ftp.nice.ch/pub/next/tools/preferences/FontSearch.0.85.NI.bs.tar.gz#/FontSearch.0.85.NI.bs/Source/MiscStringPatterns.m

This is MiscStringPatterns.m in view mode; [Download] [Up]

//
//	MiscStringPatterns.m -- Pattern matching and replacement routines
//		Written by Steve Hayman (c) 1994 by Steve Hayman.
//				Version 1.95  All rights reserved.
//		This notice may not be removed from this source code.
//
//	This object is included in the MiscKit by permission from the author
//	and its use is governed by the MiscKit license, found in the file
//	"LICENSE.rtf" in the MiscKit distribution.  Please refer to that file
//	for a list of all applicable permissions and restrictions.
//	

#import <misckit/MiscString.h>

@implementation MiscString(PatternMatching)
/*
 * Match strings against regular expressions, using re_compile 
 * Steve Hayman
 * November 23, 1993
 */
/*
 * Match a specified pattern.  Returns 1 or 0 depending on whether the
 * pattern is found in the indicated string.  Returns -1 if bogus regexp.
 * Will also optionally fill in strings with the portion of the destination
 * string before the match, the portion that matches, and the portion
 * after the match.
 */
#import <regex.h>

#ifdef DONT_COMPILE	// obsolete method
- (int) grep:(const char *)pattern caseSensitive:(BOOL)caseSens before:bstring middle:mstring after:astring
{
    
    struct regex *reg;
    int success;
    char * start, *end;
    MiscString *scratch;
    char *s = (char *)[self stringValue];
    
    if ( s == NULL )
	return NO;		// nothing matches an empty string
	
    // caseSensitive:YES means "fold case: NO", so we pass the opposite
    // to re_compile.
    
    reg = re_compile((char *)pattern, !caseSens);
    
    if ( reg  == NULL )
	return -1;		// bogus regular expression

    success = re_match( s, reg );
    
    switch( success ) {
    case 0:	// didn't match
    	free(reg);
	return 0;
    case -1:	// bogus regular expression 

    	 free(reg);

	 return -1;
	 
    default:	// matched.
	start = reg->start;
	end = reg->end;
	
	// fill in each of the various substrings, if desired
	
	// the part before the match
	if ( bstring ) {
	    scratch = [self midFrom:0 to:start - s - 1];
	    [bstring takeStringValue:scratch];
	    [scratch free];
	}
	// the part that matched
	
	if ( mstring ) {
	    scratch = [self midFrom:(start - s) to:(end - s) - 1];
	    [mstring takeStringValue:scratch];
	    [scratch free];
	}
	
	// the part after the match
	
	if ( astring ) {
	    scratch = [self midFrom: end - s  to: [self length] - 1];
	    [astring takeStringValue: scratch];
	    [scratch free];
	}
	free(reg);
	return 1;
    }
}
#endif

/*
 * Variants on the above.
 */

#ifdef DONT_COMPILE	// obsolete method
- (int) grep:(const char *)pattern caseSensitive:(BOOL)caseSens
{
    return [self grep:pattern caseSensitive:caseSens
    	 before:nil middle:nil after:nil];
}
#endif

#ifdef DONT_COMPILE	// obsolete method
- (int) grep:(const char *)pattern
{
    return [self grep:pattern caseSensitive:YES];
}
#endif

- (int) grepString:pattern caseSensitive:(BOOL)caseSens before:bstring middle:mstring after:astring
{
    return( [self grep:[pattern stringValue] caseSensitive:caseSens
    		before:bstring middle:mstring after:astring] );
    
}

- (int) grepString:pattern caseSensitive:(BOOL)caseSens
{
    return [self grepString:pattern caseSensitive:caseSens
    	 before:nil middle:nil after:nil];
}

- (int) grepString:pattern
{
    return [self grepString:pattern caseSensitive:YES];
}


- (int)replacePattern:(const char *)pattern caseSensitive:(BOOL)caseSens globally:(BOOL)glob with:(const char *)replacement
{
    id before = [[MiscString alloc] init];
    id after  = [[MiscString alloc] init];
    id middle = [[MiscString alloc] init];
    
    id newString = [[MiscString alloc] init];
    id newReplacement = [[MiscString alloc] init];
    id grepMe;
    
    int r;
    int replacements = 0;
    
    grepMe = self;
    
    // Do at least one replacement; if "glob" is TRUE, keep going
    // until we can't do any more.
    
    do {
    
	r = [grepMe grep:pattern caseSensitive:caseSens 
    		before:before middle:middle after:after];

	if ( r <= 0 )
	    break;	// no match - or no more matches
	    
	if ( [before stringValue] )
	    [newString concatenate:before];
	if ( replacement ) {
#ifdef FUTURE_FEATURE
	    /*
	     * ed has this notion of "&" on the right hand side of
	     * a substitution meaning "interpolate the text that was
	     * matched", i.e. "s/foo/&bar/" produces "foobar".
	     *
	     * In addition ed lets you mark out sub-expressions
	     * with \(\), and you can use \1, \2 ... to refer to
	     * the corresponding matched text.
	     *    s/\(A*\)\(B*\)/\2\1/
	     * which would turn AAAAABB into BBAAAAA
	     * This is ALMOST really easy to do here.
	     * To get "&" working it is ALMOST a matter of doing
	     *
	     *  [newReplacement setStringValue:replacement];
	     *  [newReplacement replacePattern:"&" 
	     *     caseSensitive:NO globally:YES
	     *     withString:middle];
	     *  [newString concatentate:newReplacement];
	     *
	     * which would be a neat use of recursion. 
	     * So why haven't I done this?  Well, you should also
	     * support "\&" meaning "a literal '&'" on the right
	     * hand side, and, well, I couldn't think of a quick
	     * two-line way to sneak that in.  Also you would get
	     * in trouble if "middle" contained a "&".
	     *
	     * Also, the regex structure contains
	     *    char *braslist[NBRA];
	     *    char *braelist[NBRA];
	     *
	     * These are pointers to the beginning and end of 
	     * parenthesized sub-expressions matched in the
	     * input text.  You could use these to implement
	     * replacements of \1, \2 ... \9, but, again, you would
	     * need to be careful that you weren't matching "\\1".
	     *
	     * Maybe next time.
	     * steve
	     */
	    
	    // "&" in the replacement string stands for the text
	    // that was matched - just like "ed"
	    [newReplacement setStringValue:replacement];
	    [newReplacement replacePattern:"&" caseSensitive:NO globally:YES
		withString:middle];
	

	    [newString concatenate:newReplacement];
#else
	    // do the simple-minded substitution in lieu of fanciness above
	    
	    [newString cat:replacement];
#endif
	}
	replacements ++;
	
	// next time around we match on the remainder of the string
	grepMe = after;
    } while ( glob );
    
   
    if ( [after stringValue] )
    	[newString concatenate:after];
    
    // If any changes were made, copy the new string.
    
    if ( replacements )
	[self takeStringValue:newString];

    [before free];
    [after free];
    [middle free];
    [newString free];
    [newReplacement free];
    
    // Return number of replacements made, or -1 if bogus regexp.
    return ( r < 0 ? r : replacements );
}

/*
 * Various other flavours of replacePattern
 */
 
- (int)replacePattern:(const char *)pattern caseSensitive:(BOOL)caseSens globally:(BOOL)glob withString:replacement
{
    return ( [self replacePattern:pattern caseSensitive:caseSens
    		globally:glob with:[replacement stringValue] ]);
}

- (int)replacePatternString:pattern caseSensitive:(BOOL)caseSens globally:(BOOL)glob with:(const char *)replacement
{
    return [self replacePattern:[pattern stringValue] caseSensitive:caseSens globally:glob with:replacement];
}
- (int)replacePatternString:pattern caseSensitive:(BOOL)caseSens globally:(BOOL)glob withString:replacement
{
    return ( [self replacePattern:[pattern stringValue] caseSensitive:caseSens
    		globally:glob
		with:[replacement stringValue]] );
}

	    
@end

These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Netfuture.ch.