2 This modules is based on the params.c module from Samba, written by Karl Auer
3 and much modifed by Christopher Hertel.
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program; if not, write to the Free Software
17 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
19 * -------------------------------------------------------------------------- **
23 * -------------------------------------------------------------------------- **
25 * This module performs lexical analysis and initial parsing of a
26 * Windows-like parameter file. It recognizes and handles four token
27 * types: section-name, parameter-name, parameter-value, and
28 * end-of-file. Comments and line continuation are handled
31 * The entry point to the module is function pm_process(). This
32 * function opens the source file, calls the Parse() function to parse
33 * the input, and then closes the file when either the EOF is reached
34 * or a fatal error is encountered.
36 * A sample parameter file might look like this:
39 * parameter one = value string
40 * parameter two = another value
42 * new parameter = some value or t'other
44 * The parameter file is divided into sections by section headers:
45 * section names enclosed in square brackets (eg. [section one]).
46 * Each section contains parameter lines, each of which consist of a
47 * parameter name and value delimited by an equal sign. Roughly, the
50 * <file> :== { <section> } EOF
52 * <section> :== <section header> { <parameter line> }
54 * <section header> :== '[' NAME ']'
56 * <parameter line> :== NAME '=' VALUE '\n'
58 * Blank lines and comment lines are ignored. Comment lines are lines
59 * beginning with either a semicolon (';') or a pound sign ('#').
61 * All whitespace in section names and parameter names is compressed
62 * to single spaces. Leading and trailing whitespace is stipped from
63 * both names and values.
65 * Only the first equals sign in a parameter line is significant.
66 * Parameter values may contain equals signs, square brackets and
67 * semicolons. Internal whitespace is retained in parameter values,
68 * with the exception of the '\r' character, which is stripped for
69 * historic reasons. Parameter names may not start with a left square
70 * bracket, an equal sign, a pound sign, or a semicolon, because these
71 * are used to identify other tokens.
73 * -------------------------------------------------------------------------- **
80 #define Realloc realloc
82 /* -------------------------------------------------------------------------- **
89 /* -------------------------------------------------------------------------- **
92 * bufr - pointer to a global buffer. This is probably a kludge,
93 * but it was the nicest kludge I could think of (for now).
94 * bSize - The size of the global buffer <bufr>.
97 static char *bufr = NULL;
100 /* -------------------------------------------------------------------------- **
104 static int EatWhitespace( FILE *InFile )
105 /* ------------------------------------------------------------------------ **
106 * Scan past whitespace (see ctype(3C)) and return the first non-whitespace
107 * character, or newline, or EOF.
109 * Input: InFile - Input source.
111 * Output: The next non-whitespace character in the input stream.
113 * Notes: Because the config files use a line-oriented grammar, we
114 * explicitly exclude the newline character from the list of
115 * whitespace characters.
116 * - Note that both EOF (-1) and the nul character ('\0') are
117 * considered end-of-file markers.
119 * ------------------------------------------------------------------------ **
124 for( c = getc( InFile ); isspace( c ) && ('\n' != c); c = getc( InFile ) )
127 } /* EatWhitespace */
129 static int EatComment( FILE *InFile )
130 /* ------------------------------------------------------------------------ **
131 * Scan to the end of a comment.
133 * Input: InFile - Input source.
135 * Output: The character that marks the end of the comment. Normally,
136 * this will be a newline, but it *might* be an EOF.
138 * Notes: Because the config files use a line-oriented grammar, we
139 * explicitly exclude the newline character from the list of
140 * whitespace characters.
141 * - Note that both EOF (-1) and the nul character ('\0') are
142 * considered end-of-file markers.
144 * ------------------------------------------------------------------------ **
149 for( c = getc( InFile ); ('\n'!=c) && (EOF!=c) && (c>0); c = getc( InFile ) )
154 static int Continuation( char *line, int pos )
155 /* ------------------------------------------------------------------------ **
156 * Scan backards within a string to discover if the last non-whitespace
157 * character is a line-continuation character ('\\').
159 * Input: line - A pointer to a buffer containing the string to be
161 * pos - This is taken to be the offset of the end of the
162 * string. This position is *not* scanned.
164 * Output: The offset of the '\\' character if it was found, or -1 to
165 * indicate that it was not.
167 * ------------------------------------------------------------------------ **
171 while( (pos >= 0) && isspace(line[pos]) )
174 return( ((pos >= 0) && ('\\' == line[pos])) ? pos : -1 );
178 static BOOL Section( FILE *InFile, BOOL (*sfunc)(char *) )
179 /* ------------------------------------------------------------------------ **
180 * Scan a section name, and pass the name to function sfunc().
182 * Input: InFile - Input source.
183 * sfunc - Pointer to the function to be called if the section
184 * name is successfully read.
186 * Output: True if the section name was read and True was returned from
187 * <sfunc>. False if <sfunc> failed or if a lexical error was
190 * ------------------------------------------------------------------------ **
196 char *func = "params.c:Section() -";
198 i = 0; /* <i> is the offset of the next free byte in bufr[] and */
199 end = 0; /* <end> is the current "end of string" offset. In most */
200 /* cases these will be the same, but if the last */
201 /* character written to bufr[] is a space, then <end> */
202 /* will be one less than <i>. */
204 c = EatWhitespace( InFile ); /* We've already got the '['. Scan */
205 /* past initial white space. */
207 while( (EOF != c) && (c > 0) )
210 /* Check that the buffer is big enough for the next character. */
211 if( i > (bSize - 2) )
214 bufr = Realloc( bufr, bSize );
217 rprintf(FERROR, "%s Memory re-allocation failure.", func);
222 /* Handle a single character. */
225 case ']': /* Found the closing bracket. */
227 if( 0 == end ) /* Don't allow an empty name. */
229 rprintf(FERROR, "%s Empty section name in configuration file.\n", func );
232 if( !sfunc( bufr ) ) /* Got a valid name. Deal with it. */
234 (void)EatComment( InFile ); /* Finish off the line. */
237 case '\n': /* Got newline before closing ']'. */
238 i = Continuation( bufr, i ); /* Check for line continuation. */
242 rprintf(FERROR, "%s Badly formed line in configuration file: %s\n",
246 end = ( (i > 0) && (' ' == bufr[i - 1]) ) ? (i - 1) : (i);
247 c = getc( InFile ); /* Continue with next line. */
250 default: /* All else are a valid name chars. */
251 if( isspace( c ) ) /* One space per whitespace region. */
255 c = EatWhitespace( InFile );
257 else /* All others copy verbatim. */
266 /* We arrive here if we've met the EOF before the closing bracket. */
267 rprintf(FERROR, "%s Unexpected EOF in the configuration file: %s\n", func, bufr );
271 static BOOL Parameter( FILE *InFile, BOOL (*pfunc)(char *, char *), int c )
272 /* ------------------------------------------------------------------------ **
273 * Scan a parameter name and value, and pass these two fields to pfunc().
275 * Input: InFile - The input source.
276 * pfunc - A pointer to the function that will be called to
277 * process the parameter, once it has been scanned.
278 * c - The first character of the parameter name, which
279 * would have been read by Parse(). Unlike a comment
280 * line or a section header, there is no lead-in
281 * character that can be discarded.
283 * Output: True if the parameter name and value were scanned and processed
284 * successfully, else False.
286 * Notes: This function is in two parts. The first loop scans the
287 * parameter name. Internal whitespace is compressed, and an
288 * equal sign (=) terminates the token. Leading and trailing
289 * whitespace is discarded. The second loop scans the parameter
290 * value. When both have been successfully identified, they are
291 * passed to pfunc() for processing.
293 * ------------------------------------------------------------------------ **
296 int i = 0; /* Position within bufr. */
297 int end = 0; /* bufr[end] is current end-of-string. */
298 int vstart = 0; /* Starting position of the parameter value. */
299 char *func = "params.c:Parameter() -";
301 /* Read the parameter name. */
302 while( 0 == vstart ) /* Loop until we've found the start of the value. */
305 if( i > (bSize - 2) ) /* Ensure there's space for next char. */
308 bufr = Realloc( bufr, bSize );
311 rprintf(FERROR, "%s Memory re-allocation failure.", func) ;
318 case '=': /* Equal sign marks end of param name. */
319 if( 0 == end ) /* Don't allow an empty name. */
321 rprintf(FERROR, "%s Invalid parameter name in config. file.\n", func );
324 bufr[end++] = '\0'; /* Mark end of string & advance. */
325 i = end; /* New string starts here. */
326 vstart = end; /* New string is parameter value. */
327 bufr[i] = '\0'; /* New string is nul, for now. */
330 case '\n': /* Find continuation char, else error. */
331 i = Continuation( bufr, i );
335 rprintf(FERROR, "%s Ignoring badly formed line in configuration file: %s\n",
339 end = ( (i > 0) && (' ' == bufr[i - 1]) ) ? (i - 1) : (i);
340 c = getc( InFile ); /* Read past eoln. */
343 case '\0': /* Shouldn't have EOF within param name. */
346 rprintf(FERROR, "%s Unexpected end-of-file at: %s\n", func, bufr );
350 if( isspace( c ) ) /* One ' ' per whitespace region. */
354 c = EatWhitespace( InFile );
356 else /* All others verbatim. */
365 /* Now parse the value. */
366 c = EatWhitespace( InFile ); /* Again, trim leading whitespace. */
367 while( (EOF !=c) && (c > 0) )
370 if( i > (bSize - 2) ) /* Make sure there's enough room. */
373 bufr = Realloc( bufr, bSize );
376 rprintf(FERROR, "%s Memory re-allocation failure.", func) ;
383 case '\r': /* Explicitly remove '\r' because the older */
384 c = getc( InFile ); /* version called fgets_slash() which also */
385 break; /* removes them. */
387 case '\n': /* Marks end of value unless there's a '\'. */
388 i = Continuation( bufr, i );
393 for( end = i; (end >= 0) && isspace(bufr[end]); end-- )
399 default: /* All others verbatim. Note that spaces do */
400 bufr[i++] = c; /* not advance <end>. This allows trimming */
401 if( !isspace( c ) ) /* of whitespace at the end of the line. */
407 bufr[end] = '\0'; /* End of value. */
409 return( pfunc( bufr, &bufr[vstart] ) ); /* Pass name & value to pfunc(). */
412 static BOOL Parse( FILE *InFile,
413 BOOL (*sfunc)(char *),
414 BOOL (*pfunc)(char *, char *) )
415 /* ------------------------------------------------------------------------ **
416 * Scan & parse the input.
418 * Input: InFile - Input source.
419 * sfunc - Function to be called when a section name is scanned.
421 * pfunc - Function to be called when a parameter is scanned.
424 * Output: True if the file was successfully scanned, else False.
426 * Notes: The input can be viewed in terms of 'lines'. There are four
428 * Blank - May contain whitespace, otherwise empty.
429 * Comment - First non-whitespace character is a ';' or '#'.
430 * The remainder of the line is ignored.
431 * Section - First non-whitespace character is a '['.
432 * Parameter - The default case.
434 * ------------------------------------------------------------------------ **
439 c = EatWhitespace( InFile );
440 while( (EOF != c) && (c > 0) )
444 case '\n': /* Blank line. */
445 c = EatWhitespace( InFile );
448 case ';': /* Comment line. */
450 c = EatComment( InFile );
453 case '[': /* Section Header. */
454 if( !Section( InFile, sfunc ) )
456 c = EatWhitespace( InFile );
459 case '\\': /* Bogus backslash. */
460 c = EatWhitespace( InFile );
463 default: /* Parameter line. */
464 if( !Parameter( InFile, pfunc, c ) )
466 c = EatWhitespace( InFile );
473 static FILE *OpenConfFile( char *FileName )
474 /* ------------------------------------------------------------------------ **
475 * Open a configuration file.
477 * Input: FileName - The pathname of the config file to be opened.
479 * Output: A pointer of type (FILE *) to the opened file, or NULL if the
480 * file could not be opened.
482 * ------------------------------------------------------------------------ **
486 char *func = "params.c:OpenConfFile() -";
488 if( NULL == FileName || 0 == *FileName )
490 rprintf(FERROR,"%s No configuration filename specified.\n", func);
494 OpenedFile = fopen( FileName, "r" );
495 if( NULL == OpenedFile )
497 rprintf(FERROR,"%s Unable to open configuration file \"%s\":\n\t%s\n",
498 func, FileName, strerror(errno));
501 return( OpenedFile );
504 BOOL pm_process( char *FileName,
505 BOOL (*sfunc)(char *),
506 BOOL (*pfunc)(char *, char *) )
507 /* ------------------------------------------------------------------------ **
508 * Process the named parameter file.
510 * Input: FileName - The pathname of the parameter file to be opened.
511 * sfunc - A pointer to a function that will be called when
512 * a section name is discovered.
513 * pfunc - A pointer to a function that will be called when
514 * a parameter name and value are discovered.
516 * Output: TRUE if the file was successfully parsed, else FALSE.
518 * ------------------------------------------------------------------------ **
523 char *func = "params.c:pm_process() -";
525 InFile = OpenConfFile( FileName ); /* Open the config file. */
529 if( NULL != bufr ) /* If we already have a buffer */
530 result = Parse( InFile, sfunc, pfunc ); /* (recursive call), then just */
533 else /* If we don't have a buffer */
534 { /* allocate one, then parse, */
535 bSize = BUFR_INC; /* then free. */
536 bufr = (char *)malloc( bSize );
539 rprintf(FERROR,"%s memory allocation failure.\n", func);
543 result = Parse( InFile, sfunc, pfunc );
551 if( !result ) /* Generic failure. */
553 rprintf(FERROR,"%s Failed. Error returned from params.c:parse().\n", func);
557 return( True ); /* Generic success. */
560 /* -------------------------------------------------------------------------- */