ncurses 5.9 - patch 20110820
[ncurses.git] / ncurses / tinfo / comp_scan.c
index 152930885ab75ffa9dc459205fd85ad44dec7ac9..ef604026a151c8173d2c1cd99b59273822df447a 100644 (file)
@@ -1,5 +1,5 @@
 /****************************************************************************
 /****************************************************************************
- * Copyright (c) 1998 Free Software Foundation, Inc.                        *
+ * Copyright (c) 1998-2010,2011 Free Software Foundation, Inc.              *
  *                                                                          *
  * Permission is hereby granted, free of charge, to any person obtaining a  *
  * copy of this software and associated documentation files (the            *
  *                                                                          *
  * Permission is hereby granted, free of charge, to any person obtaining a  *
  * copy of this software and associated documentation files (the            *
@@ -29,6 +29,7 @@
 /****************************************************************************
  *  Author: Zeyd M. Ben-Halim <zmbenhal@netcom.com> 1992,1995               *
  *     and: Eric S. Raymond <esr@snark.thyrsus.com>                         *
 /****************************************************************************
  *  Author: Zeyd M. Ben-Halim <zmbenhal@netcom.com> 1992,1995               *
  *     and: Eric S. Raymond <esr@snark.thyrsus.com>                         *
+ *     and: Thomas E. Dickey                        1996 on                 *
  ****************************************************************************/
 
 /*
  ****************************************************************************/
 
 /*
@@ -49,7 +50,7 @@
 #include <ctype.h>
 #include <tic.h>
 
 #include <ctype.h>
 #include <tic.h>
 
-MODULE_ID("$Id: comp_scan.c,v 1.34 1998/11/01 00:56:39 tom Exp $")
+MODULE_ID("$Id: comp_scan.c,v 1.93 2011/08/20 21:19:40 tom Exp $")
 
 /*
  * Maximum length of string capability we'll accept before raising an error.
 
 /*
  * Maximum length of string capability we'll accept before raising an error.
@@ -59,11 +60,17 @@ MODULE_ID("$Id: comp_scan.c,v 1.34 1998/11/01 00:56:39 tom Exp $")
 
 #define iswhite(ch)    (ch == ' '  ||  ch == '\t')
 
 
 #define iswhite(ch)    (ch == ' '  ||  ch == '\t')
 
-int    _nc_syntax;             /* termcap or terminfo? */
-long   _nc_curr_file_pos;      /* file offset of current line */
-long   _nc_comment_start;      /* start of comment range before name */
-long   _nc_comment_end;        /* end of comment range before name */
-long   _nc_start_line;         /* start line of current entry */
+NCURSES_EXPORT_VAR (int) _nc_syntax = 0;         /* termcap or terminfo? */
+NCURSES_EXPORT_VAR (int) _nc_strict_bsd = 1;  /* ncurses extended termcap? */
+NCURSES_EXPORT_VAR (long) _nc_curr_file_pos = 0; /* file offset of current line */
+NCURSES_EXPORT_VAR (long) _nc_comment_start = 0; /* start of comment range before name */
+NCURSES_EXPORT_VAR (long) _nc_comment_end = 0;   /* end of comment range before name */
+NCURSES_EXPORT_VAR (long) _nc_start_line = 0;    /* start line of current entry */
+
+NCURSES_EXPORT_VAR (struct token) _nc_curr_token =
+{
+    0, 0, 0
+};
 
 /*****************************************************************************
  *
 
 /*****************************************************************************
  *
@@ -72,25 +79,229 @@ long       _nc_start_line;         /* start line of current entry */
  *****************************************************************************/
 
 static bool first_column;      /* See 'next_char()' below */
  *****************************************************************************/
 
 static bool first_column;      /* See 'next_char()' below */
+static bool had_newline;
 static char separator;         /* capability separator */
 static int pushtype;           /* type of pushback token */
 static char separator;         /* capability separator */
 static int pushtype;           /* type of pushback token */
-static char pushname[MAX_NAME_SIZE+1];
+static char *pushname;
+
+#if NCURSES_EXT_FUNCS
+NCURSES_EXPORT_VAR (bool) _nc_disable_period = FALSE; /* used by tic -a option */
+#endif
 
 
-static int  last_char(void);
-static int  next_char(void);
-static long stream_pos(void);
-static bool end_of_stream(void);
-static void push_back(char c);
+/*****************************************************************************
+ *
+ * Character-stream handling
+ *
+ *****************************************************************************/
+
+#define LEXBUFSIZ      1024
+
+static char *bufptr;           /* otherwise, the input buffer pointer */
+static char *bufstart;         /* start of buffer so we can compute offsets */
+static FILE *yyin;             /* scanner's input file descriptor */
+
+/*
+ *     _nc_reset_input()
+ *
+ *     Resets the input-reading routines.  Used on initialization,
+ *     or after a seek has been done.  Exactly one argument must be
+ *     non-null.
+ */
+
+NCURSES_EXPORT(void)
+_nc_reset_input(FILE *fp, char *buf)
+{
+    pushtype = NO_PUSHBACK;
+    if (pushname != 0)
+       pushname[0] = '\0';
+    yyin = fp;
+    bufstart = bufptr = buf;
+    _nc_curr_file_pos = 0L;
+    if (fp != 0)
+       _nc_curr_line = 0;
+    _nc_curr_col = 0;
+}
+
+/*
+ *     int last_char()
+ *
+ *     Returns the final nonblank character on the current input buffer
+ */
+static int
+last_char(void)
+{
+    size_t len = strlen(bufptr);
+    while (len--) {
+       if (!isspace(UChar(bufptr[len])))
+           return bufptr[len];
+    }
+    return 0;
+}
+
+/*
+ *     int next_char()
+ *
+ *     Returns the next character in the input stream.  Comments and leading
+ *     white space are stripped.
+ *
+ *     The global state variable 'firstcolumn' is set TRUE if the character
+ *     returned is from the first column of the input line.
+ *
+ *     The global variable _nc_curr_line is incremented for each new line.
+ *     The global variable _nc_curr_file_pos is set to the file offset of the
+ *     beginning of each line.
+ */
+
+static int
+next_char(void)
+{
+    static char *result;
+    static size_t allocated;
+    int the_char;
+
+    if (!yyin) {
+       if (result != 0) {
+           FreeAndNull(result);
+           FreeAndNull(pushname);
+           allocated = 0;
+       }
+       /*
+        * An string with an embedded null will truncate the input.  This is
+        * intentional (we don't read binary files here).
+        */
+       if (bufptr == 0 || *bufptr == '\0')
+           return (EOF);
+       if (*bufptr == '\n') {
+           _nc_curr_line++;
+           _nc_curr_col = 0;
+       } else if (*bufptr == '\t') {
+           _nc_curr_col = (_nc_curr_col | 7);
+       }
+    } else if (!bufptr || !*bufptr) {
+       /*
+        * In theory this could be recoded to do its I/O one character at a
+        * time, saving the buffer space.  In practice, this turns out to be
+        * quite hard to get completely right.  Try it and see.  If you
+        * succeed, don't forget to hack push_back() correspondingly.
+        */
+       size_t used;
+       size_t len;
+
+       do {
+           bufstart = 0;
+           used = 0;
+           do {
+               if (used + (LEXBUFSIZ / 4) >= allocated) {
+                   allocated += (allocated + LEXBUFSIZ);
+                   result = typeRealloc(char, allocated, result);
+                   if (result == 0)
+                       return (EOF);
+                   bufstart = result;
+               }
+               if (used == 0)
+                   _nc_curr_file_pos = ftell(yyin);
+
+               if (fgets(result + used, (int) (allocated - used), yyin) != 0) {
+                   bufstart = result;
+                   if (used == 0) {
+                       if (_nc_curr_line == 0
+                           && IS_TIC_MAGIC(result)) {
+                           _nc_err_abort("This is a compiled terminal description, not a source");
+                       }
+                       _nc_curr_line++;
+                       _nc_curr_col = 0;
+                   }
+               } else {
+                   if (used != 0)
+                       strcat(result, "\n");
+               }
+               if ((bufptr = bufstart) != 0) {
+                   used = strlen(bufptr);
+                   while (iswhite(*bufptr)) {
+                       if (*bufptr == '\t') {
+                           _nc_curr_col = (_nc_curr_col | 7) + 1;
+                       } else {
+                           _nc_curr_col++;
+                       }
+                       bufptr++;
+                   }
+
+                   /*
+                    * Treat a trailing <cr><lf> the same as a <newline> so we
+                    * can read files on OS/2, etc.
+                    */
+                   if ((len = strlen(bufptr)) > 1) {
+                       if (bufptr[len - 1] == '\n'
+                           && bufptr[len - 2] == '\r') {
+                           len--;
+                           bufptr[len - 1] = '\n';
+                           bufptr[len] = '\0';
+                       }
+                   }
+               } else {
+                   return (EOF);
+               }
+           } while (bufptr[len - 1] != '\n');  /* complete a line */
+       } while (result[0] == '#');     /* ignore comments */
+    } else if (*bufptr == '\t') {
+       _nc_curr_col = (_nc_curr_col | 7);
+    }
+
+    first_column = (bufptr == bufstart);
+    if (first_column)
+       had_newline = FALSE;
+
+    _nc_curr_col++;
+    the_char = *bufptr++;
+    return UChar(the_char);
+}
+
+static void
+push_back(char c)
+/* push a character back onto the input stream */
+{
+    if (bufptr == bufstart)
+       _nc_syserr_abort("Can't backspace off beginning of line");
+    *--bufptr = c;
+    _nc_curr_col--;
+}
+
+static long
+stream_pos(void)
+/* return our current character position in the input stream */
+{
+    return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0));
+}
+
+static bool
+end_of_stream(void)
+/* are we at end of input? */
+{
+    return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0'))
+           ? TRUE : FALSE);
+}
 
 /* Assume we may be looking at a termcap-style continuation */
 
 /* Assume we may be looking at a termcap-style continuation */
-static inline int eat_escaped_newline(int ch)
+static NCURSES_INLINE int
+eat_escaped_newline(int ch)
 {
 {
-       if (ch == '\\')
-               while ((ch = next_char()) == '\n'  ||  iswhite(ch))
-                       continue;
-       return ch;
+    if (ch == '\\')
+       while ((ch = next_char()) == '\n' || iswhite(ch))
+           continue;
+    return ch;
 }
 
 }
 
+#define TOK_BUF_SIZE MAX_ENTRY_SIZE
+
+#define OkToAdd() \
+       ((tok_ptr - tok_buf) < (TOK_BUF_SIZE - 2))
+
+#define AddCh(ch) \
+       *tok_ptr++ = (char) ch; \
+       *tok_ptr = '\0'
+
+static char *tok_buf;
+
 /*
  *     int
  *     get_token()
 /*
  *     int
  *     get_token()
@@ -124,316 +335,371 @@ static inline int eat_escaped_newline(int ch)
  *
  */
 
  *
  */
 
-int _nc_get_token(void)
+NCURSES_EXPORT(int)
+_nc_get_token(bool silent)
 {
 {
-static const char terminfo_punct[] = "@%&*!#";
-long           number;
-int            type;
-int            ch;
-char *         numchk;
-char           numbuf[80];
-unsigned       found;
-static char    buffer[MAX_ENTRY_SIZE];
-char           *ptr;
-int            dot_flag = FALSE;
-long           token_start;
-
-       if (pushtype != NO_PUSHBACK)
-       {
-           int retval = pushtype;
-
-           _nc_set_type(pushname);
-           DEBUG(3, ("pushed-back token: `%s', class %d",
-                     _nc_curr_token.tk_name, pushtype));
-
-           pushtype = NO_PUSHBACK;
+    static const char terminfo_punct[] = "@%&*!#";
+
+    char *after_list;
+    char *after_name;
+    char *numchk;
+    char *tok_ptr;
+    char *s;
+    char numbuf[80];
+    int ch;
+    int dot_flag = FALSE;
+    int type;
+    long number;
+    long token_start;
+    unsigned found;
+#ifdef TRACE
+    int old_line;
+    int old_col;
+#endif
+
+    if (pushtype != NO_PUSHBACK) {
+       int retval = pushtype;
+
+       _nc_set_type(pushname != 0 ? pushname : "");
+       DEBUG(3, ("pushed-back token: `%s', class %d",
+                 _nc_curr_token.tk_name, pushtype));
+
+       pushtype = NO_PUSHBACK;
+       if (pushname != 0)
            pushname[0] = '\0';
 
            pushname[0] = '\0';
 
-           /* currtok wasn't altered by _nc_push_token() */
-           return(retval);
+       /* currtok wasn't altered by _nc_push_token() */
+       return (retval);
+    }
+
+    if (end_of_stream()) {
+       yyin = 0;
+       next_char();            /* frees its allocated memory */
+       if (tok_buf != 0) {
+           if (_nc_curr_token.tk_name == tok_buf)
+               _nc_curr_token.tk_name = 0;
        }
        }
+       return (EOF);
+    }
 
 
-       if (end_of_stream())
-           return(EOF);
+  start_token:
+    token_start = stream_pos();
+    while ((ch = next_char()) == '\n' || iswhite(ch)) {
+       if (ch == '\n')
+           had_newline = TRUE;
+       continue;
+    }
 
 
-start_token:
-       token_start = stream_pos();
-       while ((ch = next_char()) == '\n'  ||  iswhite(ch))
-           continue;
+    ch = eat_escaped_newline(ch);
+    _nc_curr_token.tk_valstring = 0;
+
+#ifdef TRACE
+    old_line = _nc_curr_line;
+    old_col = _nc_curr_col;
+#endif
+    if (ch == EOF)
+       type = EOF;
+    else {
+       /* if this is a termcap entry, skip a leading separator */
+       if (separator == ':' && ch == ':')
+           ch = next_char();
+
+       if (ch == '.'
+#if NCURSES_EXT_FUNCS
+           && !_nc_disable_period
+#endif
+           ) {
+           dot_flag = TRUE;
+           DEBUG(8, ("dot-flag set"));
 
 
-       ch = eat_escaped_newline(ch);
+           while ((ch = next_char()) == '.' || iswhite(ch))
+               continue;
+       }
 
 
-       if (ch == EOF)
+       if (ch == EOF) {
            type = EOF;
            type = EOF;
-       else {
-           /* if this is a termcap entry, skip a leading separator */
-           if (separator == ':' && ch == ':')
-               ch = next_char();
+           goto end_of_token;
+       }
 
 
-           if (ch == '.') {
-                       dot_flag = TRUE;
-                       DEBUG(8, ("dot-flag set"));
+       /* have to make some punctuation chars legal for terminfo */
+       if (!isalnum(UChar(ch))
+#if NCURSES_EXT_FUNCS
+           && !(ch == '.' && _nc_disable_period)
+#endif
+           && !strchr(terminfo_punct, (char) ch)) {
+           if (!silent)
+               _nc_warning("Illegal character (expected alphanumeric or %s) - '%s'",
+                           terminfo_punct, unctrl(UChar(ch)));
+           _nc_panic_mode(separator);
+           goto start_token;
+       }
 
 
-                       while ((ch = next_char())=='.' || iswhite(ch))
-                           continue;
-           }
+       if (tok_buf == 0)
+           tok_buf = typeMalloc(char, TOK_BUF_SIZE);
 
 
-           if (ch == EOF) {
-               type = EOF;
-               goto end_of_token;
+#ifdef TRACE
+       old_line = _nc_curr_line;
+       old_col = _nc_curr_col;
+#endif
+       tok_ptr = tok_buf;
+       AddCh(ch);
+
+       if (first_column) {
+           _nc_comment_start = token_start;
+           _nc_comment_end = _nc_curr_file_pos;
+           _nc_start_line = _nc_curr_line;
+
+           _nc_syntax = ERR;
+           after_name = 0;
+           after_list = 0;
+           while ((ch = next_char()) != '\n') {
+               if (ch == EOF) {
+                   _nc_err_abort(MSG_NO_INPUTS);
+               } else if (ch == '|') {
+                   after_list = tok_ptr;
+                   if (after_name == 0)
+                       after_name = tok_ptr;
+               } else if (ch == ':' && last_char() != ',') {
+                   _nc_syntax = SYN_TERMCAP;
+                   separator = ':';
+                   break;
+               } else if (ch == ',') {
+                   _nc_syntax = SYN_TERMINFO;
+                   separator = ',';
+                   /*
+                    * If we did not see a '|', then we found a name with no
+                    * aliases or description.
+                    */
+                   if (after_name == 0)
+                       break;
+                   /*
+                    * If we see a comma, we assume this is terminfo unless we
+                    * subsequently run into a colon.  But we don't stop
+                    * looking for a colon until hitting a newline.  This
+                    * allows commas to be embedded in description fields of
+                    * either syntax.
+                    */
+               } else
+                   ch = eat_escaped_newline(ch);
+
+               if (OkToAdd()) {
+                   AddCh(ch);
+               } else {
+                   break;
+               }
            }
            }
-
-           /* have to make some punctuation chars legal for terminfo */
-           if (!isalnum(ch) && !strchr(terminfo_punct, (char)ch)) {
-                _nc_warning("Illegal character (expected alphanumeric or %s) - %s",
-                       terminfo_punct, _tracechar((chtype)ch));
-                _nc_panic_mode(separator);
-                goto start_token;
+           *tok_ptr = '\0';
+           if (_nc_syntax == ERR) {
+               /*
+                * Grrr...what we ought to do here is barf, complaining that
+                * the entry is malformed.  But because a couple of name fields
+                * in the 8.2 termcap file end with |\, we just have to assume
+                * it's termcap syntax.
+                */
+               _nc_syntax = SYN_TERMCAP;
+               separator = ':';
+           } else if (_nc_syntax == SYN_TERMINFO) {
+               /* throw away trailing /, *$/ */
+               for (--tok_ptr;
+                    iswhite(*tok_ptr) || *tok_ptr == ',';
+                    tok_ptr--)
+                   continue;
+               tok_ptr[1] = '\0';
            }
 
            }
 
-           ptr = buffer;
-           *(ptr++) = ch;
-
-           if (first_column) {
-                       char    *desc;
-
-                       _nc_comment_start = token_start;
-                       _nc_comment_end = _nc_curr_file_pos;
-                       _nc_start_line = _nc_curr_line;
-
-                       _nc_syntax = ERR;
-                       while ((ch = next_char()) != '\n')
-                       {
-                           if (ch == EOF)
-                               _nc_err_abort("premature EOF");
-                           else if (ch == ':' && last_char() != ',')
-                           {
-                               _nc_syntax = SYN_TERMCAP;
-                               separator = ':';
-                               break;
-                           }
-                           else if (ch == ',')
-                           {
-                               _nc_syntax = SYN_TERMINFO;
-                               separator = ',';
-                               /*
-                                * Fall-through here is not an accident.
-                                * The idea is that if we see a comma, we
-                                * figure this is terminfo unless we
-                                * subsequently run into a colon -- but
-                                * we don't stop looking for that colon until
-                                * hitting a newline.  This allows commas to
-                                * be embedded in description fields of
-                                * either syntax.
-                                */
-                               /* FALLTHRU */
-                           }
-                           else
-                               ch = eat_escaped_newline(ch);
-
-                           *ptr++ = ch;
-                       }
-                       ptr[0] = '\0';
-                       if (_nc_syntax == ERR)
-                       {
-                           /*
-                            * Grrr...what we ought to do here is barf,
-                            * complaining that the entry is malformed.
-                            * But because a couple of name fields in the
-                            * 8.2 termcap file end with |\, we just have
-                            * to assume it's termcap syntax.
-                            */
-                           _nc_syntax = SYN_TERMCAP;
-                           separator = ':';
-                       }
-                       else if (_nc_syntax == SYN_TERMINFO)
-                       {
-                           /* throw away trailing /, *$/ */
-                           for (--ptr; iswhite(*ptr) || *ptr == ','; ptr--)
-                               continue;
-                           ptr[1] = '\0';
-                       }
-
-                       /*
-                        * This is the soonest we have the terminal name
-                        * fetched.  Set up for following warning messages.
-                        */
-                       ptr = strchr(buffer, '|');
-                       if (ptr == (char *)NULL)
-                           ptr = buffer + strlen(buffer);
-                       ch = *ptr;
-                       *ptr = '\0';
-                       _nc_set_type(buffer);
-                       *ptr = ch;
-
-                       /*
-                        * Compute the boundary between the aliases and the
-                        * description field for syntax-checking purposes.
-                        */
-                       desc = strrchr(buffer, '|');
-                       if (desc) {
-                           if (*desc == '\0')
-                               _nc_warning("empty longname field");
-                           else if (strchr(desc, ' ') == (char *)NULL)
-                               _nc_warning("older tic versions may treat the description field as an alias");
-                       }
-                       if (!desc)
-                           desc = buffer + strlen(buffer);
-
-                       /*
-                        * Whitespace in a name field other than the long name
-                        * can confuse rdist and some termcap tools.  Slashes
-                        * are a no-no.  Other special characters can be
-                        * dangerous due to shell expansion.
-                        */
-                       for (ptr = buffer; ptr < desc; ptr++)
-                       {
-                           if (isspace(*ptr))
-                           {
-                               _nc_warning("whitespace in name or alias field");
-                               break;
-                           }
-                           else if (*ptr == '/')
-                           {
-                               _nc_warning("slashes aren't allowed in names or aliases");
-                               break;
-                           }
-                           else if (strchr("$[]!*?", *ptr))
-                           {
-                               _nc_warning("dubious character `%c' in name or alias field", *ptr);
-                               break;
-                           }
-                       }
-
-                       ptr = buffer;
+           /*
+            * This is the soonest we have the terminal name fetched.  Set up
+            * for following warning messages.  If there's no '|', then there
+            * is no description.
+            */
+           if (after_name != 0) {
+               ch = *after_name;
+               *after_name = '\0';
+               _nc_set_type(tok_buf);
+               *after_name = (char) ch;
+           }
 
 
-                       _nc_curr_token.tk_name = buffer;
-                       type = NAMES;
+           /*
+            * Compute the boundary between the aliases and the description
+            * field for syntax-checking purposes.
+            */
+           if (after_list != 0) {
+               if (!silent) {
+                   if (*after_list == '\0')
+                       _nc_warning("empty longname field");
+                   else if (strchr(after_list, ' ') == 0)
+                       _nc_warning("older tic versions may treat the description field as an alias");
+               }
            } else {
            } else {
-                       while ((ch = next_char()) != EOF) {
-                               if (!isalnum(ch)) {
-                                       if (_nc_syntax == SYN_TERMINFO) {
-                                               if (ch != '_')
-                                                       break;
-                                       } else { /* allow ';' for "k;" */
-                                               if (ch != ';')
-                                                       break;
-                                       }
-                               }
-                               *(ptr++) = ch;
-                       }
-
-                       *ptr++ = '\0';
-                       switch (ch) {
-                       case ',':
-                       case ':':
-                               if (ch != separator)
-                                       _nc_err_abort("Separator inconsistent with syntax");
-                               _nc_curr_token.tk_name = buffer;
-                               type = BOOLEAN;
-                               break;
-                       case '@':
-                               if ((ch = next_char()) != separator)
-                                       _nc_warning("Missing separator after `%s', have %s",
-                                               buffer, _tracechar((chtype)ch));
-                               _nc_curr_token.tk_name = buffer;
-                               type = CANCEL;
-                               break;
-
-                       case '#':
-                               found  = 0;
-                               while (isalnum(ch = next_char())) {
-                                       numbuf[found++] = ch;
-                                       if (found >= sizeof(numbuf)-1)
-                                               break;
-                               }
-                               numbuf[found] = '\0';
-                               number = strtol(numbuf, &numchk, 0);
-                               if (numchk == numbuf)
-                                       _nc_warning("no value given for `%s'", buffer);
-                               if ((*numchk != '\0') || (ch != separator))
-                                       _nc_warning("Missing separator");
-                               _nc_curr_token.tk_name = buffer;
-                               _nc_curr_token.tk_valnumber = number;
-                               type = NUMBER;
-                               break;
-
-                       case '=':
-                               ch = _nc_trans_string(ptr);
-                               if (ch != separator)
-                                       _nc_warning("Missing separator");
-                               _nc_curr_token.tk_name = buffer;
-                               _nc_curr_token.tk_valstring = ptr;
-                               type = STRING;
-                               break;
-
-                       case EOF:
-                               type = EOF;
-                               break;
-                       default:
-                               /* just to get rid of the compiler warning */
-                               type = UNDEF;
-                               _nc_warning("Illegal character - %s",
-                                       _tracechar((chtype)ch));
-                       }
-               } /* end else (first_column == FALSE) */
-       } /* end else (ch != EOF) */
-
-end_of_token:
+               after_list = tok_buf + strlen(tok_buf);
+               DEBUG(1, ("missing description"));
+           }
 
 
-#ifdef TRACE
-       if (dot_flag == TRUE)
-           DEBUG(8, ("Commented out "));
-
-       if (_nc_tracing & TRACE_IEVENT)
-       {
-           fprintf(stderr, "Token: ");
-           switch (type)
-           {
-               case BOOLEAN:
-                   fprintf(stderr, "Boolean; name='%s'\n",
-                           _nc_curr_token.tk_name);
+           /*
+            * Whitespace in a name field other than the long name can confuse
+            * rdist and some termcap tools.  Slashes are a no-no.  Other
+            * special characters can be dangerous due to shell expansion.
+            */
+           for (s = tok_buf; s < after_list; ++s) {
+               if (isspace(UChar(*s))) {
+                   if (!silent)
+                       _nc_warning("whitespace in name or alias field");
                    break;
                    break;
-
-               case NUMBER:
-                   fprintf(stderr, "Number;  name='%s', value=%d\n",
-                           _nc_curr_token.tk_name,
-                           _nc_curr_token.tk_valnumber);
+               } else if (*s == '/') {
+                   if (!silent)
+                       _nc_warning("slashes aren't allowed in names or aliases");
                    break;
                    break;
-
-               case STRING:
-                   fprintf(stderr, "String;  name='%s', value=%s\n",
-                           _nc_curr_token.tk_name,
-                           _nc_visbuf(_nc_curr_token.tk_valstring));
+               } else if (strchr("$[]!*?", *s)) {
+                   if (!silent)
+                       _nc_warning("dubious character `%c' in name or alias field", *s);
                    break;
                    break;
+               }
+           }
 
 
-               case CANCEL:
-                   fprintf(stderr, "Cancel; name='%s'\n",
-                           _nc_curr_token.tk_name);
+           _nc_curr_token.tk_name = tok_buf;
+           type = NAMES;
+       } else {
+           if (had_newline && _nc_syntax == SYN_TERMCAP) {
+               _nc_warning("Missing backslash before newline");
+               had_newline = FALSE;
+           }
+           while ((ch = next_char()) != EOF) {
+               if (!isalnum(UChar(ch))) {
+                   if (_nc_syntax == SYN_TERMINFO) {
+                       if (ch != '_')
+                           break;
+                   } else {    /* allow ';' for "k;" */
+                       if (ch != ';')
+                           break;
+                   }
+               }
+               if (OkToAdd()) {
+                   AddCh(ch);
+               } else {
+                   ch = EOF;
                    break;
                    break;
+               }
+           }
 
 
-               case NAMES:
+           *tok_ptr++ = '\0';  /* separate name/value in buffer */
+           switch (ch) {
+           case ',':
+           case ':':
+               if (ch != separator)
+                   _nc_err_abort("Separator inconsistent with syntax");
+               _nc_curr_token.tk_name = tok_buf;
+               type = BOOLEAN;
+               break;
+           case '@':
+               if ((ch = next_char()) != separator && !silent)
+                   _nc_warning("Missing separator after `%s', have %s",
+                               tok_buf, unctrl(UChar(ch)));
+               _nc_curr_token.tk_name = tok_buf;
+               type = CANCEL;
+               break;
 
 
-                   fprintf(stderr, "Names; value='%s'\n",
-                           _nc_curr_token.tk_name);
-                   break;
+           case '#':
+               found = 0;
+               while (isalnum(ch = next_char())) {
+                   numbuf[found++] = (char) ch;
+                   if (found >= sizeof(numbuf) - 1)
+                       break;
+               }
+               numbuf[found] = '\0';
+               number = strtol(numbuf, &numchk, 0);
+               if (!silent) {
+                   if (numchk == numbuf)
+                       _nc_warning("no value given for `%s'", tok_buf);
+                   if ((*numchk != '\0') || (ch != separator))
+                       _nc_warning("Missing separator");
+               }
+               _nc_curr_token.tk_name = tok_buf;
+               _nc_curr_token.tk_valnumber = (int) number;
+               type = NUMBER;
+               break;
 
 
-               case EOF:
-                   fprintf(stderr, "End of file\n");
-                   break;
+           case '=':
+               ch = _nc_trans_string(tok_ptr, tok_buf + TOK_BUF_SIZE);
+               if (!silent && ch != separator)
+                   _nc_warning("Missing separator");
+               _nc_curr_token.tk_name = tok_buf;
+               _nc_curr_token.tk_valstring = tok_ptr;
+               type = STRING;
+               break;
 
 
-               default:
-                   _nc_warning("Bad token type");
+           case EOF:
+               type = EOF;
+               break;
+           default:
+               /* just to get rid of the compiler warning */
+               type = UNDEF;
+               if (!silent)
+                   _nc_warning("Illegal character - '%s'", unctrl(UChar(ch)));
            }
            }
+       }                       /* end else (first_column == FALSE) */
+    }                          /* end else (ch != EOF) */
+
+  end_of_token:
+
+#ifdef TRACE
+    if (dot_flag == TRUE)
+       DEBUG(8, ("Commented out "));
+
+    if (_nc_tracing >= DEBUG_LEVEL(8)) {
+       _tracef("parsed %d.%d to %d.%d",
+               old_line, old_col,
+               _nc_curr_line, _nc_curr_col);
+    }
+    if (_nc_tracing >= DEBUG_LEVEL(7)) {
+       switch (type) {
+       case BOOLEAN:
+           _tracef("Token: Boolean; name='%s'",
+                   _nc_curr_token.tk_name);
+           break;
+
+       case NUMBER:
+           _tracef("Token: Number;  name='%s', value=%d",
+                   _nc_curr_token.tk_name,
+                   _nc_curr_token.tk_valnumber);
+           break;
+
+       case STRING:
+           _tracef("Token: String;  name='%s', value=%s",
+                   _nc_curr_token.tk_name,
+                   _nc_visbuf(_nc_curr_token.tk_valstring));
+           break;
+
+       case CANCEL:
+           _tracef("Token: Cancel; name='%s'",
+                   _nc_curr_token.tk_name);
+           break;
+
+       case NAMES:
+
+           _tracef("Token: Names; value='%s'",
+                   _nc_curr_token.tk_name);
+           break;
+
+       case EOF:
+           _tracef("Token: End of file");
+           break;
+
+       default:
+           _nc_warning("Bad token type");
        }
        }
+    }
 #endif
 
 #endif
 
-       if (dot_flag == TRUE)           /* if commented out, use the next one */
-           type = _nc_get_token();
+    if (dot_flag == TRUE)      /* if commented out, use the next one */
+       type = _nc_get_token(silent);
 
 
-       DEBUG(3, ("token: `%s', class %d", _nc_curr_token.tk_name, type));
+    DEBUG(3, ("token: `%s', class %d",
+             ((_nc_curr_token.tk_name != 0)
+              ? _nc_curr_token.tk_name
+              : "<null>"),
+             type));
 
 
-       return(type);
+    return (type);
 }
 
 /*
 }
 
 /*
@@ -455,121 +721,181 @@ end_of_token:
  *
  */
 
  *
  */
 
-char
-_nc_trans_string(char *ptr)
+NCURSES_EXPORT(int)
+_nc_trans_string(char *ptr, char *last)
 {
 {
-int    count = 0;
-int    number;
-int    i, c;
-chtype ch, last_ch = '\0';
-bool   ignored = FALSE;
-
-       while ((ch = c = next_char()) != (chtype)separator && c != EOF) {
-           if ((_nc_syntax == SYN_TERMCAP) && c == '\n')
-               break;
-           if (ch == '^' && last_ch != '%') {
-               ch = c = next_char();
-               if (c == EOF)
-                   _nc_err_abort("Premature EOF");
-
-               if (! (is7bits(ch) && isprint(ch))) {
-                   _nc_warning("Illegal ^ character - %s",
-                       _tracechar((unsigned char)ch));
-               }
-               if (ch == '?') {
-                   *(ptr++) = '\177';
-               } else {
-                   if ((ch &= 037) == 0)
-                       ch = 128;
-                   *(ptr++) = (char)(ch);
+    int count = 0;
+    int number = 0;
+    int i, c;
+    int last_ch = '\0';
+    bool ignored = FALSE;
+    bool long_warning = FALSE;
+
+    while ((c = next_char()) != separator && c != EOF) {
+       if (ptr >= (last - 1)) {
+           if (c != EOF) {
+               while ((c = next_char()) != separator && c != EOF) {
+                   ;
                }
            }
                }
            }
-           else if (ch == '\\') {
-               ch = c = next_char();
-               if (c == EOF)
-                   _nc_err_abort("Premature EOF");
-
-               if (ch >= '0'  &&  ch <= '7') {
-                   number = ch - '0';
-                   for (i=0; i < 2; i++) {
-                       ch = c = next_char();
-                       if (c == EOF)
-                           _nc_err_abort("Premature EOF");
-
-                       if (c < '0'  ||  c > '7') {
-                           if (isdigit(c)) {
+           break;
+       }
+       if ((_nc_syntax == SYN_TERMCAP) && c == '\n')
+           break;
+       if (c == '^' && last_ch != '%') {
+           c = next_char();
+           if (c == EOF)
+               _nc_err_abort(MSG_NO_INPUTS);
+
+           if (!(is7bits(c) && isprint(c))) {
+               _nc_warning("Illegal ^ character - '%s'", unctrl(UChar(c)));
+           }
+           if (c == '?' && (_nc_syntax != SYN_TERMCAP)) {
+               *(ptr++) = '\177';
+               if (_nc_tracing)
+                   _nc_warning("Allow ^? as synonym for \\177");
+           } else {
+               if ((c &= 037) == 0)
+                   c = 128;
+               *(ptr++) = (char) (c);
+           }
+       } else if (c == '\\') {
+           bool strict_bsd = ((_nc_syntax == SYN_TERMCAP) && _nc_strict_bsd);
+
+           c = next_char();
+           if (c == EOF)
+               _nc_err_abort(MSG_NO_INPUTS);
+
+#define isoctal(c) ((c) >= '0' && (c) <= '7')
+
+           if (isoctal(c) || (strict_bsd && isdigit(c))) {
+               number = c - '0';
+               for (i = 0; i < 2; i++) {
+                   c = next_char();
+                   if (c == EOF)
+                       _nc_err_abort(MSG_NO_INPUTS);
+
+                   if (!isoctal(c)) {
+                       if (isdigit(c)) {
+                           if (!strict_bsd) {
                                _nc_warning("Non-octal digit `%c' in \\ sequence", c);
                                /* allow the digit; it'll do less harm */
                                _nc_warning("Non-octal digit `%c' in \\ sequence", c);
                                /* allow the digit; it'll do less harm */
-                           } else {
-                               push_back((char)c);
-                               break;
                            }
                            }
+                       } else {
+                           push_back((char) c);
+                           break;
                        }
                        }
-
-                       number = number * 8 + c - '0';
                    }
 
                    }
 
-                   if (number == 0)
-                       number = 0200;
-                   *(ptr++) = (char) number;
-               } else {
-                   switch (c) {
-                       case 'E':
-                       case 'e':       *(ptr++) = '\033';      break;
-
-                       case 'a':       *(ptr++) = '\007';      break;
-
-                       case 'l':
-                       case 'n':       *(ptr++) = '\n';        break;
+                   number = number * 8 + c - '0';
+               }
 
 
-                       case 'r':       *(ptr++) = '\r';        break;
+               number = UChar(number);
+               if (number == 0 && !strict_bsd)
+                   number = 0200;
+               *(ptr++) = (char) number;
+           } else {
+               switch (c) {
+               case 'E':
+                   *(ptr++) = '\033';
+                   break;
 
 
-                       case 'b':       *(ptr++) = '\010';      break;
+               case 'n':
+                   *(ptr++) = '\n';
+                   break;
 
 
-                       case 's':       *(ptr++) = ' ';         break;
+               case 'r':
+                   *(ptr++) = '\r';
+                   break;
 
 
-                       case 'f':       *(ptr++) = '\014';      break;
+               case 'b':
+                   *(ptr++) = '\010';
+                   break;
 
 
-                       case 't':       *(ptr++) = '\t';        break;
+               case 'f':
+                   *(ptr++) = '\014';
+                   break;
 
 
-                       case '\\':      *(ptr++) = '\\';        break;
+               case 't':
+                   *(ptr++) = '\t';
+                   break;
 
 
-                       case '^':       *(ptr++) = '^';         break;
+               case '\\':
+                   *(ptr++) = '\\';
+                   break;
 
 
-                       case ',':       *(ptr++) = ',';         break;
+               case '^':
+                   *(ptr++) = '^';
+                   break;
 
 
-                       case ':':       *(ptr++) = ':';         break;
+               case ',':
+                   *(ptr++) = ',';
+                   break;
 
 
-                       case '\n':
-                           continue;
+               case '\n':
+                   continue;
 
 
+               default:
+                   if ((_nc_syntax == SYN_TERMINFO) || !_nc_strict_bsd) {
+                       switch (c) {
+                       case 'a':
+                           c = '\007';
+                           break;
+                       case 'e':
+                           c = '\033';
+                           break;
+                       case 'l':
+                           c = '\n';
+                           break;
+                       case 's':
+                           c = ' ';
+                           break;
+                       case ':':
+                           c = ':';
+                           break;
                        default:
                        default:
-                           _nc_warning("Illegal character %s in \\ sequence",
-                                   _tracechar((unsigned char)ch));
-                           *(ptr++) = (char)ch;
-                   } /* endswitch (ch) */
-               } /* endelse (ch < '0' ||  ch > '7') */
-           } /* end else if (ch == '\\') */
-           else if (ch == '\n' && (_nc_syntax == SYN_TERMINFO)) {
-               /* newlines embedded in a terminfo string are ignored */
-               ignored = TRUE;
-           } else {
-               *(ptr++) = (char)ch;
-           }
+                           _nc_warning("Illegal character '%s' in \\ sequence",
+                                       unctrl(UChar(c)));
+                           break;
+                       }
+                   }
+                   /* FALLTHRU */
+               case '|':
+                   *(ptr++) = (char) c;
+               }               /* endswitch (c) */
+           }                   /* endelse (c < '0' ||  c > '7') */
+       }
+       /* end else if (c == '\\') */
+       else if (c == '\n' && (_nc_syntax == SYN_TERMINFO)) {
+           /*
+            * Newlines embedded in a terminfo string are ignored, provided
+            * that the next line begins with whitespace.
+            */
+           ignored = TRUE;
+       } else {
+           *(ptr++) = (char) c;
+       }
 
 
-           if (!ignored) {
-               last_ch = ch;
-               count ++;
+       if (!ignored) {
+           if (_nc_curr_col <= 1) {
+               push_back((char) c);
+               c = '\n';
+               break;
            }
            }
-           ignored = FALSE;
+           last_ch = c;
+           count++;
+       }
+       ignored = FALSE;
 
 
-           if (count > MAXCAPLEN)
-               _nc_warning("Very long string found.  Missing separator?");
-       } /* end while */
+       if (count > MAXCAPLEN && !long_warning) {
+           _nc_warning("Very long string found.  Missing separator?");
+           long_warning = TRUE;
+       }
+    }                          /* end while */
 
 
-       *ptr = '\0';
+    *ptr = '\0';
 
 
-       return(ch);
+    return (c);
 }
 
 /*
 }
 
 /*
@@ -579,178 +905,53 @@ bool     ignored = FALSE;
  *     get_token() call.
  */
 
  *     get_token() call.
  */
 
-void _nc_push_token(int tokclass)
+NCURSES_EXPORT(void)
+_nc_push_token(int tokclass)
 {
     /*
 {
     /*
-     * This implementation is kind of bogus, it will fail if we ever do
-     * more than one pushback at a time between get_token() calls.  It
-     * relies on the fact that curr_tok is static storage that nothing
-     * but get_token() touches.
+     * This implementation is kind of bogus, it will fail if we ever do more
+     * than one pushback at a time between get_token() calls.  It relies on the
+     * fact that _nc_curr_token is static storage that nothing but
+     * _nc_get_token() touches.
      */
     pushtype = tokclass;
      */
     pushtype = tokclass;
+    if (pushname == 0)
+       pushname = typeMalloc(char, MAX_NAME_SIZE + 1);
     _nc_get_type(pushname);
 
     DEBUG(3, ("pushing token: `%s', class %d",
     _nc_get_type(pushname);
 
     DEBUG(3, ("pushing token: `%s', class %d",
-             _nc_curr_token.tk_name, pushtype));
+             ((_nc_curr_token.tk_name != 0)
+              ? _nc_curr_token.tk_name
+              : "<null>"),
+             pushtype));
 }
 
 /*
  * Panic mode error recovery - skip everything until a "ch" is found.
  */
 }
 
 /*
  * Panic mode error recovery - skip everything until a "ch" is found.
  */
-void _nc_panic_mode(char ch)
-{
-       int c;
-
-       for (;;) {
-               c = next_char();
-               if (c == ch)
-                       return;
-               if (c == EOF)
-                       return;
-       }
-}
-
-/*****************************************************************************
- *
- * Character-stream handling
- *
- *****************************************************************************/
-
-#define LEXBUFSIZ      1024
-
-static char *bufptr;           /* otherwise, the input buffer pointer */
-static char *bufstart;         /* start of buffer so we can compute offsets */
-static FILE *yyin;             /* scanner's input file descriptor */
-
-/*
- *     _nc_reset_input()
- *
- *     Resets the input-reading routines.  Used on initialization,
- *     or after a seek has been done.  Exactly one argument must be
- *     non-null.
- */
-
-void _nc_reset_input(FILE *fp, char *buf)
-{
-       pushtype = NO_PUSHBACK;
-       pushname[0] = '\0';
-       yyin = fp;
-       bufstart = bufptr = buf;
-       _nc_curr_file_pos = 0L;
-       if (fp != 0)
-               _nc_curr_line = 0;
-       _nc_curr_col = 0;
-}
-
-/*
- *     int last_char()
- *
- *     Returns the final nonblank character on the current input buffer
- */
-static int
-last_char(void)
+NCURSES_EXPORT(void)
+_nc_panic_mode(char ch)
 {
 {
-       size_t len = strlen(bufptr);
-       while (len--) {
-               if (!isspace(bufptr[len]))
-                       return bufptr[len];
-       }
-       return 0;
+    int c;
+
+    for (;;) {
+       c = next_char();
+       if (c == ch)
+           return;
+       if (c == EOF)
+           return;
+    }
 }
 
 }
 
-/*
- *     int next_char()
- *
- *     Returns the next character in the input stream.  Comments and leading
- *     white space are stripped.
- *
- *     The global state variable 'firstcolumn' is set TRUE if the character
- *     returned is from the first column of the input line.
- *
- *     The global variable _nc_curr_line is incremented for each new line.
- *     The global variable _nc_curr_file_pos is set to the file offset of the
- *     beginning of each line.
- */
-
-static int
-next_char(void)
+#if NO_LEAKS
+NCURSES_EXPORT(void)
+_nc_comp_scan_leaks(void)
 {
 {
-    if (!yyin)
-    {
-       if (*bufptr == '\0')
-           return(EOF);
-       if (*bufptr == '\n') {
-           _nc_curr_line++;
-           _nc_curr_col = 0;
-       }
+    if (pushname != 0) {
+       FreeAndNull(pushname);
     }
     }
-    else if (!bufptr || !*bufptr)
-    {
-       /*
-        * In theory this could be recoded to do its I/O one
-        * character at a time, saving the buffer space.  In
-        * practice, this turns out to be quite hard to get
-        * completely right.  Try it and see.  If you succeed,
-        * don't forget to hack push_back() correspondingly.
-        */
-       static char line[LEXBUFSIZ];
-       size_t len;
-
-       do {
-              _nc_curr_file_pos = ftell(yyin);
-
-              if ((bufstart = fgets(line, LEXBUFSIZ, yyin)) != NULL) {
-                  _nc_curr_line++;
-                  _nc_curr_col = 0;
-              }
-              bufptr = bufstart;
-          } while
-              (bufstart != NULL && line[0] == '#');
-
-       if (bufstart == NULL)
-           return (EOF);
-
-       while (iswhite(*bufptr))
-           bufptr++;
-
-       /*
-        * Treat a trailing <cr><lf> the same as a <newline> so we can read
-        * files on OS/2, etc.
-        */
-       if ((len = strlen(bufptr)) > 1) {
-           if (bufptr[len-1] == '\n'
-            && bufptr[len-2] == '\r') {
-               bufptr[len-2] = '\n';
-               bufptr[len-1] = '\0';
-           }
-       }
+    if (tok_buf != 0) {
+       FreeAndNull(tok_buf);
     }
     }
-
-    first_column = (bufptr == bufstart);
-
-    _nc_curr_col++;
-    return(*bufptr++);
-}
-
-static void push_back(char c)
-/* push a character back onto the input stream */
-{
-    if (bufptr == bufstart)
-           _nc_syserr_abort("Can't backspace off beginning of line");
-    *--bufptr = c;
-}
-
-static long stream_pos(void)
-/* return our current character position in the input stream */
-{
-    return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0));
-}
-
-static bool end_of_stream(void)
-/* are we at end of input? */
-{
-    return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0'))
-           ? TRUE : FALSE);
 }
 }
-
-/* comp_scan.c ends here */
+#endif