1 /* -------------------------------------------------------------------------- **
2 * Microsoft Network Services for Unix, AKA., Andrew Tridgell's SAMBA.
4 * This module Copyright (C) 1990-1998 Karl Auer
6 * Rewritten almost completely by Christopher R. Hertel
7 * at the University of Minnesota, September, 1997.
8 * This module Copyright (C) 1997-1998 by the University of Minnesota
9 * -------------------------------------------------------------------------- **
11 * This program is free software; you can redistribute it and/or modify
12 * it under the terms of the GNU General Public License as published by
13 * the Free Software Foundation; either version 2 of the License, or
14 * (at your option) any later version.
16 * This program is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 * GNU General Public License for more details.
21 * You should have received a copy of the GNU General Public License
22 * along with this program; if not, write to the Free Software
23 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
25 * -------------------------------------------------------------------------- **
29 * -------------------------------------------------------------------------- **
31 * This module performs lexical analysis and initial parsing of a
32 * Windows-like parameter file. It recognizes and handles four token
33 * types: section-name, parameter-name, parameter-value, and
34 * end-of-file. Comments and line continuation are handled
37 * The entry point to the module is function pm_process(). This
38 * function opens the source file, calls the Parse() function to parse
39 * the input, and then closes the file when either the EOF is reached
40 * or a fatal error is encountered.
42 * A sample parameter file might look like this:
45 * parameter one = value string
46 * parameter two = another value
48 * new parameter = some value or t'other
50 * The parameter file is divided into sections by section headers:
51 * section names enclosed in square brackets (eg. [section one]).
52 * Each section contains parameter lines, each of which consist of a
53 * parameter name and value delimited by an equal sign. Roughly, the
56 * <file> :== { <section> } EOF
58 * <section> :== <section header> { <parameter line> }
60 * <section header> :== '[' NAME ']'
62 * <parameter line> :== NAME '=' VALUE '\n'
64 * Blank lines and comment lines are ignored. Comment lines are lines
65 * beginning with either a semicolon (';') or a pound sign ('#').
67 * All whitespace in section names and parameter names is compressed
68 * to single spaces. Leading and trailing whitespace is stipped from
69 * both names and values.
71 * Only the first equals sign in a parameter line is significant.
72 * Parameter values may contain equals signs, square brackets and
73 * semicolons. Internal whitespace is retained in parameter values,
74 * with the exception of the '\r' character, which is stripped for
75 * historic reasons. Parameter names may not start with a left square
76 * bracket, an equal sign, a pound sign, or a semicolon, because these
77 * are used to identify other tokens.
79 * -------------------------------------------------------------------------- **
84 /* -------------------------------------------------------------------------- **
91 /* -------------------------------------------------------------------------- **
94 * DEBUGLEVEL - The ubiquitous DEBUGLEVEL. This determines which DEBUG()
95 * messages will be produced.
96 * bufr - pointer to a global buffer. This is probably a kludge,
97 * but it was the nicest kludge I could think of (for now).
98 * bSize - The size of the global buffer <bufr>.
101 extern int DEBUGLEVEL
;
103 static char *bufr
= NULL
;
104 static int bSize
= 0;
106 /* we can't use FILE* due to the 256 fd limit - use this cheap hack
114 static int mygetc(myFILE
*f
)
116 if (f
->p
>= f
->buf
+f
->size
) return EOF
;
117 /* be sure to return chars >127 as positive values */
118 return (int)( *(f
->p
++) & 0x00FF );
121 static void myfile_close(myFILE
*f
)
124 if (f
->buf
) free(f
->buf
);
128 /* -------------------------------------------------------------------------- **
132 static int EatWhitespace( myFILE
*InFile
)
133 /* ------------------------------------------------------------------------ **
134 * Scan past whitespace (see ctype(3C)) and return the first non-whitespace
135 * character, or newline, or EOF.
137 * Input: InFile - Input source.
139 * Output: The next non-whitespace character in the input stream.
141 * Notes: Because the config files use a line-oriented grammar, we
142 * explicitly exclude the newline character from the list of
143 * whitespace characters.
144 * - Note that both EOF (-1) and the nul character ('\0') are
145 * considered end-of-file markers.
147 * ------------------------------------------------------------------------ **
152 for( c
= mygetc( InFile
); isspace( c
) && ('\n' != c
); c
= mygetc( InFile
) )
155 } /* EatWhitespace */
157 static int EatComment( myFILE
*InFile
)
158 /* ------------------------------------------------------------------------ **
159 * Scan to the end of a comment.
161 * Input: InFile - Input source.
163 * Output: The character that marks the end of the comment. Normally,
164 * this will be a newline, but it *might* be an EOF.
166 * Notes: Because the config files use a line-oriented grammar, we
167 * explicitly exclude the newline character from the list of
168 * whitespace characters.
169 * - Note that both EOF (-1) and the nul character ('\0') are
170 * considered end-of-file markers.
172 * ------------------------------------------------------------------------ **
177 for( c
= mygetc( InFile
); ('\n'!=c
) && (EOF
!=c
) && (c
>0); c
= mygetc( InFile
) )
182 /*****************************************************************************
183 * Scan backards within a string to discover if the last non-whitespace
184 * character is a line-continuation character ('\\').
186 * Input: line - A pointer to a buffer containing the string to be
188 * pos - This is taken to be the offset of the end of the
189 * string. This position is *not* scanned.
191 * Output: The offset of the '\\' character if it was found, or -1 to
192 * indicate that it was not.
194 *****************************************************************************/
196 static int Continuation( char *line
, int pos
)
201 while( (pos
>= 0) && isspace(line
[pos
]) )
204 /* we should recognize if `\` is part of a multibyte character or not. */
207 skip
= get_character_len(line
[pos2
]);
210 } else if (pos
== pos2
) {
211 return( ((pos
>= 0) && ('\\' == line
[pos
])) ? pos
: -1 );
220 static BOOL
Section( myFILE
*InFile
, BOOL (*sfunc
)(char *) )
221 /* ------------------------------------------------------------------------ **
222 * Scan a section name, and pass the name to function sfunc().
224 * Input: InFile - Input source.
225 * sfunc - Pointer to the function to be called if the section
226 * name is successfully read.
228 * Output: True if the section name was read and True was returned from
229 * <sfunc>. False if <sfunc> failed or if a lexical error was
232 * ------------------------------------------------------------------------ **
238 char *func
= "params.c:Section() -";
240 i
= 0; /* <i> is the offset of the next free byte in bufr[] and */
241 end
= 0; /* <end> is the current "end of string" offset. In most */
242 /* cases these will be the same, but if the last */
243 /* character written to bufr[] is a space, then <end> */
244 /* will be one less than <i>. */
246 c
= EatWhitespace( InFile
); /* We've already got the '['. Scan */
247 /* past initial white space. */
249 while( (EOF
!= c
) && (c
> 0) )
252 /* Check that the buffer is big enough for the next character. */
253 if( i
> (bSize
- 2) )
256 bufr
= Realloc( bufr
, bSize
);
259 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
264 /* Handle a single character. */
267 case ']': /* Found the closing bracket. */
269 if( 0 == end
) /* Don't allow an empty name. */
271 DEBUG(0, ("%s Empty section name in configuration file.\n", func
));
274 if( !sfunc( bufr
) ) /* Got a valid name. Deal with it. */
276 (void)EatComment( InFile
); /* Finish off the line. */
279 case '\n': /* Got newline before closing ']'. */
280 i
= Continuation( bufr
, i
); /* Check for line continuation. */
284 DEBUG(0, ("%s Badly formed line in configuration file: %s\n",
288 end
= ( (i
> 0) && (' ' == bufr
[i
- 1]) ) ? (i
- 1) : (i
);
289 c
= mygetc( InFile
); /* Continue with next line. */
292 default: /* All else are a valid name chars. */
293 if( isspace( c
) ) /* One space per whitespace region. */
297 c
= EatWhitespace( InFile
);
299 else /* All others copy verbatim. */
303 c
= mygetc( InFile
);
308 /* We arrive here if we've met the EOF before the closing bracket. */
309 DEBUG(0, ("%s Unexpected EOF in the configuration file: %s\n", func
, bufr
));
313 static BOOL
Parameter( myFILE
*InFile
, BOOL (*pfunc
)(char *, char *), int c
)
314 /* ------------------------------------------------------------------------ **
315 * Scan a parameter name and value, and pass these two fields to pfunc().
317 * Input: InFile - The input source.
318 * pfunc - A pointer to the function that will be called to
319 * process the parameter, once it has been scanned.
320 * c - The first character of the parameter name, which
321 * would have been read by Parse(). Unlike a comment
322 * line or a section header, there is no lead-in
323 * character that can be discarded.
325 * Output: True if the parameter name and value were scanned and processed
326 * successfully, else False.
328 * Notes: This function is in two parts. The first loop scans the
329 * parameter name. Internal whitespace is compressed, and an
330 * equal sign (=) terminates the token. Leading and trailing
331 * whitespace is discarded. The second loop scans the parameter
332 * value. When both have been successfully identified, they are
333 * passed to pfunc() for processing.
335 * ------------------------------------------------------------------------ **
338 int i
= 0; /* Position within bufr. */
339 int end
= 0; /* bufr[end] is current end-of-string. */
340 int vstart
= 0; /* Starting position of the parameter value. */
341 char *func
= "params.c:Parameter() -";
343 /* Read the parameter name. */
344 while( 0 == vstart
) /* Loop until we've found the start of the value. */
347 if( i
> (bSize
- 2) ) /* Ensure there's space for next char. */
350 bufr
= Realloc( bufr
, bSize
);
353 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
360 case '=': /* Equal sign marks end of param name. */
361 if( 0 == end
) /* Don't allow an empty name. */
363 DEBUG(0, ("%s Invalid parameter name in config. file.\n", func
));
366 bufr
[end
++] = '\0'; /* Mark end of string & advance. */
367 i
= end
; /* New string starts here. */
368 vstart
= end
; /* New string is parameter value. */
369 bufr
[i
] = '\0'; /* New string is nul, for now. */
372 case '\n': /* Find continuation char, else error. */
373 i
= Continuation( bufr
, i
);
377 DEBUG(1,("%s Ignoring badly formed line in configuration file: %s\n",
381 end
= ( (i
> 0) && (' ' == bufr
[i
- 1]) ) ? (i
- 1) : (i
);
382 c
= mygetc( InFile
); /* Read past eoln. */
385 case '\0': /* Shouldn't have EOF within param name. */
388 DEBUG(1,("%s Unexpected end-of-file at: %s\n", func
, bufr
));
392 if( isspace( c
) ) /* One ' ' per whitespace region. */
396 c
= EatWhitespace( InFile
);
398 else /* All others verbatim. */
402 c
= mygetc( InFile
);
407 /* Now parse the value. */
408 c
= EatWhitespace( InFile
); /* Again, trim leading whitespace. */
409 while( (EOF
!=c
) && (c
> 0) )
412 if( i
> (bSize
- 2) ) /* Make sure there's enough room. */
415 bufr
= Realloc( bufr
, bSize
);
418 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
425 case '\r': /* Explicitly remove '\r' because the older */
426 c
= mygetc( InFile
); /* version called fgets_slash() which also */
427 break; /* removes them. */
429 case '\n': /* Marks end of value unless there's a '\'. */
430 i
= Continuation( bufr
, i
);
435 for( end
= i
; (end
>= 0) && isspace(bufr
[end
]); end
-- )
437 c
= mygetc( InFile
);
441 default: /* All others verbatim. Note that spaces do */
442 bufr
[i
++] = c
; /* not advance <end>. This allows trimming */
443 if( !isspace( c
) ) /* of whitespace at the end of the line. */
445 c
= mygetc( InFile
);
449 bufr
[end
] = '\0'; /* End of value. */
451 return( pfunc( bufr
, &bufr
[vstart
] ) ); /* Pass name & value to pfunc(). */
454 static BOOL
Parse( myFILE
*InFile
,
455 BOOL (*sfunc
)(char *),
456 BOOL (*pfunc
)(char *, char *) )
457 /* ------------------------------------------------------------------------ **
458 * Scan & parse the input.
460 * Input: InFile - Input source.
461 * sfunc - Function to be called when a section name is scanned.
463 * pfunc - Function to be called when a parameter is scanned.
466 * Output: True if the file was successfully scanned, else False.
468 * Notes: The input can be viewed in terms of 'lines'. There are four
470 * Blank - May contain whitespace, otherwise empty.
471 * Comment - First non-whitespace character is a ';' or '#'.
472 * The remainder of the line is ignored.
473 * Section - First non-whitespace character is a '['.
474 * Parameter - The default case.
476 * ------------------------------------------------------------------------ **
481 c
= EatWhitespace( InFile
);
482 while( (EOF
!= c
) && (c
> 0) )
486 case '\n': /* Blank line. */
487 c
= EatWhitespace( InFile
);
490 case ';': /* Comment line. */
492 c
= EatComment( InFile
);
495 case '[': /* Section Header. */
496 if( !Section( InFile
, sfunc
) )
498 c
= EatWhitespace( InFile
);
501 case '\\': /* Bogus backslash. */
502 c
= EatWhitespace( InFile
);
505 default: /* Parameter line. */
506 if( !Parameter( InFile
, pfunc
, c
) )
508 c
= EatWhitespace( InFile
);
515 static myFILE
*OpenConfFile( char *FileName
)
516 /* ------------------------------------------------------------------------ **
517 * Open a configuration file.
519 * Input: FileName - The pathname of the config file to be opened.
521 * Output: A pointer of type (char **) to the lines of the file
523 * ------------------------------------------------------------------------ **
526 char *func
= "params.c:OpenConfFile() -";
527 extern BOOL in_client
;
528 int lvl
= in_client
?1:0;
531 ret
= (myFILE
*)malloc(sizeof(*ret
));
532 if (!ret
) return NULL
;
534 ret
->buf
= file_load(FileName
, &ret
->size
);
535 if( NULL
== ret
->buf
)
538 ("%s Unable to open configuration file \"%s\":\n\t%s\n",
539 func
, FileName
, strerror(errno
)) );
548 BOOL
pm_process( char *FileName
,
549 BOOL (*sfunc
)(char *),
550 BOOL (*pfunc
)(char *, char *) )
551 /* ------------------------------------------------------------------------ **
552 * Process the named parameter file.
554 * Input: FileName - The pathname of the parameter file to be opened.
555 * sfunc - A pointer to a function that will be called when
556 * a section name is discovered.
557 * pfunc - A pointer to a function that will be called when
558 * a parameter name and value are discovered.
560 * Output: TRUE if the file was successfully parsed, else FALSE.
562 * ------------------------------------------------------------------------ **
567 char *func
= "params.c:pm_process() -";
569 InFile
= OpenConfFile( FileName
); /* Open the config file. */
573 DEBUG( 3, ("%s Processing configuration file \"%s\"\n", func
, FileName
) );
575 if( NULL
!= bufr
) /* If we already have a buffer */
576 result
= Parse( InFile
, sfunc
, pfunc
); /* (recursive call), then just */
579 else /* If we don't have a buffer */
580 { /* allocate one, then parse, */
581 bSize
= BUFR_INC
; /* then free. */
582 bufr
= (char *)malloc( bSize
);
585 DEBUG(0,("%s memory allocation failure.\n", func
));
586 myfile_close(InFile
);
589 result
= Parse( InFile
, sfunc
, pfunc
);
595 myfile_close(InFile
);
597 if( !result
) /* Generic failure. */
599 DEBUG(0,("%s Failed. Error returned from params.c:parse().\n", func
));
603 return( True
); /* Generic success. */
606 /* -------------------------------------------------------------------------- */