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 static char *bufr
= NULL
;
102 static int bSize
= 0;
104 /* we can't use FILE* due to the 256 fd limit - use this cheap hack
112 static int mygetc(myFILE
*f
)
114 if (f
->p
>= f
->buf
+f
->size
) return EOF
;
115 /* be sure to return chars >127 as positive values */
116 return (int)( *(f
->p
++) & 0x00FF );
119 static void myfile_close(myFILE
*f
)
126 /* -------------------------------------------------------------------------- **
130 static int EatWhitespace( myFILE
*InFile
)
131 /* ------------------------------------------------------------------------ **
132 * Scan past whitespace (see ctype(3C)) and return the first non-whitespace
133 * character, or newline, or EOF.
135 * Input: InFile - Input source.
137 * Output: The next non-whitespace character in the input stream.
139 * Notes: Because the config files use a line-oriented grammar, we
140 * explicitly exclude the newline character from the list of
141 * whitespace characters.
142 * - Note that both EOF (-1) and the nul character ('\0') are
143 * considered end-of-file markers.
145 * ------------------------------------------------------------------------ **
150 for( c
= mygetc( InFile
); isspace( c
) && ('\n' != c
); c
= mygetc( InFile
) )
153 } /* EatWhitespace */
155 static int EatComment( myFILE
*InFile
)
156 /* ------------------------------------------------------------------------ **
157 * Scan to the end of a comment.
159 * Input: InFile - Input source.
161 * Output: The character that marks the end of the comment. Normally,
162 * this will be a newline, but it *might* be an EOF.
164 * Notes: Because the config files use a line-oriented grammar, we
165 * explicitly exclude the newline character from the list of
166 * whitespace characters.
167 * - Note that both EOF (-1) and the nul character ('\0') are
168 * considered end-of-file markers.
170 * ------------------------------------------------------------------------ **
175 for( c
= mygetc( InFile
); ('\n'!=c
) && (EOF
!=c
) && (c
>0); c
= mygetc( InFile
) )
180 /*****************************************************************************
181 * Scan backards within a string to discover if the last non-whitespace
182 * character is a line-continuation character ('\\').
184 * Input: line - A pointer to a buffer containing the string to be
186 * pos - This is taken to be the offset of the end of the
187 * string. This position is *not* scanned.
189 * Output: The offset of the '\\' character if it was found, or -1 to
190 * indicate that it was not.
192 *****************************************************************************/
194 static int Continuation( char *line
, int pos
)
199 while( (pos
>= 0) && isspace((int)line
[pos
]) )
202 /* we should recognize if `\` is part of a multibyte character or not. */
205 skip
= get_character_len(line
[pos2
]);
208 } else if (pos
== pos2
) {
209 return( ((pos
>= 0) && ('\\' == line
[pos
])) ? pos
: -1 );
218 static BOOL
Section( myFILE
*InFile
, BOOL (*sfunc
)(char *) )
219 /* ------------------------------------------------------------------------ **
220 * Scan a section name, and pass the name to function sfunc().
222 * Input: InFile - Input source.
223 * sfunc - Pointer to the function to be called if the section
224 * name is successfully read.
226 * Output: True if the section name was read and True was returned from
227 * <sfunc>. False if <sfunc> failed or if a lexical error was
230 * ------------------------------------------------------------------------ **
236 char *func
= "params.c:Section() -";
238 i
= 0; /* <i> is the offset of the next free byte in bufr[] and */
239 end
= 0; /* <end> is the current "end of string" offset. In most */
240 /* cases these will be the same, but if the last */
241 /* character written to bufr[] is a space, then <end> */
242 /* will be one less than <i>. */
244 c
= EatWhitespace( InFile
); /* We've already got the '['. Scan */
245 /* past initial white space. */
247 while( (EOF
!= c
) && (c
> 0) )
250 /* Check that the buffer is big enough for the next character. */
251 if( i
> (bSize
- 2) )
255 tb
= Realloc( bufr
, bSize
+BUFR_INC
);
258 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
265 /* Handle a single character. */
268 case ']': /* Found the closing bracket. */
270 if( 0 == end
) /* Don't allow an empty name. */
272 DEBUG(0, ("%s Empty section name in configuration file.\n", func
));
275 if( !sfunc( unix_to_dos(bufr
) ) ) /* Got a valid name. Deal with it. */
277 (void)EatComment( InFile
); /* Finish off the line. */
280 case '\n': /* Got newline before closing ']'. */
281 i
= Continuation( bufr
, i
); /* Check for line continuation. */
285 DEBUG(0, ("%s Badly formed line in configuration file: %s\n",
289 end
= ( (i
> 0) && (' ' == bufr
[i
- 1]) ) ? (i
- 1) : (i
);
290 c
= mygetc( InFile
); /* Continue with next line. */
293 default: /* All else are a valid name chars. */
294 if( isspace( c
) ) /* One space per whitespace region. */
298 c
= EatWhitespace( InFile
);
300 else /* All others copy verbatim. */
304 c
= mygetc( InFile
);
309 /* We arrive here if we've met the EOF before the closing bracket. */
310 DEBUG(0, ("%s Unexpected EOF in the configuration file: %s\n", func
, bufr
));
314 static BOOL
Parameter( myFILE
*InFile
, BOOL (*pfunc
)(char *, char *), int c
)
315 /* ------------------------------------------------------------------------ **
316 * Scan a parameter name and value, and pass these two fields to pfunc().
318 * Input: InFile - The input source.
319 * pfunc - A pointer to the function that will be called to
320 * process the parameter, once it has been scanned.
321 * c - The first character of the parameter name, which
322 * would have been read by Parse(). Unlike a comment
323 * line or a section header, there is no lead-in
324 * character that can be discarded.
326 * Output: True if the parameter name and value were scanned and processed
327 * successfully, else False.
329 * Notes: This function is in two parts. The first loop scans the
330 * parameter name. Internal whitespace is compressed, and an
331 * equal sign (=) terminates the token. Leading and trailing
332 * whitespace is discarded. The second loop scans the parameter
333 * value. When both have been successfully identified, they are
334 * passed to pfunc() for processing.
336 * ------------------------------------------------------------------------ **
339 int i
= 0; /* Position within bufr. */
340 int end
= 0; /* bufr[end] is current end-of-string. */
341 int vstart
= 0; /* Starting position of the parameter value. */
342 char *func
= "params.c:Parameter() -";
344 /* Read the parameter name. */
345 while( 0 == vstart
) /* Loop until we've found the start of the value. */
348 if( i
> (bSize
- 2) ) /* Ensure there's space for next char. */
352 tb
= Realloc( bufr
, bSize
+ BUFR_INC
);
355 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
364 case '=': /* Equal sign marks end of param name. */
365 if( 0 == end
) /* Don't allow an empty name. */
367 DEBUG(0, ("%s Invalid parameter name in config. file.\n", func
));
370 bufr
[end
++] = '\0'; /* Mark end of string & advance. */
371 i
= end
; /* New string starts here. */
372 vstart
= end
; /* New string is parameter value. */
373 bufr
[i
] = '\0'; /* New string is nul, for now. */
376 case '\n': /* Find continuation char, else error. */
377 i
= Continuation( bufr
, i
);
381 DEBUG(1,("%s Ignoring badly formed line in configuration file: %s\n",
385 end
= ( (i
> 0) && (' ' == bufr
[i
- 1]) ) ? (i
- 1) : (i
);
386 c
= mygetc( InFile
); /* Read past eoln. */
389 case '\0': /* Shouldn't have EOF within param name. */
392 DEBUG(1,("%s Unexpected end-of-file at: %s\n", func
, bufr
));
396 if( isspace( c
) ) /* One ' ' per whitespace region. */
400 c
= EatWhitespace( InFile
);
402 else /* All others verbatim. */
406 c
= mygetc( InFile
);
411 /* Now parse the value. */
412 c
= EatWhitespace( InFile
); /* Again, trim leading whitespace. */
413 while( (EOF
!=c
) && (c
> 0) )
416 if( i
> (bSize
- 2) ) /* Make sure there's enough room. */
419 bufr
= Realloc( bufr
, bSize
);
422 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
429 case '\r': /* Explicitly remove '\r' because the older */
430 c
= mygetc( InFile
); /* version called fgets_slash() which also */
431 break; /* removes them. */
433 case '\n': /* Marks end of value unless there's a '\'. */
434 i
= Continuation( bufr
, i
);
439 for( end
= i
; (end
>= 0) && isspace((int)bufr
[end
]); end
-- )
441 c
= mygetc( InFile
);
445 default: /* All others verbatim. Note that spaces do */
446 bufr
[i
++] = c
; /* not advance <end>. This allows trimming */
447 if( !isspace( c
) ) /* of whitespace at the end of the line. */
449 c
= mygetc( InFile
);
453 bufr
[end
] = '\0'; /* End of value. */
455 return( pfunc( bufr
, &bufr
[vstart
] ) ); /* Pass name & value to pfunc(). */
458 static BOOL
Parse( myFILE
*InFile
,
459 BOOL (*sfunc
)(char *),
460 BOOL (*pfunc
)(char *, char *) )
461 /* ------------------------------------------------------------------------ **
462 * Scan & parse the input.
464 * Input: InFile - Input source.
465 * sfunc - Function to be called when a section name is scanned.
467 * pfunc - Function to be called when a parameter is scanned.
470 * Output: True if the file was successfully scanned, else False.
472 * Notes: The input can be viewed in terms of 'lines'. There are four
474 * Blank - May contain whitespace, otherwise empty.
475 * Comment - First non-whitespace character is a ';' or '#'.
476 * The remainder of the line is ignored.
477 * Section - First non-whitespace character is a '['.
478 * Parameter - The default case.
480 * ------------------------------------------------------------------------ **
485 c
= EatWhitespace( InFile
);
486 while( (EOF
!= c
) && (c
> 0) )
490 case '\n': /* Blank line. */
491 c
= EatWhitespace( InFile
);
494 case ';': /* Comment line. */
496 c
= EatComment( InFile
);
499 case '[': /* Section Header. */
500 if( !Section( InFile
, sfunc
) )
502 c
= EatWhitespace( InFile
);
505 case '\\': /* Bogus backslash. */
506 c
= EatWhitespace( InFile
);
509 default: /* Parameter line. */
510 if( !Parameter( InFile
, pfunc
, c
) )
512 c
= EatWhitespace( InFile
);
519 static myFILE
*OpenConfFile( char *FileName
)
520 /* ------------------------------------------------------------------------ **
521 * Open a configuration file.
523 * Input: FileName - The pathname of the config file to be opened.
525 * Output: A pointer of type (char **) to the lines of the file
527 * ------------------------------------------------------------------------ **
530 char *func
= "params.c:OpenConfFile() -";
531 extern BOOL in_client
;
532 int lvl
= in_client
?1:0;
535 ret
= (myFILE
*)malloc(sizeof(*ret
));
536 if (!ret
) return NULL
;
538 ret
->buf
= file_load(FileName
, &ret
->size
);
539 if( NULL
== ret
->buf
)
542 ("%s Unable to open configuration file \"%s\":\n\t%s\n",
543 func
, FileName
, strerror(errno
)) );
552 BOOL
pm_process( char *FileName
,
553 BOOL (*sfunc
)(char *),
554 BOOL (*pfunc
)(char *, char *) )
555 /* ------------------------------------------------------------------------ **
556 * Process the named parameter file.
558 * Input: FileName - The pathname of the parameter file to be opened.
559 * sfunc - A pointer to a function that will be called when
560 * a section name is discovered.
561 * pfunc - A pointer to a function that will be called when
562 * a parameter name and value are discovered.
564 * Output: TRUE if the file was successfully parsed, else FALSE.
566 * ------------------------------------------------------------------------ **
571 char *func
= "params.c:pm_process() -";
573 InFile
= OpenConfFile( FileName
); /* Open the config file. */
577 DEBUG( 3, ("%s Processing configuration file \"%s\"\n", func
, FileName
) );
579 if( NULL
!= bufr
) /* If we already have a buffer */
580 result
= Parse( InFile
, sfunc
, pfunc
); /* (recursive call), then just */
583 else /* If we don't have a buffer */
584 { /* allocate one, then parse, */
585 bSize
= BUFR_INC
; /* then free. */
586 bufr
= (char *)malloc( bSize
);
589 DEBUG(0,("%s memory allocation failure.\n", func
));
590 myfile_close(InFile
);
593 result
= Parse( InFile
, sfunc
, pfunc
);
599 myfile_close(InFile
);
601 if( !result
) /* Generic failure. */
603 DEBUG(0,("%s Failed. Error returned from params.c:parse().\n", func
));
607 return( True
); /* Generic success. */
610 /* -------------------------------------------------------------------------- */