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 )
199 while( (pos >= 0) && isspace(line[pos]) )
202 return (((pos >= 0) && ('\\' == line[pos])) ? pos : -1 );
207 static BOOL Section( myFILE *InFile, BOOL (*sfunc)(char *) )
208 /* ------------------------------------------------------------------------ **
209 * Scan a section name, and pass the name to function sfunc().
211 * Input: InFile - Input source.
212 * sfunc - Pointer to the function to be called if the section
213 * name is successfully read.
215 * Output: True if the section name was read and True was returned from
216 * <sfunc>. False if <sfunc> failed or if a lexical error was
219 * ------------------------------------------------------------------------ **
225 char *func = "params.c:Section() -";
227 i = 0; /* <i> is the offset of the next free byte in bufr[] and */
228 end = 0; /* <end> is the current "end of string" offset. In most */
229 /* cases these will be the same, but if the last */
230 /* character written to bufr[] is a space, then <end> */
231 /* will be one less than <i>. */
233 c = EatWhitespace( InFile ); /* We've already got the '['. Scan */
234 /* past initial white space. */
236 while( (EOF != c) && (c > 0) )
239 /* Check that the buffer is big enough for the next character. */
240 if( i > (bSize - 2) )
243 bufr = Realloc( bufr, bSize );
246 DEBUG(0, ("%s Memory re-allocation failure.", func) );
251 /* Handle a single character. */
254 case ']': /* Found the closing bracket. */
256 if( 0 == end ) /* Don't allow an empty name. */
258 DEBUG(0, ("%s Empty section name in configuration file.\n", func ));
261 if( !sfunc(bufr) ) /* Got a valid name. Deal with it. */
263 (void)EatComment( InFile ); /* Finish off the line. */
266 case '\n': /* Got newline before closing ']'. */
267 i = Continuation( bufr, i ); /* Check for line continuation. */
271 DEBUG(0, ("%s Badly formed line in configuration file: %s\n",
275 end = ( (i > 0) && (' ' == bufr[i - 1]) ) ? (i - 1) : (i);
276 c = mygetc( InFile ); /* Continue with next line. */
279 default: /* All else are a valid name chars. */
280 if( isspace( c ) ) /* One space per whitespace region. */
284 c = EatWhitespace( InFile );
286 else /* All others copy verbatim. */
290 c = mygetc( InFile );
295 /* We arrive here if we've met the EOF before the closing bracket. */
296 DEBUG(0, ("%s Unexpected EOF in the configuration file: %s\n", func, bufr ));
300 static BOOL Parameter( myFILE *InFile, BOOL (*pfunc)(char *, char *), int c )
301 /* ------------------------------------------------------------------------ **
302 * Scan a parameter name and value, and pass these two fields to pfunc().
304 * Input: InFile - The input source.
305 * pfunc - A pointer to the function that will be called to
306 * process the parameter, once it has been scanned.
307 * c - The first character of the parameter name, which
308 * would have been read by Parse(). Unlike a comment
309 * line or a section header, there is no lead-in
310 * character that can be discarded.
312 * Output: True if the parameter name and value were scanned and processed
313 * successfully, else False.
315 * Notes: This function is in two parts. The first loop scans the
316 * parameter name. Internal whitespace is compressed, and an
317 * equal sign (=) terminates the token. Leading and trailing
318 * whitespace is discarded. The second loop scans the parameter
319 * value. When both have been successfully identified, they are
320 * passed to pfunc() for processing.
322 * ------------------------------------------------------------------------ **
325 int i = 0; /* Position within bufr. */
326 int end = 0; /* bufr[end] is current end-of-string. */
327 int vstart = 0; /* Starting position of the parameter value. */
328 char *func = "params.c:Parameter() -";
330 /* Read the parameter name. */
331 while( 0 == vstart ) /* Loop until we've found the start of the value. */
334 if( i > (bSize - 2) ) /* Ensure there's space for next char. */
337 bufr = Realloc( bufr, bSize );
340 DEBUG(0, ("%s Memory re-allocation failure.", func) );
347 case '=': /* Equal sign marks end of param name. */
348 if( 0 == end ) /* Don't allow an empty name. */
350 DEBUG(0, ("%s Invalid parameter name in config. file.\n", func ));
353 bufr[end++] = '\0'; /* Mark end of string & advance. */
354 i = end; /* New string starts here. */
355 vstart = end; /* New string is parameter value. */
356 bufr[i] = '\0'; /* New string is nul, for now. */
359 case '\n': /* Find continuation char, else error. */
360 i = Continuation( bufr, i );
364 DEBUG(1,("%s Ignoring badly formed line in configuration file: %s\n",
368 end = ( (i > 0) && (' ' == bufr[i - 1]) ) ? (i - 1) : (i);
369 c = mygetc( InFile ); /* Read past eoln. */
372 case '\0': /* Shouldn't have EOF within param name. */
375 DEBUG(1,("%s Unexpected end-of-file at: %s\n", func, bufr ));
379 if( isspace( c ) ) /* One ' ' per whitespace region. */
383 c = EatWhitespace( InFile );
385 else /* All others verbatim. */
389 c = mygetc( InFile );
394 /* Now parse the value. */
395 c = EatWhitespace( InFile ); /* Again, trim leading whitespace. */
396 while( (EOF !=c) && (c > 0) )
399 if( i > (bSize - 2) ) /* Make sure there's enough room. */
402 bufr = Realloc( bufr, bSize );
405 DEBUG(0, ("%s Memory re-allocation failure.", func) );
412 case '\r': /* Explicitly remove '\r' because the older */
413 c = mygetc( InFile ); /* version called fgets_slash() which also */
414 break; /* removes them. */
416 case '\n': /* Marks end of value unless there's a '\'. */
417 i = Continuation( bufr, i );
422 for( end = i; (end >= 0) && isspace(bufr[end]); end-- )
424 c = mygetc( InFile );
428 default: /* All others verbatim. Note that spaces do */
429 bufr[i++] = c; /* not advance <end>. This allows trimming */
430 if( !isspace( c ) ) /* of whitespace at the end of the line. */
432 c = mygetc( InFile );
436 bufr[end] = '\0'; /* End of value. */
438 return( pfunc( bufr, &bufr[vstart] ) ); /* Pass name & value to pfunc(). */
441 static BOOL Parse( myFILE *InFile,
442 BOOL (*sfunc)(char *),
443 BOOL (*pfunc)(char *, char *) )
444 /* ------------------------------------------------------------------------ **
445 * Scan & parse the input.
447 * Input: InFile - Input source.
448 * sfunc - Function to be called when a section name is scanned.
450 * pfunc - Function to be called when a parameter is scanned.
453 * Output: True if the file was successfully scanned, else False.
455 * Notes: The input can be viewed in terms of 'lines'. There are four
457 * Blank - May contain whitespace, otherwise empty.
458 * Comment - First non-whitespace character is a ';' or '#'.
459 * The remainder of the line is ignored.
460 * Section - First non-whitespace character is a '['.
461 * Parameter - The default case.
463 * ------------------------------------------------------------------------ **
468 c = EatWhitespace( InFile );
469 while( (EOF != c) && (c > 0) )
473 case '\n': /* Blank line. */
474 c = EatWhitespace( InFile );
477 case ';': /* Comment line. */
479 c = EatComment( InFile );
482 case '[': /* Section Header. */
483 if( !Section( InFile, sfunc ) )
485 c = EatWhitespace( InFile );
488 case '\\': /* Bogus backslash. */
489 c = EatWhitespace( InFile );
492 default: /* Parameter line. */
493 if( !Parameter( InFile, pfunc, c ) )
495 c = EatWhitespace( InFile );
502 static myFILE *OpenConfFile( char *FileName )
503 /* ------------------------------------------------------------------------ **
504 * Open a configuration file.
506 * Input: FileName - The pathname of the config file to be opened.
508 * Output: A pointer of type (char **) to the lines of the file
510 * ------------------------------------------------------------------------ **
513 char *func = "params.c:OpenConfFile() -";
514 extern BOOL in_client;
515 int lvl = in_client?1:0;
518 ret = (myFILE *)malloc(sizeof(*ret));
519 if (!ret) return NULL;
521 ret->buf = file_load(FileName, &ret->size);
522 if( NULL == ret->buf )
525 ("%s Unable to open configuration file \"%s\":\n\t%s\n",
526 func, FileName, strerror(errno)) );
535 BOOL pm_process( char *FileName,
536 BOOL (*sfunc)(char *),
537 BOOL (*pfunc)(char *, char *) )
538 /* ------------------------------------------------------------------------ **
539 * Process the named parameter file.
541 * Input: FileName - The pathname of the parameter file to be opened.
542 * sfunc - A pointer to a function that will be called when
543 * a section name is discovered.
544 * pfunc - A pointer to a function that will be called when
545 * a parameter name and value are discovered.
547 * Output: TRUE if the file was successfully parsed, else FALSE.
549 * ------------------------------------------------------------------------ **
554 char *func = "params.c:pm_process() -";
556 InFile = OpenConfFile( FileName ); /* Open the config file. */
560 DEBUG( 3, ("%s Processing configuration file \"%s\"\n", func, FileName) );
562 if( NULL != bufr ) /* If we already have a buffer */
563 result = Parse( InFile, sfunc, pfunc ); /* (recursive call), then just */
566 else /* If we don't have a buffer */
567 { /* allocate one, then parse, */
568 bSize = BUFR_INC; /* then free. */
569 bufr = (char *)malloc( bSize );
572 DEBUG(0,("%s memory allocation failure.\n", func));
573 myfile_close(InFile);
576 result = Parse( InFile, sfunc, pfunc );
582 myfile_close(InFile);
584 if( !result ) /* Generic failure. */
586 DEBUG(0,("%s Failed. Error returned from params.c:parse().\n", func));
590 return( True ); /* Generic success. */
593 /* -------------------------------------------------------------------------- */