename,0,"Can't allocate %d of memory to hold this file.",
-      filesize+1);
+  if( filesize>100000000 || filebuf==0 ){
+    ErrorMsg(ps.filename,0,"Input file too large.");
     gp->errorcnt++;
     fclose(fp);
     return;
@@ -2587,12 +2729,12 @@ void Parse(struct lemon *gp)
             if( c=='\n' ) lineno++;
             prevc = c;
             cp++;
-          }
-        }else if( c=='/' && cp[1]=='/' ){  /* Skip C++ style comments too */
+          }
+        }else if( c=='/' && cp[1]=='/' ){  /* Skip C++ style comments too */
           cp = &cp[2];
           while( (c= *cp)!=0 && c!='\n' ) cp++;
           if( c ) lineno++;
-        }else if( c=='\'' || c=='\"' ){    /* String a character literals */
+        }else if( c=='\'' || c=='\"' ){    /* String a character literals */
           int startchar, prevc;
           startchar = c;
           prevc = 0;
@@ -2600,8 +2742,8 @@ void Parse(struct lemon *gp)
             if( c=='\n' ) lineno++;
             if( prevc=='\\' ) prevc = 0;
             else              prevc = c;
-          }
-        }
+          }
+        }
       }
       if( c==0 ){
         ErrorMsg(ps.filename,ps.tokenlineno,
@@ -2716,10 +2858,10 @@ PRIVATE char *file_makename(struct lemon *lemp, const char *suffix)
     fprintf(stderr,"Can't allocate space for a filename.\n");
     exit(1);
   }
-  strcpy(name,lemp->filename);
+  lemon_strcpy(name,lemp->filename);
   cp = strrchr(name,'.');
   if( cp ) *cp = 0;
-  strcat(name,suffix);
+  lemon_strcat(name,suffix);
   return name;
 }
 
@@ -2776,11 +2918,13 @@ void Reprint(struct lemon *lemp)
     printf(" ::=");
     for(i=0; i<rp->nrhs; i++){
       sp = rp->rhs[i];
-      printf(" %s", sp->name);
       if( sp->type==MULTITERMINAL ){
+        printf(" %s", sp->subsym[0]->name);
         for(j=1; j<sp->nsubsym; j++){
           printf("|%s", sp->subsym[j]->name);
         }
+      }else{
+        printf(" %s", sp->name);
       }
       /* if( rp->rhsalias[i] ) printf("(%s)",rp->rhsalias[i]); */
     }
@@ -2802,11 +2946,13 @@ void ConfigPrint(FILE *fp, struct config *cfp)
     if( i==cfp->dot ) fprintf(fp," *");
     if( i==rp->nrhs ) break;
     sp = rp->rhs[i];
-    fprintf(fp," %s", sp->name);
     if( sp->type==MULTITERMINAL ){
+      fprintf(fp," %s", sp->subsym[0]->name);
       for(j=1; j<sp->nsubsym; j++){
         fprintf(fp,"|%s",sp->subsym[j]->name);
       }
+    }else{
+      fprintf(fp," %s", sp->name);
     }
   }
 }
@@ -2916,7 +3062,7 @@ void ReportOutput(struct lemon *lemp)
     while( cfp ){
       char buf[20];
       if( cfp->dot==cfp->rp->nrhs ){
-        sprintf(buf,"(%d)",cfp->rp->index);
+        lemon_sprintf(buf,"(%d)",cfp->rp->index);
         fprintf(fp,"    %5s ",buf);
       }else{
         fprintf(fp,"          ");
@@ -2981,7 +3127,7 @@ PRIVATE char *pathsearch(char *argv0, char *name, int modemask)
     c = *cp;
     *cp = 0;
     path = (char *)malloc( lemonStrlen(argv0) + lemonStrlen(name) + 2 );
-    if( path ) sprintf(path,"%s/%s",argv0,name);
+    if( path ) lemon_sprintf(path,"%s/%s",argv0,name);
     *cp = c;
   }else{
     pathlist = getenv("PATH");
@@ -2990,13 +3136,13 @@ PRIVATE char *pathsearch(char *argv0, char *name, int modemask)
     path = (char *)malloc( lemonStrlen(pathlist)+lemonStrlen(name)+2 );
     if( (pathbuf != 0) && (path!=0) ){
       pathbufptr = pathbuf;
-      strcpy(pathbuf, pathlist);
+      lemon_strcpy(pathbuf, pathlist);
       while( *pathbuf ){
         cp = strchr(pathbuf,':');
         if( cp==0 ) cp = &pathbuf[lemonStrlen(pathbuf)];
         c = *cp;
         *cp = 0;
-        sprintf(path,"%s/%s",pathbuf,name);
+        lemon_sprintf(path,"%s/%s",pathbuf,name);
         *cp = c;
         if( c==0 ) pathbuf[0] = 0;
         else pathbuf = &cp[1];
@@ -3087,9 +3233,9 @@ PRIVATE FILE *tplt_open(struct lemon *lemp)
 
   cp = strrchr(lemp->filename,'.');
   if( cp ){
-    sprintf(buf,"%.*s.lt",(int)(cp-lemp->filename),lemp->filename);
+    lemon_sprintf(buf,"%.*s.lt",(int)(cp-lemp->filename),lemp->filename);
   }else{
-    sprintf(buf,"%s.lt",lemp->filename);
+    lemon_sprintf(buf,"%s.lt",lemp->filename);
   }
   if( access(buf,004)==0 ){
     tpltname = buf;
@@ -3240,9 +3386,9 @@ PRIVATE char *append_str(const char *zText, int n, int p1, int p2){
   while( n-- > 0 ){
     c = *(zText++);
     if( c=='%' && n>0 && zText[0]=='d' ){
-      sprintf(zInt, "%d", p1);
+      lemon_sprintf(zInt, "%d", p1);
       p1 = p2;
-      strcpy(&z[used], zInt);
+      lemon_strcpy(&z[used], zInt);
       used += lemonStrlen(&z[used]);
       zText++;
       n--;
@@ -3391,7 +3537,7 @@ void print_stack_union(
   int maxdtlength;          /* Maximum length of any ".datatype" field. */
   char *stddt;              /* Standardized name for a datatype */
   int i,j;                  /* Loop counters */
-  int hash;                 /* For hashing the name of a type */
+  unsigned hash;            /* For hashing the name of a type */
   const char *name;         /* Name of the parser */
 
   /* Allocate and initialize types[] and allocate stddt[] */
@@ -3458,7 +3604,7 @@ void print_stack_union(
         break;
       }
       hash++;
-      if( hash>=arraysize ) hash = 0;
+      if( hash>=(unsigned)arraysize ) hash = 0;
     }
     if( types[hash]==0 ){
       sp->dtnum = hash + 1;
@@ -3467,7 +3613,7 @@ void print_stack_union(
         fprintf(stderr,"Out of memory.\n");
         exit(1);
       }
-      strcpy(types[hash],stddt);
+      lemon_strcpy(types[hash],stddt);
     }
   }
 
@@ -3553,9 +3699,11 @@ static void writeRuleText(FILE *out, struct rule *rp){
   fprintf(out,"%s ::=", rp->lhs->name);
   for(j=0; j<rp->nrhs; j++){
     struct symbol *sp = rp->rhs[j];
-    fprintf(out," %s", sp->name);
-    if( sp->type==MULTITERMINAL ){
+    if( sp->type!=MULTITERMINAL ){
+      fprintf(out," %s", sp->name);
+    }else{
       int k;
+      fprintf(out," %s", sp->subsym[0]->name);
       for(k=1; k<sp->nsubsym; k++){
         fprintf(out,"|%s",sp->subsym[k]->name);
       }
@@ -3856,7 +4004,7 @@ void ReportTable(
   /* Generate a table containing the symbolic name of every symbol
   */
   for(i=0; i<lemp->nsymbol; i++){
-    sprintf(line,"\"%s\",",lemp->symbols[i]->name);
+    lemon_sprintf(line,"\"%s\",",lemp->symbols[i]->name);
     fprintf(out,"  %-15s",line);
     if( (i&3)==3 ){ fprintf(out,"\n"); lineno++; }
   }
@@ -4021,12 +4169,15 @@ void ReportHeader(struct lemon *lemp)
   else                    prefix = "";
   in = file_open(lemp,".h","rb");
   if( in ){
+    int nextChar;
     for(i=1; i<lemp->nterminal && fgets(line,LINESIZE,in); i++){
-      sprintf(pattern,"#define %s%-30s %2d\n",prefix,lemp->symbols[i]->name,i);
+      lemon_sprintf(pattern,"#define %s%-30s %3d\n",
+                    prefix,lemp->symbols[i]->name,i);
       if( strcmp(line,pattern) ) break;
     }
+    nextChar = fgetc(in);
     fclose(in);
-    if( i==lemp->nterminal ){
+    if( i==lemp->nterminal && nextChar==EOF ){
       /* No change in the file.  Don't rewrite it. */
       return;
     }
@@ -4034,7 +4185,7 @@ void ReportHeader(struct lemon *lemp)
   out = file_open(lemp,".h","wb");
   if( out ){
     for(i=1; i<lemp->nterminal; i++){
-      fprintf(out,"#define %s%-30s %2d\n",prefix,lemp->symbols[i]->name,i);
+      fprintf(out,"#define %s%-30s %3d\n",prefix,lemp->symbols[i]->name,i);
     }
     fclose(out);  
   }
@@ -4232,10 +4383,10 @@ int SetUnion(char *s1, char *s2)
 ** Code for processing tables in the LEMON parser generator.
 */
 
-PRIVATE int strhash(const char *x)
+PRIVATE unsigned strhash(const char *x)
 {
-  int h = 0;
-  while( *x) h = h*13 + *(x++);
+  unsigned h = 0;
+  while( *x ) h = h*13 + *(x++);
   return h;
 }
 
@@ -4251,7 +4402,7 @@ const char *Strsafe(const char *y)
   if( y==0 ) return 0;
   z = Strsafe_find(y);
   if( z==0 && (cpy=(char *)malloc( lemonStrlen(y)+1 ))!=0 ){
-    strcpy(cpy,y);
+    lemon_strcpy(cpy,y);
     z = cpy;
     Strsafe_insert(z);
   }
@@ -4290,8 +4441,7 @@ void Strsafe_init(){
   if( x1a ){
     x1a->size = 1024;
     x1a->count = 0;
-    x1a->tbl = (x1node*)malloc( 
-      (sizeof(x1node) + sizeof(x1node*))*1024 );
+    x1a->tbl = (x1node*)calloc(1024, sizeof(x1node) + sizeof(x1node*));
     if( x1a->tbl==0 ){
       free(x1a);
       x1a = 0;
@@ -4307,8 +4457,8 @@ void Strsafe_init(){
 int Strsafe_insert(const char *data)
 {
   x1node *np;
-  int h;
-  int ph;
+  unsigned h;
+  unsigned ph;
 
   if( x1a==0 ) return 0;
   ph = strhash(data);
@@ -4328,8 +4478,7 @@ int Strsafe_insert(const char *data)
     struct s_x1 array;
     array.size = size = x1a->size*2;
     array.count = x1a->count;
-    array.tbl = (x1node*)malloc(
-      (sizeof(x1node) + sizeof(x1node*))*size );
+    array.tbl = (x1node*)calloc(size, sizeof(x1node) + sizeof(x1node*));
     if( array.tbl==0 ) return 0;  /* Fail due to malloc failure */
     array.ht = (x1node**)&(array.tbl[size]);
     for(i=0; i<size; i++) array.ht[i] = 0;
@@ -4362,7 +4511,7 @@ int Strsafe_insert(const char *data)
 ** if no such key. */
 const char *Strsafe_find(const char *key)
 {
-  int h;
+  unsigned h;
   x1node *np;
 
   if( x1a==0 ) return 0;
@@ -4404,11 +4553,15 @@ struct symbol *Symbol_new(const char *x)
   return sp;
 }
 
-/* Compare two symbols for working purposes
+/* Compare two symbols for sorting purposes.  Return negative,
+** zero, or positive if a is less then, equal to, or greater
+** than b.
 **
 ** Symbols that begin with upper case letters (terminals or tokens)
 ** must sort before symbols that begin with lower case letters
-** (non-terminals).  Other than that, the order does not matter.
+** (non-terminals).  And MULTITERMINAL symbols (created using the
+** %token_class directive) must sort at the very end. Other than
+** that, the order does not matter.
 **
 ** We find experimentally that leaving the symbols in their original
 ** order (the order they appeared in the grammar file) gives the
@@ -4416,12 +4569,11 @@ struct symbol *Symbol_new(const char *x)
 */
 int Symbolcmpp(const void *_a, const void *_b)
 {
-  const struct symbol **a = (const struct symbol **) _a;
-  const struct symbol **b = (const struct symbol **) _b;
-  int i1 = (**a).index + 10000000*((**a).name[0]>'Z');
-  int i2 = (**b).index + 10000000*((**b).name[0]>'Z');
-  assert( i1!=i2 || strcmp((**a).name,(**b).name)==0 );
-  return i1-i2;
+  const struct symbol *a = *(const struct symbol **) _a;
+  const struct symbol *b = *(const struct symbol **) _b;
+  int i1 = a->type==MULTITERMINAL ? 3 : a->name[0]>'Z' ? 2 : 1;
+  int i2 = b->type==MULTITERMINAL ? 3 : b->name[0]>'Z' ? 2 : 1;
+  return i1==i2 ? a->index - b->index : i1 - i2;
 }
 
 /* There is one instance of the following structure for each
@@ -4456,8 +4608,7 @@ void Symbol_init(){
   if( x2a ){
     x2a->size = 128;
     x2a->count = 0;
-    x2a->tbl = (x2node*)malloc( 
-      (sizeof(x2node) + sizeof(x2node*))*128 );
+    x2a->tbl = (x2node*)calloc(128, sizeof(x2node) + sizeof(x2node*));
     if( x2a->tbl==0 ){
       free(x2a);
       x2a = 0;
@@ -4473,8 +4624,8 @@ void Symbol_init(){
 int Symbol_insert(struct symbol *data, const char *key)
 {
   x2node *np;
-  int h;
-  int ph;
+  unsigned h;
+  unsigned ph;
 
   if( x2a==0 ) return 0;
   ph = strhash(key);
@@ -4494,8 +4645,7 @@ int Symbol_insert(struct symbol *data, const char *key)
     struct s_x2 array;
     array.size = size = x2a->size*2;
     array.count = x2a->count;
-    array.tbl = (x2node*)malloc(
-      (sizeof(x2node) + sizeof(x2node*))*size );
+    array.tbl = (x2node*)calloc(size, sizeof(x2node) + sizeof(x2node*));
     if( array.tbl==0 ) return 0;  /* Fail due to malloc failure */
     array.ht = (x2node**)&(array.tbl[size]);
     for(i=0; i<size; i++) array.ht[i] = 0;
@@ -4530,7 +4680,7 @@ int Symbol_insert(struct symbol *data, const char *key)
 ** if no such key. */
 struct symbol *Symbol_find(const char *key)
 {
-  int h;
+  unsigned h;
   x2node *np;
 
   if( x2a==0 ) return 0;
@@ -4604,9 +4754,9 @@ PRIVATE int statecmp(struct config *a, struct config *b)
 }
 
 /* Hash a state */
-PRIVATE int statehash(struct config *a)
+PRIVATE unsigned statehash(struct config *a)
 {
-  int h=0;
+  unsigned h=0;
   while( a ){
     h = h*571 + a->rp->index*37 + a->dot;
     a = a->bp;
@@ -4655,8 +4805,7 @@ void State_init(){
   if( x3a ){
     x3a->size = 128;
     x3a->count = 0;
-    x3a->tbl = (x3node*)malloc( 
-      (sizeof(x3node) + sizeof(x3node*))*128 );
+    x3a->tbl = (x3node*)calloc(128, sizeof(x3node) + sizeof(x3node*));
     if( x3a->tbl==0 ){
       free(x3a);
       x3a = 0;
@@ -4672,8 +4821,8 @@ void State_init(){
 int State_insert(struct state *data, struct config *key)
 {
   x3node *np;
-  int h;
-  int ph;
+  unsigned h;
+  unsigned ph;
 
   if( x3a==0 ) return 0;
   ph = statehash(key);
@@ -4693,8 +4842,7 @@ int State_insert(struct state *data, struct config *key)
     struct s_x3 array;
     array.size = size = x3a->size*2;
     array.count = x3a->count;
-    array.tbl = (x3node*)malloc(
-      (sizeof(x3node) + sizeof(x3node*))*size );
+    array.tbl = (x3node*)calloc(size, sizeof(x3node) + sizeof(x3node*));
     if( array.tbl==0 ) return 0;  /* Fail due to malloc failure */
     array.ht = (x3node**)&(array.tbl[size]);
     for(i=0; i<size; i++) array.ht[i] = 0;
@@ -4729,7 +4877,7 @@ int State_insert(struct state *data, struct config *key)
 ** if no such key. */
 struct state *State_find(struct config *key)
 {
-  int h;
+  unsigned h;
   x3node *np;
 
   if( x3a==0 ) return 0;
@@ -4751,7 +4899,7 @@ struct state **State_arrayof()
   int i,size;
   if( x3a==0 ) return 0;
   size = x3a->count;
-  array = (struct state **)malloc( sizeof(struct state *)*size );
+  array = (struct state **)calloc(size, sizeof(struct state *));
   if( array ){
     for(i=0; i<size; i++) array[i] = x3a->tbl[i].data;
   }
@@ -4759,9 +4907,9 @@ struct state **State_arrayof()
 }
 
 /* Hash a configuration */
-PRIVATE int confighash(struct config *a)
+PRIVATE unsigned confighash(struct config *a)
 {
-  int h=0;
+  unsigned h=0;
   h = h*571 + a->rp->index*37 + a->dot;
   return h;
 }
@@ -4797,8 +4945,7 @@ void Configtable_init(){
   if( x4a ){
     x4a->size = 64;
     x4a->count = 0;
-    x4a->tbl = (x4node*)malloc( 
-      (sizeof(x4node) + sizeof(x4node*))*64 );
+    x4a->tbl = (x4node*)calloc(64, sizeof(x4node) + sizeof(x4node*));
     if( x4a->tbl==0 ){
       free(x4a);
       x4a = 0;
@@ -4814,8 +4961,8 @@ void Configtable_init(){
 int Configtable_insert(struct config *data)
 {
   x4node *np;
-  int h;
-  int ph;
+  unsigned h;
+  unsigned ph;
 
   if( x4a==0 ) return 0;
   ph = confighash(data);
@@ -4835,8 +4982,7 @@ int Configtable_insert(struct config *data)
     struct s_x4 array;
     array.size = size = x4a->size*2;
     array.count = x4a->count;
-    array.tbl = (x4node*)malloc(
-      (sizeof(x4node) + sizeof(x4node*))*size );
+    array.tbl = (x4node*)calloc(size, sizeof(x4node) + sizeof(x4node*));
     if( array.tbl==0 ) return 0;  /* Fail due to malloc failure */
     array.ht = (x4node**)&(array.tbl[size]);
     for(i=0; i<size; i++) array.ht[i] = 0;