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 * -------------------------------------------------------------------------- **
83 #include "system/iconv.h"
85 /* -------------------------------------------------------------------------- **
92 /* we can't use FILE* due to the 256 fd limit - use this cheap hack
102 static int mygetc(myFILE
*f
)
104 if (f
->p
>= f
->buf
+f
->size
) return EOF
;
105 /* be sure to return chars >127 as positive values */
106 return (int)( *(f
->p
++) & 0x00FF );
109 static void myfile_close(myFILE
*f
)
114 /* -------------------------------------------------------------------------- **
118 static int EatWhitespace( myFILE
*InFile
)
119 /* ------------------------------------------------------------------------ **
120 * Scan past whitespace (see ctype(3C)) and return the first non-whitespace
121 * character, or newline, or EOF.
123 * Input: InFile - Input source.
125 * Output: The next non-whitespace character in the input stream.
127 * Notes: Because the config files use a line-oriented grammar, we
128 * explicitly exclude the newline character from the list of
129 * whitespace characters.
130 * - Note that both EOF (-1) and the nul character ('\0') are
131 * considered end-of-file markers.
133 * ------------------------------------------------------------------------ **
138 for( c
= mygetc( InFile
); isspace( c
) && ('\n' != c
); c
= mygetc( InFile
) )
141 } /* EatWhitespace */
143 static int EatComment( myFILE
*InFile
)
144 /* ------------------------------------------------------------------------ **
145 * Scan to the end of a comment.
147 * Input: InFile - Input source.
149 * Output: The character that marks the end of the comment. Normally,
150 * this will be a newline, but it *might* be an EOF.
152 * Notes: Because the config files use a line-oriented grammar, we
153 * explicitly exclude the newline character from the list of
154 * whitespace characters.
155 * - Note that both EOF (-1) and the nul character ('\0') are
156 * considered end-of-file markers.
158 * ------------------------------------------------------------------------ **
163 for( c
= mygetc( InFile
); ('\n'!=c
) && (EOF
!=c
) && (c
>0); c
= mygetc( InFile
) )
168 /*****************************************************************************
169 * Scan backards within a string to discover if the last non-whitespace
170 * character is a line-continuation character ('\\').
172 * Input: line - A pointer to a buffer containing the string to be
174 * pos - This is taken to be the offset of the end of the
175 * string. This position is *not* scanned.
177 * Output: The offset of the '\\' character if it was found, or -1 to
178 * indicate that it was not.
180 *****************************************************************************/
182 static int Continuation(char *line
, int pos
)
185 while( (pos
>= 0) && isspace((int)line
[pos
]))
188 return (((pos
>= 0) && ('\\' == line
[pos
])) ? pos
: -1 );
192 static BOOL
Section( myFILE
*InFile
, BOOL (*sfunc
)(const char *, void *), void *userdata
)
193 /* ------------------------------------------------------------------------ **
194 * Scan a section name, and pass the name to function sfunc().
196 * Input: InFile - Input source.
197 * sfunc - Pointer to the function to be called if the section
198 * name is successfully read.
200 * Output: True if the section name was read and True was returned from
201 * <sfunc>. False if <sfunc> failed or if a lexical error was
204 * ------------------------------------------------------------------------ **
210 const char *func
= "params.c:Section() -";
212 i
= 0; /* <i> is the offset of the next free byte in bufr[] and */
213 end
= 0; /* <end> is the current "end of string" offset. In most */
214 /* cases these will be the same, but if the last */
215 /* character written to bufr[] is a space, then <end> */
216 /* will be one less than <i>. */
218 c
= EatWhitespace( InFile
); /* We've already got the '['. Scan */
219 /* past initial white space. */
221 while( (EOF
!= c
) && (c
> 0) )
224 /* Check that the buffer is big enough for the next character. */
225 if( i
> (InFile
->bSize
- 2) )
229 tb
= talloc_realloc(InFile
, InFile
->bufr
, char, InFile
->bSize
+ BUFR_INC
);
232 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
236 InFile
->bSize
+= BUFR_INC
;
239 /* Handle a single character. */
242 case ']': /* Found the closing bracket. */
243 InFile
->bufr
[end
] = '\0';
244 if( 0 == end
) /* Don't allow an empty name. */
246 DEBUG(0, ("%s Empty section name in configuration file.\n", func
));
249 if( !sfunc(InFile
->bufr
,userdata
) ) /* Got a valid name. Deal with it. */
251 (void)EatComment( InFile
); /* Finish off the line. */
254 case '\n': /* Got newline before closing ']'. */
255 i
= Continuation( InFile
->bufr
, i
); /* Check for line continuation. */
258 InFile
->bufr
[end
] = '\0';
259 DEBUG(0, ("%s Badly formed line in configuration file: %s\n",
260 func
, InFile
->bufr
));
263 end
= ( (i
> 0) && (' ' == InFile
->bufr
[i
- 1]) ) ? (i
- 1) : (i
);
264 c
= mygetc( InFile
); /* Continue with next line. */
267 default: /* All else are a valid name chars. */
268 if( isspace( c
) ) /* One space per whitespace region. */
270 InFile
->bufr
[end
] = ' ';
272 c
= EatWhitespace( InFile
);
274 else /* All others copy verbatim. */
276 InFile
->bufr
[i
++] = c
;
278 c
= mygetc( InFile
);
283 /* We arrive here if we've met the EOF before the closing bracket. */
284 DEBUG(0, ("%s Unexpected EOF in the configuration file\n", func
));
288 static BOOL
Parameter( myFILE
*InFile
, BOOL (*pfunc
)(const char *, const char *, void *), int c
, void *userdata
)
289 /* ------------------------------------------------------------------------ **
290 * Scan a parameter name and value, and pass these two fields to pfunc().
292 * Input: InFile - The input source.
293 * pfunc - A pointer to the function that will be called to
294 * process the parameter, once it has been scanned.
295 * c - The first character of the parameter name, which
296 * would have been read by Parse(). Unlike a comment
297 * line or a section header, there is no lead-in
298 * character that can be discarded.
300 * Output: True if the parameter name and value were scanned and processed
301 * successfully, else False.
303 * Notes: This function is in two parts. The first loop scans the
304 * parameter name. Internal whitespace is compressed, and an
305 * equal sign (=) terminates the token. Leading and trailing
306 * whitespace is discarded. The second loop scans the parameter
307 * value. When both have been successfully identified, they are
308 * passed to pfunc() for processing.
310 * ------------------------------------------------------------------------ **
313 int i
= 0; /* Position within bufr. */
314 int end
= 0; /* bufr[end] is current end-of-string. */
315 int vstart
= 0; /* Starting position of the parameter value. */
316 const char *func
= "params.c:Parameter() -";
318 /* Read the parameter name. */
319 while( 0 == vstart
) /* Loop until we've found the start of the value. */
322 if( i
> (InFile
->bSize
- 2) ) /* Ensure there's space for next char. */
326 tb
= talloc_realloc(InFile
, InFile
->bufr
, char, InFile
->bSize
+ BUFR_INC
);
329 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
333 InFile
->bSize
+= BUFR_INC
;
338 case '=': /* Equal sign marks end of param name. */
339 if( 0 == end
) /* Don't allow an empty name. */
341 DEBUG(0, ("%s Invalid parameter name in config. file.\n", func
));
344 InFile
->bufr
[end
++] = '\0'; /* Mark end of string & advance. */
345 i
= end
; /* New string starts here. */
346 vstart
= end
; /* New string is parameter value. */
347 InFile
->bufr
[i
] = '\0'; /* New string is nul, for now. */
350 case '\n': /* Find continuation char, else error. */
351 i
= Continuation( InFile
->bufr
, i
);
354 InFile
->bufr
[end
] = '\0';
355 DEBUG(1,("%s Ignoring badly formed line in configuration file: %s\n",
356 func
, InFile
->bufr
));
359 end
= ( (i
> 0) && (' ' == InFile
->bufr
[i
- 1]) ) ? (i
- 1) : (i
);
360 c
= mygetc( InFile
); /* Read past eoln. */
363 case '\0': /* Shouldn't have EOF within param name. */
365 InFile
->bufr
[i
] = '\0';
366 DEBUG(1,("%s Unexpected end-of-file at: %s\n", func
, InFile
->bufr
));
370 if( isspace( c
) ) /* One ' ' per whitespace region. */
372 InFile
->bufr
[end
] = ' ';
374 c
= EatWhitespace( InFile
);
376 else /* All others verbatim. */
378 InFile
->bufr
[i
++] = c
;
380 c
= mygetc( InFile
);
385 /* Now parse the value. */
386 c
= EatWhitespace( InFile
); /* Again, trim leading whitespace. */
387 while( (EOF
!=c
) && (c
> 0) )
390 if( i
> (InFile
->bSize
- 2) ) /* Make sure there's enough room. */
394 tb
= talloc_realloc(InFile
, InFile
->bufr
, char, InFile
->bSize
+ BUFR_INC
);
397 DEBUG(0, ("%s Memory re-allocation failure.", func
) );
401 InFile
->bSize
+= BUFR_INC
;
406 case '\r': /* Explicitly remove '\r' because the older */
407 c
= mygetc( InFile
); /* version called fgets_slash() which also */
408 break; /* removes them. */
410 case '\n': /* Marks end of value unless there's a '\'. */
411 i
= Continuation( InFile
->bufr
, i
);
416 for( end
= i
; (end
>= 0) && isspace((int)InFile
->bufr
[end
]); end
-- )
418 c
= mygetc( InFile
);
422 default: /* All others verbatim. Note that spaces do */
423 InFile
->bufr
[i
++] = c
; /* not advance <end>. This allows trimming */
424 if( !isspace( c
) ) /* of whitespace at the end of the line. */
426 c
= mygetc( InFile
);
430 InFile
->bufr
[end
] = '\0'; /* End of value. */
432 return( pfunc( InFile
->bufr
, &InFile
->bufr
[vstart
], userdata
) ); /* Pass name & value to pfunc(). */
435 static BOOL
Parse( myFILE
*InFile
,
436 BOOL (*sfunc
)(const char *, void *),
437 BOOL (*pfunc
)(const char *, const char *, void *),
439 /* ------------------------------------------------------------------------ **
440 * Scan & parse the input.
442 * Input: InFile - Input source.
443 * sfunc - Function to be called when a section name is scanned.
445 * pfunc - Function to be called when a parameter is scanned.
448 * Output: True if the file was successfully scanned, else False.
450 * Notes: The input can be viewed in terms of 'lines'. There are four
452 * Blank - May contain whitespace, otherwise empty.
453 * Comment - First non-whitespace character is a ';' or '#'.
454 * The remainder of the line is ignored.
455 * Section - First non-whitespace character is a '['.
456 * Parameter - The default case.
458 * ------------------------------------------------------------------------ **
463 c
= EatWhitespace( InFile
);
464 while( (EOF
!= c
) && (c
> 0) )
468 case '\n': /* Blank line. */
469 c
= EatWhitespace( InFile
);
472 case ';': /* Comment line. */
474 c
= EatComment( InFile
);
477 case '[': /* Section Header. */
478 if( !Section( InFile
, sfunc
, userdata
) )
480 c
= EatWhitespace( InFile
);
483 case '\\': /* Bogus backslash. */
484 c
= EatWhitespace( InFile
);
487 default: /* Parameter line. */
488 if( !Parameter( InFile
, pfunc
, c
, userdata
) )
490 c
= EatWhitespace( InFile
);
497 static myFILE
*OpenConfFile( const char *FileName
)
498 /* ------------------------------------------------------------------------ **
499 * Open a configuration file.
501 * Input: FileName - The pathname of the config file to be opened.
503 * Output: A pointer of type (char **) to the lines of the file
505 * ------------------------------------------------------------------------ **
508 const char *func
= "params.c:OpenConfFile() -";
511 ret
= talloc(talloc_autofree_context(), myFILE
);
512 if (!ret
) return NULL
;
514 ret
->buf
= file_load(FileName
, &ret
->size
, ret
);
515 if( NULL
== ret
->buf
)
518 ("%s Unable to open configuration file \"%s\":\n\t%s\n",
519 func
, FileName
, strerror(errno
)) );
530 BOOL
pm_process( const char *FileName
,
531 BOOL (*sfunc
)(const char *, void *),
532 BOOL (*pfunc
)(const char *, const char *, void *),
534 /* ------------------------------------------------------------------------ **
535 * Process the named parameter file.
537 * Input: FileName - The pathname of the parameter file to be opened.
538 * sfunc - A pointer to a function that will be called when
539 * a section name is discovered.
540 * pfunc - A pointer to a function that will be called when
541 * a parameter name and value are discovered.
543 * Output: TRUE if the file was successfully parsed, else FALSE.
545 * ------------------------------------------------------------------------ **
550 const char *func
= "params.c:pm_process() -";
552 InFile
= OpenConfFile( FileName
); /* Open the config file. */
556 DEBUG( 3, ("%s Processing configuration file \"%s\"\n", func
, FileName
) );
558 if( NULL
!= InFile
->bufr
) /* If we already have a buffer */
559 result
= Parse( InFile
, sfunc
, pfunc
, userdata
); /* (recursive call), then just */
562 else /* If we don't have a buffer */
563 { /* allocate one, then parse, */
564 InFile
->bSize
= BUFR_INC
; /* then free. */
565 InFile
->bufr
= talloc_array(InFile
, char, InFile
->bSize
);
566 if( NULL
== InFile
->bufr
)
568 DEBUG(0,("%s memory allocation failure.\n", func
));
569 myfile_close(InFile
);
572 result
= Parse( InFile
, sfunc
, pfunc
, userdata
);
577 myfile_close(InFile
);
579 if( !result
) /* Generic failure. */
581 DEBUG(0,("%s Failed. Error returned from params.c:parse().\n", func
));
585 return( True
); /* Generic success. */
588 /* -------------------------------------------------------------------------- */