[ANN] RegexKitLite 2.0

FROM : John Engelhart
DATE : Tue Jul 08 00:13:23 2008

I've just pushed out RegexKitLite 2.0.

Documentation: http://regexkit.sourceforge.net/RegexKitLite/index.html
Download: http://downloads.sourceforge.net/regexkit/RegexKitLite-2.0.tar.bz2
  (~40K)
Xcode 3.0 integrated docs: feed://regexkit.sourceforge.net/RegexKitLiteDocSets.atom

RegexKitLite is a lightweight NSString category extension that uses 
the ICU shared library that ships with Mac OS X as its regular 
expression pattern matching engine.

The major new features of RegexKitLite 2.0 are:

o The ability to split a string in to a NSArray with a regular 
expression.

o Search and replace using common $n capture substitution in the 
replacement string.

The following NSString methods were added:

- (NSArray *)componentsSeparatedByRegex:(NSString *)regex;
- (NSString *)stringByReplacingOccurrencesOfRegex:(NSString *)regex 
withString:(NSString *)replacement;

and the following NSMutableString method:

- (NSUInteger)replaceOccurrencesOfRegex:(NSString *)regex withString:
(NSString *)replacement;

Search and replace supports $n capture substitution, such as:

replacedString = [@"Search and replace all these words." 
stringByReplacingOccurrencesOfRegex:@"\\b(\\w+)\\b" withString:@"<<
$1>>"];

results in a replacedString of: @"<<Replace>> <<all>> <<these>> 
<<words>>."

Since regular expressions are often at the heart of a loop that is 
processing a large volume of data, I've made an effort to make things 
as fast as reasonably possible.  RegexKitLite tries to cache data that 
is expensive to create, such as compiling a regular expression or 
performing a conversion of a string from its native encoding in to 
UTF-16, which is the only encoding ICU knows how to deal with.  Little 
things like calling CFStringLength() or assigning a value indirectly 
through a __strong GC pointer add up pretty quickly when they're 
executed three million times in the course of processing a file.