params.c

来自「samba-3.0.22.tar.gz 编译smb服务器的源码」· C语言 代码 · 共 605 行 · 第 1/2 页

/* -------------------------------------------------------------------------- **
 * Microsoft Network Services for Unix, AKA., Andrew Tridgell's SAMBA.
 *
 * This module Copyright (C) 1990-1998 Karl Auer
 *
 * Rewritten almost completely by Christopher R. Hertel
 * at the University of Minnesota, September, 1997.
 * This module Copyright (C) 1997-1998 by the University of Minnesota
 * -------------------------------------------------------------------------- **
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *
 * -------------------------------------------------------------------------- **
 *
 * Module name: params
 *
 * -------------------------------------------------------------------------- **
 *
 *  This module performs lexical analysis and initial parsing of a
 *  Windows-like parameter file.  It recognizes and handles four token
 *  types:  section-name, parameter-name, parameter-value, and
 *  end-of-file.  Comments and line continuation are handled
 *  internally.
 *
 *  The entry point to the module is function pm_process().  This
 *  function opens the source file, calls the Parse() function to parse
 *  the input, and then closes the file when either the EOF is reached
 *  or a fatal error is encountered.
 *
 *  A sample parameter file might look like this:
 *
 *  [section one]
 *  parameter one = value string
 *  parameter two = another value
 *  [section two]
 *  new parameter = some value or t'other
 *
 *  The parameter file is divided into sections by section headers:
 *  section names enclosed in square brackets (eg. [section one]).
 *  Each section contains parameter lines, each of which consist of a
 *  parameter name and value delimited by an equal sign.  Roughly, the
 *  syntax is:
 *
 *    <file>            :==  { <section> } EOF
 *
 *    <section>         :==  <section header> { <parameter line> }
 *
 *    <section header>  :==  '[' NAME ']'
 *
 *    <parameter line>  :==  NAME '=' VALUE '\n'
 *
 *  Blank lines and comment lines are ignored.  Comment lines are lines
 *  beginning with either a semicolon (';') or a pound sign ('#').
 *
 *  All whitespace in section names and parameter names is compressed
 *  to single spaces.  Leading and trailing whitespace is stipped from
 *  both names and values.
 *
 *  Only the first equals sign in a parameter line is significant.
 *  Parameter values may contain equals signs, square brackets and
 *  semicolons.  Internal whitespace is retained in parameter values,
 *  with the exception of the '\r' character, which is stripped for
 *  historic reasons.  Parameter names may not start with a left square
 *  bracket, an equal sign, a pound sign, or a semicolon, because these
 *  are used to identify other tokens.
 *
 * -------------------------------------------------------------------------- **
 */

#include "includes.h"

extern BOOL in_client;

/* -------------------------------------------------------------------------- **
 * Constants...
 */

#define BUFR_INC 1024


/* -------------------------------------------------------------------------- **
 * Variables...
 *
 *  DEBUGLEVEL  - The ubiquitous DEBUGLEVEL.  This determines which DEBUG()
 *                messages will be produced.
 *  bufr        - pointer to a global buffer.  This is probably a kludge,
 *                but it was the nicest kludge I could think of (for now).
 *  bSize       - The size of the global buffer <bufr>.
 */

static char *bufr  = NULL;
static int   bSize = 0;

/* we can't use FILE* due to the 256 fd limit - use this cheap hack
   instead */
typedef struct {
	char *buf;
	char *p;
	size_t size;
	char *end_section_p;
} myFILE;

static int mygetc(myFILE *f)
{
	if (f->p >= f->buf+f->size)
		return EOF;
        /* be sure to return chars >127 as positive values */
	return (int)( *(f->p++) & 0x00FF );
}

static void myfile_close(myFILE *f)
{
	if (!f)
		return;
	SAFE_FREE(f->buf);
	SAFE_FREE(f);
}

/* Find the end of the section. We must use mb functions for this. */
static int FindSectionEnd(myFILE *f)
{
	f->end_section_p = strchr_m(f->p, ']');
	return f->end_section_p ? 1 : 0;
}

static int AtSectionEnd(myFILE *f)
{
	if (f->p == f->end_section_p + 1) {
		f->end_section_p = NULL;
		return 1;
	}
	return 0;
}

/* -------------------------------------------------------------------------- **
 * Functions...
 */
  /* ------------------------------------------------------------------------ **
   * Scan past whitespace (see ctype(3C)) and return the first non-whitespace
   * character, or newline, or EOF.
   *
   *  Input:  InFile  - Input source.
   *
   *  Output: The next non-whitespace character in the input stream.
   *
   *  Notes:  Because the config files use a line-oriented grammar, we
   *          explicitly exclude the newline character from the list of
   *          whitespace characters.
   *        - Note that both EOF (-1) and the nul character ('\0') are
   *          considered end-of-file markers.
   *
   * ------------------------------------------------------------------------ **
   */
 
static int EatWhitespace( myFILE *InFile )
{
	int c;

	for( c = mygetc( InFile ); isspace( c ) && ('\n' != c); c = mygetc( InFile ) )
		;
	return( c );
}

  /* ------------------------------------------------------------------------ **
   * Scan to the end of a comment.
   *
   *  Input:  InFile  - Input source.
   *
   *  Output: The character that marks the end of the comment.  Normally,
   *          this will be a newline, but it *might* be an EOF.
   *
   *  Notes:  Because the config files use a line-oriented grammar, we
   *          explicitly exclude the newline character from the list of
   *          whitespace characters.
   *        - Note that both EOF (-1) and the nul character ('\0') are
   *          considered end-of-file markers.
   *
   * ------------------------------------------------------------------------ **
   */

static int EatComment( myFILE *InFile )
{
	int c;

	for( c = mygetc( InFile ); ('\n'!=c) && (EOF!=c) && (c>0); c = mygetc( InFile ) )
		;
	return( c );
}

/*****************************************************************************
 * Scan backards within a string to discover if the last non-whitespace
 * character is a line-continuation character ('\\').
 *
 *  Input:  line  - A pointer to a buffer containing the string to be
 *                  scanned.
 *          pos   - This is taken to be the offset of the end of the
 *                  string.  This position is *not* scanned.
 *
 *  Output: The offset of the '\\' character if it was found, or -1 to
 *          indicate that it was not.
 *
 *****************************************************************************/

static int Continuation(char *line, int pos )
{
	pos--;
	while( (pos >= 0) && isspace((int)line[pos]))
		pos--;

	return (((pos >= 0) && ('\\' == line[pos])) ? pos : -1 );
}

/* ------------------------------------------------------------------------ **
 * Scan a section name, and pass the name to function sfunc().
 *
 *  Input:  InFile  - Input source.
 *          sfunc   - Pointer to the function to be called if the section
 *                    name is successfully read.
 *
 *  Output: True if the section name was read and True was returned from
 *          <sfunc>.  False if <sfunc> failed or if a lexical error was
 *          encountered.
 *
 * ------------------------------------------------------------------------ **
 */

static BOOL Section( myFILE *InFile, BOOL (*sfunc)(const char *) )
{
	int   c;
	int   i;
	int   end;
	const char *func  = "params.c:Section() -";

	i = 0;      /* <i> is the offset of the next free byte in bufr[] and  */
	end = 0;    /* <end> is the current "end of string" offset.  In most  */
		    /* cases these will be the same, but if the last          */
		    /* character written to bufr[] is a space, then <end>     */
		    /* will be one less than <i>.                             */


	/* Find the end of the section. We must use mb functions for this. */
	if (!FindSectionEnd(InFile)) {
		DEBUG(0, ("%s No terminating ']' character in section.\n", func) );
		return False;
	}

	c = EatWhitespace( InFile );    /* We've already got the '['.  Scan */
					/* past initial white space.        */

	while( (EOF != c) && (c > 0) ) {
		/* Check that the buffer is big enough for the next character. */
		if( i > (bSize - 2) ) {
			char *tb;
      
			tb = (char *)SMB_REALLOC( bufr, bSize +BUFR_INC );
			if( NULL == tb ) {
				DEBUG(0, ("%s Memory re-allocation failure.", func) );
				return False;
			}
			bufr = tb;
			bSize += BUFR_INC;
		}

		/* Handle a single character other than section end. */
		switch( c ) {
			case '\n': /* Got newline before closing ']'.    */
				i = Continuation( bufr, i );    /* Check for line continuation.     */
				if( i < 0 ) {
					bufr[end] = '\0';
					DEBUG(0, ("%s Badly formed line in configuration file: %s\n", func, bufr ));
					return False;
				}
				end = ( (i > 0) && (' ' == bufr[i - 1]) ) ? (i - 1) : (i);
					c = mygetc( InFile );             /* Continue with next line.         */
				break;

			default: /* All else are a valid name chars.   */
				if(isspace( c )) {
					/* One space per whitespace region. */
					bufr[end] = ' ';
					i = end + 1;
					c = EatWhitespace( InFile );
				} else {
					bufr[i++] = c;
					end = i;
					c = mygetc( InFile );
				}
		}

		if (AtSectionEnd(InFile)) {
			/* Got to the closing bracket. */