comp_scan.c

ncurses 库 可能有用酒用 没用就算了 我觉得还可以用

/****************************************************************************
 * Copyright (c) 1998-2004,2005 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            *
 * "Software"), to deal in the Software without restriction, including      *
 * without limitation the rights to use, copy, modify, merge, publish,      *
 * distribute, distribute with modifications, sublicense, and/or sell       *
 * copies of the Software, and to permit persons to whom the Software is    *
 * furnished to do so, subject to the following conditions:                 *
 *                                                                          *
 * The above copyright notice and this permission notice shall be included  *
 * in all copies or substantial portions of the Software.                   *
 *                                                                          *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS  *
 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF               *
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.   *
 * IN NO EVENT SHALL THE ABOVE COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,   *
 * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR    *
 * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR    *
 * THE USE OR OTHER DEALINGS IN THE SOFTWARE.                               *
 *                                                                          *
 * Except as contained in this notice, the name(s) of the above copyright   *
 * holders shall not be used in advertising or otherwise to promote the     *
 * sale, use or other dealings in this Software without prior written       *
 * authorization.                                                           *
 ****************************************************************************/

/****************************************************************************
 *  Author: Zeyd M. Ben-Halim <[e-mail]> 1992,1995               *
 *     and: Eric S. Raymond <[e-mail]>                         *
 *     and: Thomas E. Dickey                        1996 on                 *
 ****************************************************************************/

/*
 *	comp_scan.c --- Lexical scanner for terminfo compiler.
 *
 *	_nc_reset_input()
 *	_nc_get_token()
 *	_nc_panic_mode()
 *	int _nc_syntax;
 *	int _nc_curr_line;
 *	long _nc_curr_file_pos;
 *	long _nc_comment_start;
 *	long _nc_comment_end;
 */

#include <curses.priv.h>

#include <ctype.h>
#include <term_entry.h>
#include <tic.h>

MODULE_ID("$Id: comp_scan.c,v 1.76 2005/06/04 22:04:45 tom Exp $")

/*
 * Maximum length of string capability we'll accept before raising an error.
 * Yes, there is a real capability in /etc/termcap this long, an "is".
 */
#define MAXCAPLEN	600

#define iswhite(ch)	(ch == ' '  ||  ch == '\t')

NCURSES_EXPORT_VAR(int)
_nc_syntax = 0;			/* termcap or terminfo? */
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
};

/*****************************************************************************
 *
 * Token-grabbing machinery
 *
 *****************************************************************************/

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 *pushname;

#if NCURSES_EXT_FUNCS
NCURSES_EXPORT_VAR(bool)
_nc_disable_period = FALSE;	/* used by tic -a option */
#endif

/*****************************************************************************
 *
 * 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);
		}
		if (used == 0)
		    _nc_curr_file_pos = ftell(yyin);

		if (fgets(result + used, (int) (allocated - used), yyin) != 0) {
		    bufstart = result;
		    if (used == 0) {
			_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 */
static inline int
eat_escaped_newline(int ch)
{
    if (ch == '\\')
	while ((ch = next_char()) == '\n' || iswhite(ch))
	    continue;
    return ch;
}

/*
 *	int
 *	get_token()
 *
 *	Scans the input for the next token, storing the specifics in the
 *	global structure 'curr_token' and returning one of the following:
 *
 *		NAMES		A line beginning in column 1.  'name'
 *				will be set to point to everything up to but
 *				not including the first separator on the line.
 *		BOOLEAN		An entry consisting of a name followed by
 *				a separator.  'name' will be set to point to
 *				the name of the capability.
 *		NUMBER		An entry of the form
 *					name#digits,
 *				'name' will be set to point to the capability
 *				name and 'valnumber' to the number given.
 *		STRING		An entry of the form
 *					name=characters,
 *				'name' is set to the capability name and
 *				'valstring' to the string of characters, with
 *				input translations done.
 *		CANCEL		An entry of the form
 *					name@,
 *				'name' is set to the capability name and
 *				'valnumber' to -1.
 *		EOF		The end of the file has been reached.
 *
 *	A `separator' is either a comma or a semicolon, depending on whether
 *	we are in termcap or terminfo mode.
 *
 */

NCURSES_EXPORT(int)
_nc_get_token(bool silent)
{
    static const char terminfo_punct[] = "@%&*!#";
    static char *buffer;

    char *after_list;
    char *after_name;
    char *numchk;
    char *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';

	/* currtok wasn't altered by _nc_push_token() */
	return (retval);
    }

    if (end_of_stream()) {
	yyin = 0;
	next_char();		/* frees its allocated memory */
	if (buffer != 0) {
	    if (_nc_curr_token.tk_name == buffer)
		_nc_curr_token.tk_name = 0;
	    FreeAndNull(buffer);
	}
	return (EOF);
    }

  start_token:
    token_start = stream_pos();
    while ((ch = next_char()) == '\n' || iswhite(ch)) {
	if (ch == '\n')
	    had_newline = TRUE;
	continue;
    }

    ch = eat_escaped_newline(ch);

#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"));

	    while ((ch = next_char()) == '.' || iswhite(ch))
		continue;
	}

	if (ch == EOF) {
	    type = EOF;
	    goto end_of_token;
	}

	/* 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((chtype) ch));
	    _nc_panic_mode(separator);
	    goto start_token;
	}

	if (buffer == 0)
	    buffer = typeMalloc(char, MAX_ENTRY_SIZE);

#ifdef TRACE
	old_line = _nc_curr_line;
	old_col = _nc_curr_col;
#endif
	ptr = buffer;
	*(ptr++) = 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 = ptr;
		    if (after_name == 0)
			after_name = ptr;
		} else if (ch == ':' && last_char() != ',') {
		    _nc_syntax = SYN_TERMCAP;
		    separator = ':';
		    break;