RegexKitLite

Lightweight Objective-C Regular Expressions for Mac OS X using the ICU Library
2010-04-18, v4.0

Introduction to RegexKitLite

This document introduces RegexKitLite for Mac OS X. RegexKitLite enables easy access to regular expressions by providing a number of additions to the standard Foundation NSString class. RegexKitLite acts as a bridge between the NSString class and the regular expression engine in the International Components for Unicode, or ICU, dynamic shared library that is shipped with Mac OS X.

Highlights

Who Should Read This Document

This document is intended for readers who would like to be able to use regular expressions in their Objective-C applications, whether those applications are for the Mac OS X desktop, or for the iPhone.

This document, and RegexKitLite, is also intended for anyone who has the need to search and manipulate NSString objects. If you've ever used the NSScanner, NSCharacterSet, and NSPredicate classes, or any of the the NSString rangeOf… methods, RegexKitLite is for you.

Regular expressions are a powerful way to search, match and extract, and manipulate strings. RegexKitLite can perform many of the same operations that NSScanner, NSCharacterSet, and NSPredicate perform, and usually do it with far fewer lines of code. As an example, RegexKitLite Cookbook - Parsing CSV Data contains an example that is just over a dozen lines, but is a full featured CSV, or Comma Separated Value, parser that takes a CSV input and turns it in to a NSArray of NSArrays.

Organization of This Document

This document follows the conventions and styles used by Apples documentation and is divided in to two main parts:

Additional information can be found in the following sections:

Supporting RegexKitLite through Financial Donations

A significant amount of time and effort has gone in to the development of RegexKitLite. Even though it is distributed under the terms of the BSD License, you are encouraged to contribute financially if you are using RegexKitLite in a profitable commercial application. Should you decide to contribute to RegexKitLite, please keep the following in mind:

  • What it would have cost you in terms of hours, or consultant fees, to develop similar functionality.
  • The Ohloh.net metrics do not factor in the cost of writing documentation, which is where most of the effort is spent.
  • The target audience for RegexKitLite is very small, so there are relatively few units "sold".

You can contribute by visiting SourceForge.net's donation page for RegexKitLite.

Metrics by Ohloh.net
Support This Project
Important:

You are always required to acknowledge the use of RegexKitLite in your product as specified in the terms of the BSD License.

Download

You can download RegexKitLite distribution that corresponds to this version of the documentation here— RegexKitLite-4.0.tar.bz2 (139.1K). To be automatically notified when a new version of RegexKitLite is available, add the RegexKitLite documentation feed to Xcode.

PDF Documentation

This document is available in PDF format: RegexKitLite-4.0.pdf (1.1M).

Note:

If you wish to print this document, it is recommend that you use the PDF version.

Reporting Bugs

You can file bug reports, or review submitted bugs, at the following URL: http://sourceforge.net/tracker/?group_id=204582&atid=990188

Note:

Anonymous bug reports are no longer accepted due to spam. A SourceForge.net account is required to file a bug report.

Contacting The Author

The author can be contacted at regexkitlite@gmail.com.

RegexKitLite Overview

While RegexKitLite is not a descendant of the RegexKit.framework source code, it does provide a small subset of RegexKits NSString methods for performing various regular expression tasks. These include determining the range that a regular expression matches within a string, easily creating a new string from the results of a match, splitting a string in to a NSArray with a regular expression, and performing search and replace operations with regular expressions using common $n substitution syntax.

RegexKitLite uses the regular expression provided by the ICU library that ships with Mac OS X. The two files, RegexKitLite.h and RegexKitLite.m, and linking against the /usr/lib/libicucore.dylib ICU shared library is all that is required. Adding RegexKitLite to your project only adds a few kilobytes of overhead to your applications size and typically only requires a few kilobytes of memory at run-time. Since a regular expression must first be compiled by the ICU library before it can be used, RegexKitLite keeps a small 4-way set associative cache with a least recently used replacement policy of the compiled regular expressions.

See Also

Official Support from Apple for ICU Regular Expressions

Mac OS X

As of Mac OS X 10.6, the author is not aware of any official support from Apple for linking to the libicucore.dylib library. On the other hand, the author is unaware of any official prohibition against it, either. Linking to the ICU library and making use of the ICU regular expression API is slightly different than making use of private, undocumented API's. There are a number of very good reasons why you shouldn't use private, undocumented API's, such as:

The ICU library, on the other hand, contains a "published, public API" in which the ICU developers have committed to supporting in later releases, and RegexKitLite uses only these public APIs. One could argue that Apple is not obligated to continue to include the ICU library in later versions of Mac OS X, but this seems unlikely for a number of reasons which will not be discussed here. With the introduction of iPhone OS 3.2, Apple now officially supports iPhone applications linking to the ICU library for the purpose of using its regular expression functionality. This is encouraging news for Mac OS X developers if one assumes that Apple will try to keep some kind of parity between the iPhone OS and Mac OS X API's.

iPhone OS < 3.2

Prior to iPhone OS 3.2, there was never any official support from Apple for linking to the libicucore.dylib library. It was unclear if linking to the library would violate the iPhone OS SDK Agreement prohibition against using undocumented API's, but a large number of iPhone applications choose to use RegexKitLite, and the author is not aware of a single rejection because of it.

iPhone OS ≥ 3.2

Starting with iPhone OS 3.2, Apple now officially allows iPhone OS applications to link with the ICU library. The ICU library contains a lot of functionality for dealing with internationalization and localization, but Apple only officially permits the use of the ICU Regular Expression functionality.

Apple also provides a way to use ICU based regular expressions from Foundation by adding a new option to NSStringCompareOptionsNSRegularExpressionSearch. This new option can be used with the NSString rangeOfString:options: method, and the following example of its usage is given:

// finds phone number in format nnn-nnn-nnnn NSRange r; NSString *regEx = @"{3}-[0-9]{3}-[0-9]{4}"; r = [textView.text rangeOfString:regEx options:NSRegularExpressionSearch]; if (r.location != NSNotFound) { NSLog(@"Phone number is %@", [textView.text substringWithRange:r]); } else { NSLog(@"Not found."); }

At this time, rangeOfString:options: is the only regular expression functionality Apple has added to Foundation and capture groups in a regular expression are not supported. Apple also gives the following note:

Note:

As noted in "ICU Regular-Expression Support," the ICU libraries related to regular expressions are included in iPhone OS 3.2. However, you should only use the ICU facilities if the NSString alternative is not sufficient for your needs.

RegexKitLite provides a much richer API, such as the automatic extraction of a match as a NSString. Using RegexKitLite, the example can be rewritten as:

// finds phone number in format nnn-nnn-nnnn NSString *regEx = @"{3}-[0-9]{3}-[0-9]{4}"; NSString *match = [textView.text stringByMatching:regEx]; if ([match isEqual:@""] == NO) { NSLog(@"Phone number is %@", match); } else { NSLog(@"Not found."); }

What's more, RegexKitLite provides easy access to all the matches of a regular expression in a NSString:

// finds phone number in format nnn-nnn-nnnn NSString *regEx = @"{3}-[0-9]{3}-[0-9]{4}"; for(NSString *match in [textView.text componentsMatchedByRegex:regEx]) { NSLog(@"Phone number is %@", match); }

To do the same thing using just NSRegularExpressionSearch would require significantly more code and effort on your part. RegexKitLite also provides powerful search and replace functionality:

// finds phone number in format nnn-nnn-nnnn NSString *regEx = @"({3})-([0-9]{3}-[0-9]{4})"; // and transforms the phone number in to the format of (nnn) nnn-nnnn NSString *replaced = [textView.text stringByReplacingOccurrencesOfRegex:regEx withString:@"($1) $2"];

RegexKitLite also has a number of performance enhancing features built in such as caching compiled regular expressions. Although the author does not have any benchmarks comparing NSRegularExpressionSearch to RegexKitLite, it is likely that RegexKitLite outperforms NSRegularExpressionSearch.

See Also

The iPhone 4.0 SDK Agreement

While iPhone OS 3.2 included official, Apple sanctioned use of linking of the ICU library for the purposes of using the ICU regular expression engine, the iPhone OS 4.0 SDK included the following change to the iPhone OS SDK Agreement:

3.3.1

Applications may only use Documented APIs in the manner prescribed by Apple and must not use or call any private APIs. Applications must be originally written in Objective-C, C, C++, or JavaScript as executed by the iPhone OS WebKit engine, and only code written in C, C++, and Objective-C may compile and directly link against the Documented APIs (e.g., Applications that link to Documented APIs through an intermediary translation or compatibility layer or tool are prohibited).

This raises a number of obvious questions:

There is considerable speculation as to what is covered by this change, but at the time of this writing, there is no empirical evidence or official guidelines from Apple to make any kind of an informed decision as to whether or not the use of RegexKitLite would violate 3.3.1. It is the authors opinion that RegexKitLite could be considered as a compatibility layer between NSString and the now Documented APIs for regular expressions in the ICU library.

It is widely speculated that the motivation for the change to 3.3.1 was to prevent the development of Flash applications for the iPhone. The author believes that most reasonable people would consider the application of compatibility layer in this context to mean something entirely different than what it means when applied to RegexKitLite.

At this time, the author is not aware of a single iPhone application that has been rejected due to the use of RegexKitLite. If your application is rejected due to the use of RegexKitLite, please let the author know by emailing regexkitlite@gmail.com. As always, CAVEAT EMPTOR.

The Difference Between RegexKit.framework and RegexKitLite

RegexKit.framework and RegexKitLite are two different projects. In retrospect, RegexKitLite should have been given a more distinctive name. Below is a table summarizing some of the key differences between the two:

 RegexKit.frameworkRegexKitLite
Regex LibraryPCREICU
Library IncludedYes, built into framework object file.No, provided by Mac OS X.
Library Linked AsStatically linked into framework.Dynamically linked to /usr/lib/libicucore.dylib.
Compiled SizeApproximately 371KB per architecture.Very small, approximately 16KB—20KB per architecture.
StyleExternal, linked to framework.Compiled directly in to final executable.
Feature SetLarge, with additions to many classes.Minimal, NSString only.
-
Version 0.6.0. About half of the 371KB is the PCRE library.
The default distribution framework shared library file is 1.4MB in size and includes the ppc, ppc64, i386, and x86_64 architectures.
If 64-bit support is removed, the framework shared library file size drops to 664KB.
-
Since the ICU library is part of Mac OS X, it does not add to the final size.

Compiled Regular Expression Cache

The NSString that contains the regular expression must be compiled in to an ICU URegularExpression. This can be an expensive, time consuming step, and the compiled regular expression can be reused again in another search, even if the strings to be searched are different. Therefore RegexKitLite keeps a small cache of recently compiled regular expressions.

The cache is organized as a 4-way set associative cache, and the size of the cache can be tuned with the pre-processor define RKL_CACHE_SIZE. The default cache size, which should always be a prime number, is set to 13. Since the cache is 4-way set associative, the total number of compiled regular expressions that can be cached is RKL_CACHE_SIZE times four, for a total of 13 * 4, or 52. The NSString regexString is mapped to a cache set using modular arithmetic: Cache set ≡ [regexString hash] mod RKL_CACHE_SIZE, i.e. cacheSet = [regexString hash] % 13;. Since RegexKitLite uses Core Foundation, this is actually coded as cacheSet = CFHash(regexString) % RKL_CACHE_SIZE;.

Each of the four "ways" of a cache set are checked to see if it contains a NSString that was used to create the compiled regular expression that is identical to the NSString for the regular expression that is being checked. If there is an exact match, then the matching "way" is updated as the most recently used, and the compiled regular expression is used as-is. Otherwise, the least recently used, or LRU, "way" in the cache set is cleared and replaced with the compiled regular expression for the regular expression that wasn't in the cache.

In addition to the compiled regular expression cache, RegexKitLite keeps a small lookaside cache that maps a regular expressions NSString pointer and RKLRegexOptions directly to a cached compiled regular expression. When a regular expressions NSString pointer and RKLRegexOptions is in the lookaside cache, RegexKitLite can bypass calling CFHash(regexString) and checking each of the four "ways" in a cache set since the lookaside cache has provided the exact cached compiled regular expression. The lookaside cache is quite small at just 64 bytes and it was added because Shark.app profiling during performance tuning showed that CFHash(), while quite fast, was the primary bottleneck when retrieving already compiled and cached regular expressions, typically accounting for ≅40% of the look up time.

Regular Expressions in Mutable Strings

When a regular expression is compiled, an immutable copy of the string is kept. For immutable NSString objects, the copy is usually the same object with its reference count increased by one. Only NSMutableString objects will cause a new, immutable NSString to be created.

If the regular expression being used is stored in a NSMutableString, the cached regular expression will continue to be used as long as the NSMutableString remains unchanged. Once mutated, the changed NSMutableString will no longer be a match for the cached compiled regular expression that was being used by it previously. Even if the newly mutated strings hash is congruent to the previous unmutated strings hash modulo RKL_CACHE_SIZE, that is to say they share the same cache set (i.e., ([mutatedString hash] % RKL_CACHE_SIZE) == ([unmutatedString hash] % RKL_CACHE_SIZE)), the immutable copy of the regular expression string used to create the compiled regular expression is used to ensure true equality. The newly mutated string will have to go through the whole regular expression compile and cache creation process.

This means that NSMutableString objects can be safely used as regular expressions, and any mutations to those objects will immediately be detected and reflected in the regular expression used for matching.

Searching Mutable Strings

Unfortunately, the ICU regular expression API requires that the compiled regular expression be "set" to the string to be searched. To search a different string, the compiled regular expression must be "set" to the new string. Therefore, RegexKitLite tracks the last NSString that each compiled regular expression was set to, recording the pointer to the NSString object, its hash, and its length. If any of these parameters are different from the last parameters used for a compiled regular expression, the compiled regular expression is "set" to the new string. Since mutating a string will likely change its hash value, it's generally safe to search NSMutableString objects, and in most cases the mutation will reset the compiled regular expression to the updated contents of the NSMutableString.

Caution:

Care must be taken when mutable strings are searched and there exists the possibility that the string has mutated between searches. See NSString RegexKitLite Additions Reference - Cached Information and Mutable Strings for more information.

Last Match Information

When performing a match, the arguments used to perform the match are kept. If those same arguments are used again, the actual matching operation is skipped because the compiled regular expression already contains the results for the given arguments. This is mostly useful when a regular expression contains multiple capture groups, and the results for different capture groups for the same match are needed. This means that there is only a small penalty for iterating over all the capture groups in a regular expression for a match, and essentially becomes the direct ICU regular expression API equivalent of uregex_start() and uregex_end().

See Also

UTF-16 Conversion Cache

RegexKitLite is ideal when the string being matched is a non-ASCII, Unicode string. This is because the regular expression engine used, ICU, can only operate on UTF-16 encoded strings. Since Cocoa keeps essentially all non-ASCII strings encoded in UTF-16 form internally, this means that RegexKitLite can operate directly on the strings buffer without having to make a temporary copy and transcode the string in to ICU's required format.

Like all object oriented programming, the internal representation of an objects information is private. However, the ICU regular expression engine requires that the text to be search be encoded as a UTF-16 string. For pragmatic purposes, Core Foundation has several public functions that can provide direct access to the buffer used to hold the contents of the string, but such direct access is only available if the private buffer is already encoded in the requested direct access format. As a rough rule of thumb, 8-bit simple strings, such as ASCII, are kept in their 8-bit format. Non 8-bit simple strings are stored as UTF-16 strings. Of course, this is an implementation private detail, so this behavior should never be relied upon. It is mentioned because of the tremendous impact on matching performance and efficiency it can have if a string must be converted to UTF-16.

For strings in which direct access to the UTF-16 string is available, RegexKitLite uses that buffer. This is the ideal case as no extra work needs to be performed, such as converting the string in to a UTF-16 string, and allocating memory to hold the temporary conversion. Of course, direct access is not always available, and occasionally the string to be searched will need to be converted in to a UTF-16 string.

RegexKitLite has two conversion cache types. Each conversion cache type contains four buffers each, and buffers are re-used on a least recently used basis. If the selected cache type does not contain the contents of the NSString that is currently being searched in any of its buffers, the least recently used buffer is cleared and the current NSString takes it place. The first conversion cache type is fixed in size and set by the C pre-processor define RKL_FIXED_LENGTH, which defaults to 2048. Any string whose length is less than RKL_FIXED_LENGTH will use the fixed size conversion cache type. The second conversion cache type, for strings whose length is longer than RKL_FIXED_LENGTH, will use a dynamically sized conversion buffer. The memory allocation for the dynamically sized conversion buffer is resized for each conversion with realloc() to the size needed to hold the entire contents of the UTF-16 converted string.

This strategy was chosen for its relative simplicity. Keeping track of dynamically created resources is required to prevent memory leaks. As designed, there are only four pointers to dynamically allocated memory: the four pointers to hold the conversion contents of strings whose length is larger than RKL_FIXED_LENGTH. However, since realloc() is used to manage those memory allocations, it becomes very difficult to accidentally leak the buffers. Having the fixed sized buffers means that the memory allocation system isn't bothered with many small requests, most of which are transient in nature to begin with. The current strategy tries to strike the best balance between performance and simplicity.

Mutable Strings

When converted in to a UTF-16 string, the hash of the NSString is recorded, along with the pointer to the NSString object and the strings length. In order for the RegexKitLite to use the cached conversion, all of these parameters must be equal to their values of the NSString to be searched. If there is any difference, the cached conversion is discarded and the current NSString, or NSMutableString as the case may be, is reconverted in to a UTF-16 string.

Caution:

Care must be taken when mutable strings are searched and there exists the possibility that the string has mutated between searches. See NSString RegexKitLite Additions Reference - Cached Information and Mutable Strings for more information.

Multithreading Safety

RegexKitLite is also multithreading safe. Access to the compiled regular expression cache and the conversion cache is protected by a single OSSpinLock to ensure that only one thread has access at a time. The lock remains held while the regular expression match is performed since the compiled regular expression returned by the ICU library is not safe to use from multiple threads. Once the match has completed, the lock is released, and another thread is free to lock the cache and perform a match.

Important:

While it is safe to use the same regular expression from any thread at any time, the usual multithreading caveats apply. For example, it is not safe to mutate a NSMutableString in one thread while performing a match on the mutating string in another.

If Blocks functionality is enabled, and a RegexKitLite method that takes a Block as one of its parameters is used, RegexKitLite takes a slightly different approach in order to support the asynchronous, and possibly re-entrant, nature of Blocks.

First, an autoreleased Block helper proxy object is created and is used to keep track of any Block local resources needed to perform a Block-based enumeration.

Then the regular expression cache is checked exactly as before. Once a compiled regular expression is obtained, the ICU function uregex_clone is used to create a Block local copy of the regular expression. After the Block local copy has been made, the global compiled regular expression cache lock is unlocked.

If the string to be searched requires conversion to UTF-16, then a one time use Block local UTF-16 conversion of the string is created.

These changes mean that RegexKitLite Block-based enumeration methods are just as multithreading safe and easy to use as non-Block-based enumeration methods, such as the ability to continue to use RegexKitLite methods without any restrictions from within the Block used for enumeration.

64-bit Support

RegexKitLite is 64-bit clean. Internally, RegexKitLite uses Cocoas standard NSInteger and NSUInteger types for representing integer values. The size of these types change between 32-bit and 64-bit automatically, depending on the target architecture. ICU, on the other hand, uses a signed 32-bit int type for many of its arguments, such as string offset values. Because of this, the maximum length of a string that RegexKitLite will accept is the maximum value that can be represented by a signed 32-bit integer, which is approximately 2 gigabytes. Strings that are longer this limit will raise NSRangeException. This limitation may be significant to those who are switching to 64-bit because the size of the data they need to process exceeds what can be represented with 32-bits.

Note:

Several numeric constants throughout this document will have either L or UL appended to them— for example 0UL, or 2L. This is to ensure that they are treated as 64-bit long or unsigned long values, respectively, when targeting a 64-bit architecture.

Using RegexKitLite

The goal of RegexKitLite is not to be a comprehensive Objective-C regular expression framework, but to provide a set of easy to use primitives from which additional functionality can be created. To this end, RegexKitLite provides the following two core primitives from which everything else is built:

There is often a need to create a new string of the characters that were matched by a regular expression. RegexKitLite provides the following method which conveniently combines sending the receiver substringWithRange: with the range returned by rangeOfRegex:.

RegexKitLite 2.0 adds the ability to split strings by dividing them with a regular expression, and the ability to perform search and replace operations using common $n substitution syntax. replaceOccurrencesOfRegex:withString: is used to modify the contents of NSMutableString objects directly and stringByReplacingOccurrencesOfRegex:withString: will create a new, immutable NSString from the receiver.

RegexKitLite 3.0 adds several new methods that return a NSArray containing the aggregated results of a number of individual regex operations.

RegexKitLite 4.0 adds several new methods that take advantage of the new blocks language extension.

There are no additional classes that supply the regular expression matching functionality, everything is accomplished with the two methods above. These methods are added to the existing NSString class via an Objective-C category extension. See RegexKitLite NSString Additions Reference for a complete list of methods.

The real workhorse is the rangeOfRegex:options:inRange:capture:error: method. The receiver of the message is an ordinary NSString class member that you wish to perform a regular expression match on. The parameters of the method are a NSString containing the regular expression regex, any RKLRegexOptions match options, the NSRange range of the receiver that is to be searched, the capture number from the regular expression regex that you would like the result for, and an optional error parameter that will contain a NSError object if a problem occurs with the details of the error.

Important:

The C language assigns special meaning to the \ character when inside a quoted " " string in your source code. The \ character is the escape character, and the character that follows has a different meaning than normal. The most common example of this is \n, which translates in to the new-line character. Because of this, you are required to 'escape' any uses of \ by prepending it with another \. In practical terms this means doubling any \ in a regular expression, which unfortunately is quite common, that are inside of quoted " " strings in your source code. Failure to do so will result in numerous warnings from the compiler about unknown escape sequences. To match a single literal \ with a regular expression requires no less than four backslashes: "\\\\".

See Also

Finding the Range of a Match

A simple example:

NSString *searchString = @"This is neat."; NSString *regexString = @"(\\w+)\\s+(\\w+)\\s+(\\w+)"; NSRange matchedRange = NSMakeRange(NSNotFound, 0UL); NSError *error = NULL; matchedRange = [searchString rangeOfRegex:regexString options:RKLNoOptions inRange:searchRange capture:2L error:&error]; NSLog(@"matchedRange: %@", NSStringFromRange(matchedRange)); // 2008-03-18 03:51:16.530 test[51583:813] matchedRange: {5, 2}
Continues…
 

In the previous example, the NSRange that capture number 2 matched is {5, 2}, which corresponds to the word is in searchString. Once the NSRange is known, you can create a new string containing just the matching text:

…example
NSString *matchedString = [searchString substringWithRange:matchedRange]; NSLog(@"matchedString: '%@'", matchedString); // 2008-03-18 03:51:16.532 test[51583:813] matchedString: 'is'

RegexKitLite can conveniently combine the two steps above with stringByMatching:. This example also demonstrates the use of one of the simpler convenience methods, where some of the arguments are automatically filled in with default values:

NSString *searchString = @"This is neat."; NSString *regexString = @"(\\w+)\\s+(\\w+)\\s+(\\w+)"; NSString *matchedString = [searchString stringByMatching:regexString capture:2L]; NSLog(@"matchedString: '%@'", matchedString); // 2008-03-18 03:53:42.949 test[51583:813] matchedString: 'is'
See Also

Search and Replace

You can perform search and replace operations on NSString objects and use common $n capture group substitution in the replacement string:

NSString *searchString = @"This is neat."; NSString *regexString = @"\\b(\\w+)\\b"; NSString *replaceWithString = @"{$1}"; NSString *replacedString = NULL; replacedString = [searchString stringByReplacingOccurrencesOfRegex:regexString withString:replaceWithString]; NSLog(@"replaced string: '%@'", replacedString); // 2008-07-01 19:03:03.195 test[68775:813] replaced string: '{This} {is} {neat}.'
Important:
Search and replace methods will raise a RKLICURegexException if the replacementString contains $n capture references where n is greater than the number of capture groups in the regular expression.

In this example, the regular expression \b(\w+)\b has a single capture group, which is created with the use of () parenthesis. The text that was matched inside the parenthesis is available for use in the replacement text by using $n, where n is the parenthesized capture group you would like to use. Additional capture groups are numbered sequentially in the order that they appear from left to right. Capture group 0 (zero) is also available and is equivalent to all the text that the regular expression matched.

Mutable strings can be manipulated directly:

NSMutableString *mutableString = [NSMutableString stringWithString:@"This is neat."]; NSString *regexString = @"\\b(\\w+)\\b"; NSString *replaceWithString = @"{$1}"; NSUInteger replacedCount = 0UL; replacedCount = [mutableString replaceOccurrencesOfRegex:regexString withString:replaceWithString]; NSLog(@"count: %lu string: '%@'", (u_long)replacedCount, mutableString); // 2008-07-01 21:25:43.433 test[69689:813] count: 3 string: '{This} {is} {neat}.'

Search and Replace using Blocks

RegexKitLite 4.0 adds support for performing the same search and replacement on strings, except now the contents of the replacement string are created by the Block that is passed as the argument. For each match that is found in the string, the Block argument is called and passed the details of the match which includes a C array of NSString objects, one for each capture, along with a C array of NSRange structures with the range information for the current match. The text that was matched will be replaced with the NSString object that the Block is required to return. % This allows you complete control over the contents of the replaced text, such as doing complex transformations of the matched text, which is much more flexible and powerful than the simple, fixed replacement functionality provided by stringByReplacingOccurrencesOfRegex:withString:. The example below is essentially the same as the previous search and replace examples, except this example uses the capitalizedString method to capitalize the matched result, which is then used in the string that is returned as the replacement text. Note that the first letter in each word in replacedString is now capitalized.
NSString *searchString = @"This is neat."; NSString *regexString = @"\\b(\\w+)\\b"; NSString *replacedString = NULL; replacedString = [searchString stringByReplacingOccurrencesOfRegex:regexString usingBlock: ^NSString *(NSInteger captureCount, NSString * const capturedStrings, const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { return([NSString stringWithFormat:@"{%@}", [capturedStrings[1] capitalizedString]]); }]; // 2010-04-14 21:00:42.726 test[35053:a0f] replaced string: '{This} {Is} {Neat}.'
See Also

Splitting Strings

Strings can be split with a regular expression using the componentsSeparatedByRegex: methods. This functionality is nearly identical to the preexisting NSString method componentsSeparatedByString:, except instead of only being able to use a fixed string as a separator, you can use a regular expression:

NSString *searchString = @"This is neat."; NSString *regexString = @"\\s+"; NSArray *splitArray = NULL; splitArray = [searchString componentsSeparatedByRegex:regexString]; // splitArray == { @"This", @"is", @"neat." } NSLog(@"splitArray: %@", splitArray);
Continues…
 

The output from NSLog() when run from a shell:

splitArray
shell% ./splitArray 2008-07-01 20:58:39.025 splitArray[69618:813] splitArray: ( This, is, "neat." ) shell%

Unfortunately our example string @"This is neat." doesn't allow us to show off the power of regular expressions. As you can probably imagine, splitting the string with the regular expression \s+ allows for one or more white space characters to be matched. This can be much more flexible than just a fixed string of @" ", which will split on a single space only. If our example string contained extra spaces, say @"This   is     neat.", the result would have been the same.

See Also

Creating an Array of Every Match

RegexKitLite 3.0 adds several methods that conveniently perform a number of individual RegexKitLite operations and aggregate the results in to a NSArray. Since the result is a NSArray, the standard Cocoa collection enumeration patterns can be used, such as NSEnumerator and Objective-C 2.0's for…in feature. One of the most common tasks is to extract all of the matches of a regular expression from a string. componentsMatchedByRegex: returns the entire text matched by a regular expression even if the regular expression contains additional capture groups, effectively capture group 0. For example:

NSString *searchString = @"$10.23, $1024.42, $3099"; NSString *regexString = @"\\$((\\d+)(?:\\.(\\d+)|\\.?))"; NSArray *matchArray = NULL; matchArray = [searchString componentsMatchedByRegex:regexString]; // matchArray == { @"$10.23", @"$1024.42", @"$3099" }; NSLog(@"matchArray: %@", matchArray);
Continues…
 

The output from NSLog() when run from a shell:

matchArray
shell% ./matchArray 2009-05-06 03:20:03.546 matchArray[69939:10b] matchArray: ( "$10.23", "$1024.42", "$3099" ) shell%

As the example above demonstrates, componentsMatchedByRegex: returns the entire text that the regular expression matched even though the regular expression contains capture groups. arrayOfCaptureComponentsMatchedByRegex: can be used if you need to get the text that the individual capture groups matched as well:

NSString *searchString = @"$10.23, $1024.42, $3099"; NSString *regexString = @"\\$((\\d+)(?:\\.(\\d+)|\\.?))"; NSArray *capturesArray = NULL; capturesArray = [searchString arrayOfCaptureComponentsMatchedByRegex:regexString]; /* capturesArray == [NSArray arrayWithObjects: [NSArray arrayWithObjects: @"$10.23", @"10.23", @"10", @"23", NULL], [NSArray arrayWithObjects:@"$1024.42", @"1024.42", @"1024", @"42", NULL], [NSArray arrayWithObjects: @"$3099", @"3099", @"3099", @"", NULL], NULL]; */ NSLog(@"capturesArray: %@", capturesArray);
Continues…
 

The output from NSLog() when run from a shell:

capturesArray
shell% ./capturesArray 2009-05-06 03:25:46.852 capturesArray[69981:10b] capturesArray: ( ( "$10.23", "10.23", 10, 23 ), ( "$1024.42", "1024.42", 1024, 42 ), ( "$3099", 3099, 3099, "" ) ) shell%
See Also

Enumerating Matches

The RegexKitLite componentsMatchedByRegex: method enables you to quickly create a NSArray containing all the matches of a regular expression in a string. To enumerate the contents of the NSArray, you can send the array an objectEnumerator message.

See Also

An example using componentsMatchedByRegex: and a NSEnumerator:

File name:main.m
#import <Foundation/Foundation.h> #import "RegexKitLite.h" int main(int argc, char *argv[]) { NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; NSString *searchString = @"one\ntwo\n\nfour\n"; NSArray *matchArray = NULL; NSEnumerator *matchEnumerator = NULL; NSString *regexString = @"(?m)^.*$"; NSLog(@"searchString: '%@'", searchString); NSLog(@"regexString : '%@'", regexString); matchArray = [searchString componentsMatchedByRegex:regexString]; matchEnumerator = [matchArray objectEnumerator]; NSUInteger line = 0UL; NSString *matchedString = NULL; while((matchedString = [matchEnumerator nextObject]) != NULL) { NSLog(@"%lu: %lu '%@'", (u_long)++line, (u_long)[matchedString length], matchedString); } [pool release]; return(0); }

The following shell transcript demonstrates compiling the example and executing it. Line number three clearly demonstrates that matches of zero length are possible. Without the additional logic in nextObject to handle this special case, the enumerator would never advance past the match.

Note:

In the shell transcript below, the NSLog() line that prints searchString has been annotated with the '' character to help visually identify the corresponding \n new-line characters in searchString.

shell% cd examples shell% gcc -I.. -g -o main main.m ../RegexKitLite.m -framework Foundation -licucore shell% ./main 2008-03-21 15:56:17.469 main[44050:807] searchString: 'one two four ' 2008-03-21 15:56:17.520 main[44050:807] regexString : '(?m)^.*$' 2008-03-21 15:56:17.575 main[44050:807] 1: 3 'one' 2008-03-21 15:56:17.580 main[44050:807] 2: 3 'two' 2008-03-21 15:56:17.584 main[44050:807] 3: 0 '' 2008-03-21 15:56:17.590 main[44050:807] 4: 4 'four' shell%

Enumerating Matches with Objective-C 2.0

You can enumerate all the matches of a regular expression in a string using Objective-C 2.0's for…in feature. Compared to using a NSEnumerator, using for…in not only takes fewer lines of code to accomplish the same thing, it is usually faster as well.

See Also

An example using the Objective-C 2.0 for…in feature:

File name:for_in.m
#import <Foundation/Foundation.h> #import "RegexKitLite.h" int main(int argc, char *argv[]) { NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; NSString *searchString = @"one\ntwo\n\nfour\n"; NSString *regexString = @"(?m)^.*$"; NSUInteger line = 0UL; NSLog(@"searchString: '%@'", searchString); NSLog(@"regexString : '%@'", regexString); for(NSString *matchedString in [searchString componentsMatchedByRegex:regexString]) { NSLog(@"%lu: %lu '%@'", (u_long)++line, (u_long)[matchedString length], matchedString); } [pool release]; return(0); }
Note:

The output of the preceding example is identical to the NSEnumerator shell output.

Enumerating Matches using Blocks

A third way to enumerate all the matches of a regular expression in a string is to use one of the Blocks-based enumeration methods.
See Also
An example using enumerateStringsMatchedByRegex:usingBlock::
#import <Foundation/Foundation.h> #import "RegexKitLite.h" int main(int argc, char *argv) { NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; NSString *searchString = @"one\ntwo\n\nfour\n"; NSString *regexString = @"(?m)^.*$"; __block NSUInteger line = 0UL; NSLog(@"searchString: '%@'", searchString); NSLog(@"regexString : '%@'", regexString); [searchString enumerateStringsMatchedByRegex:regexString usingBlock: ^(NSInteger captureCount, NSString * const capturedStrings[captureCount], const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { NSLog(@"%lu: %lu '%@'", ++line, [capturedStrings[0] length], capturedStrings[0]); }]; [pool release]; return(0); }
Note:

The output of the preceding example is identical to the NSEnumerator shell output.

DTrace

Important:

DTrace support is not enabled by default. To enable DTrace support, use the RKL_DTRACE pre-processor flag: -DRKL_DTRACE

RegexKitLite has two DTrace probe points that provide information about its internal caches:

Each of the probe points supply information via a number of arguments that are accessible through the DTrace variables arg0argn.

The first argument, eventID via arg0, is a unique event ID that is incremented each time the RegexKitLite mutex lock is acquired. All the probes that fire while the mutex is held will share the same event ID. This can help if you are trying to correlate multiple events across different CPUs.

Important:

Most uses of the dtrace command require superuser privileges. The examples given here use sudo to execute dtrace as the root user.

The following is available in examples/compiledRegexCache.d and demonstrates the use of all the arguments available via the RegexKitLite:::compiledRegexCache probe point:

File name:compiledRegexCache.d
#!/usr/sbin/dtrace -s RegexKitLite*:::compiledRegexCache { this->eventID = (unsigned long)arg0; this->regexUTF8 = copyinstr(arg1); this->options = (unsigned int)arg2; this->captures = (int)arg3; this->hitMiss = (int)arg4; this->icuStatusCode = (int)arg5; this->icuErrorMessage = (arg6 == 0) ? "" : copyinstr(arg6); this->hitRate = (double *)copyin(arg7, sizeof(double)); printf("%5d: %-60.60s Opt: %#8.8x Cap: %2d Hit: %2d Rate: %6.2f%% code: %5d msg: %s\n", this->eventID, this->regexUTF8, this->options, this->captures, this->hitMiss, *this->hitRate, this->icuStatusCode, this->icuErrorMessage); }

Below is an example of the output, which has been trimmed for brevity, from compiledRegexCache.d:

compiledRegexCache.d
shell% sudo dtrace -Z -q -s compiledRegexCache.d 110: (\[{2})(.+?)(]{2}) Opt: 0x00000000 Cap: 3 Hit: 0 Rate: 63.64% code: 0 msg: 111: (\[{2})(.+?)(]{2}) Opt: 0x00000000 Cap: 3 Hit: 1 Rate: 63.96% code: 0 msg: 131: (\w+ Opt: 0x00000000 Cap: -1 Hit: -1 Rate: 63.36% code: 66310 msg: U_REGEX_MISMATCHED_PAREN 164: \b\s* Opt: 0x00000000 Cap: 0 Hit: 0 Rate: 60.98% code: 0 msg: 165: \$((\d+)(?:\.(\d+)|\.?)) Opt: 0x00000000 Cap: 3 Hit: 1 Rate: 61.21% code: 0 msg: 166: \b(https?)://([a-zA-Z0-9\-.]+)((?:/[a-zA-Z0-9\-._?,'+\&%$… Opt: 0x00000000 Cap: 3 Hit: 0 Rate: 60.84% code: 0 msg: shell%

An example that prints the number of times that a compiled regular expression was not in the cache per second:

shell% sudo dtrace -Z -q -n 'RegexKitLite*:::compiledRegexCache /arg4 == 0/ { @miss[pid, execname] = count(); }' -n 'tick-1sec { printa("%-8d %-40s %@d/sec\n", @miss); trunc(@miss); }' 67003 RegexKitLite_tests 16/sec 67008 RegexKitLite_tests 50/sec ^C shell%
See Also

The following is available in examples/utf16ConversionCache.d and demonstrates the use of all the arguments available via the RegexKitLite:::utf16ConversionCache probe point.

File name:utf16ConversionCache.d
#!/usr/sbin/dtrace -s enum { RKLCacheHitLookupFlag = 1 << 0, RKLConversionRequiredLookupFlag = 1 << 1, RKLSetTextLookupFlag = 1 << 2, RKLDynamicBufferLookupFlag = 1 << 3, RKLErrorLookupFlag = 1 << 4 }; RegexKitLite*:::utf16ConversionCache { this->eventID = (unsigned long)arg0; this->lookupResultFlags = (unsigned int)arg1; this->hitRate = (double *)copyin(arg2, sizeof(double)); this->stringPtr = (void *)arg3; this->NSRange_location = (unsigned long)arg4; this->NSRange_length = (unsigned long)arg5; this->length = (long)arg6; printf("%5lu: flags: %#8.8x {Hit: %d Conv: %d SetText: %d Dyn: %d Error: %d} rate: %6.2f%% string: %#8.8p NSRange {%6lu, %6lu} length: %ld\n", this->eventID, this->lookupResultFlags, (this->lookupResultFlags & RKLCacheHitLookupFlag) != 0, (this->lookupResultFlags & RKLConversionRequiredLookupFlag) != 0, (this->lookupResultFlags & RKLSetTextLookupFlag) != 0, (this->lookupResultFlags & RKLDynamicBufferLookupFlag) != 0, (this->lookupResultFlags & RKLErrorLookupFlag) != 0, *this->hitRate, this->stringPtr, this->NSRange_location, this->NSRange_length, this->length); }

Below is an example of the output, which has been trimmed for brevity, from utf16ConversionCache.d:

utf16ConversionCache.d
shell% sudo dtrace -Z -q -s utf16ConversionCache.d 85: flags: 0x00000000 {Hit: 0 Conv: 0 SetText: 0 Dyn: 0 Error: 0} rate: 59.18% string: 0x0010f530 NSRange { 0, 18} length: 18 86: flags: 0x00000004 {Hit: 0 Conv: 0 SetText: 1 Dyn: 0 Error: 0} rate: 59.18% string: 0x0010f530 NSRange { 0, 18} length: 18 87: flags: 0x00000006 {Hit: 0 Conv: 1 SetText: 1 Dyn: 0 Error: 0} rate: 58.00% string: 0x00054930 NSRange { 1, 37} length: 39 88: flags: 0x00000003 {Hit: 1 Conv: 1 SetText: 0 Dyn: 0 Error: 0} rate: 58.82% string: 0x00054930 NSRange { 1, 37} length: 39 109: flags: 0x00000006 {Hit: 0 Conv: 1 SetText: 1 Dyn: 0 Error: 0} rate: 53.62% string: 0x00054d00 NSRange { 0, 56} length: 56 110: flags: 0x00000006 {Hit: 0 Conv: 1 SetText: 1 Dyn: 0 Error: 0} rate: 52.86% string: 0x00054680 NSRange { 0, 1064} length: 1064 111: flags: 0x00000007 {Hit: 1 Conv: 1 SetText: 1 Dyn: 0 Error: 0} rate: 53.52% string: 0x00054680 NSRange { 46, 978} length: 1064 shell%

An example that prints the number of times that a string required a conversion to UTF-16 and was not in the cache per second:

shell% sudo dtrace -Z -q -n 'RegexKitLite*:::utf16ConversionCache /(arg1 & 0x3) == 0x2/ { @miss[pid, execname] = count(); }' -n 'tick-1sec { printa("%-8d %-40s %@d/sec\n", @miss); trunc(@miss); }' 67020 RegexKitLite_tests 73/sec 67037 RegexKitLite_tests 64/sec ^C shell%
See Also

ICU Syntax

In this section:

ICU Regular Expression Syntax

For your convenience, the regular expression syntax from the ICU documentation is included below. When in doubt, you should refer to the official ICU User Guide - Regular Expressions documentation page.

See Also
Metacharacters
CharacterDescription
\aMatch a BELL, \u0007
\AMatch at the beginning of the input. Differs from ^ in that \A will not match after a new-line within the input.
\b, outside of a [Set] Match if the current position is a word boundary. Boundaries occur at the transitions between word \w and non-word \W characters, with combining marks ignored.
\b, within a [Set]Match a BACKSPACE, \u0008.
\BMatch if the current position is not a word boundary.
\cxMatch a Control-x character.
\dMatch any character with the Unicode General Category of Nd (Number, Decimal Digit).
\DMatch any character that is not a decimal digit.
\eMatch an ESCAPE, \u001B.
\ETerminates a \Q\E quoted sequence.
\fMatch a FORM FEED, \u000C.
\GMatch if the current position is at the end of the previous match.
\nMatch a LINE FEED, \u000A.
\N{Unicode Character Name}Match the named Unicode Character.
\p{Unicode Property Name}Match any character with the specified Unicode Property.
\P{Unicode Property Name}Match any character not having the specified Unicode Property.
\QQuotes all following characters until \E.
\rMatch a CARRIAGE RETURN, \u000D.
\sMatch a white space character. White space is defined as [\t\n\f\r\p{Z}].
\SMatch a non-white space character.
\tMatch a HORIZONTAL TABULATION, \u0009.
\uhhhhMatch the character with the hex value hhhh.
\UhhhhhhhhMatch the character with the hex value hhhhhhhh. Exactly eight hex digits must be provided, even though the largest Unicode code point is \U0010ffff.
\wMatch a word character. Word characters are [\p{Ll}\p{Lu}\p{Lt}\p{Lo}\p{Nd}].
\WMatch a non-word character.
\x{h}Match the character with hex value hhhh. From one to six hex digits may be supplied.
\xhhMatch the character with two digit hex value hh.
\XMatch a Grapheme Cluster.
\ZMatch if the current position is at the end of input, but before the final line terminator, if one exists.
\zMatch if the current position is at the end of input.
\n
Back Reference. Match whatever the nth capturing group matched. n must be a number ≥ 1 and ≤ total number of capture groups in the pattern.
Note:
Octal escapes, such as \012, are not supported.
[pattern]Match any one character from the set. See ICU Regular Expression Character Classes for a full description of what may appear in the pattern.
.Match any character.
^Match at the beginning of a line.
$Match at the end of a line.
\Quotes the following character. Characters that must be quoted to be treated as literals are * ? + [ ( ) { } ^ $ | \ . /
Operators
OperatorDescription
|Alternation. A|B matches either A or B.
*Match zero or more times. Match as many times as possible.
+Match one or more times. Match as many times as possible.
?Match zero or one times. Prefer one.
{n}Match exactly n times.
{n,}Match at least n times. Match as many times as possible.
{n,m}Match between n and m times. Match as many times as possible, but not more than m.
*?Match zero or more times. Match as few times as possible.
+?Match one or more times. Match as few times as possible.
??Match zero or one times. Prefer zero.
{n}?Match exactly n times.
{n,}?Match at least n times, but no more than required for an overall pattern match.
{n,m}?Match between n and m times. Match as few times as possible, but not less than n.
*+Match zero or more times. Match as many times as possible when first encountered, do not retry with fewer even if overall match fails. Possessive match.
++Match one or more times. Possessive match.
?+Match zero or one times. Possessive match.
{n}+Match exactly n times. Possessive match.
{n,}+Match at least n times. Possessive match.
{n,m}+Match between n and m times. Possessive match.
()Capturing parentheses. Range of input that matched the parenthesized subexpression is available after the match.
(?:)Non-capturing parentheses. Groups the included pattern, but does not provide capturing of matching text. Somewhat more efficient than capturing parentheses.
(?>)Atomic-match parentheses. First match of the parenthesized subexpression is the only one tried; if it does not lead to an overall pattern match, back up the search for a match to a position before the (?> .
(?#)Free-format comment (?#comment).
(?=)Look-ahead assertion. True if the parenthesized pattern matches at the current input position, but does not advance the input position.
(?!)Negative look-ahead assertion. True if the parenthesized pattern does not match at the current input position. Does not advance the input position.
(?<=)Look-behind assertion. True if the parenthesized pattern matches text preceding the current input position, with the last character of the match being the input character just before the current position. Does not alter the input position. The length of possible strings matched by the look-behind pattern must not be unbounded (no * or + operators).
(?<!)Negative Look-behind assertion. True if the parenthesized pattern does not match text preceding the current input position, with the last character of the match being the input character just before the current position. Does not alter the input position. The length of possible strings matched by the look-behind pattern must not be unbounded (no * or + operators).
(?ismwx-ismwx:)Flag settings. Evaluate the parenthesized expression with the specified flags enabled or -disabled.
(?ismwx-ismwx)Flag settings. Change the flag settings. Changes apply to the portion of the pattern following the setting. For example, (?i) changes to a case insensitive match.
See Also

ICU Regular Expression Character Classes

The following was originally from ICU User Guide - UnicodeSet, but has been adapted to fit the needs of this documentation. Specifically, the ICU UnicodeSet documentation describes an ICU C++ object— UnicodeSet. The term UnicodeSet was effectively replaced with Character Class, which is more appropriate in the context of regular expressions. As always, you should refer to the original, official documentation when in doubt.

See Also

Overview

A character class is a regular expression pattern that represents a set of Unicode characters or character strings. The following table contains some example character class patterns:

PatternDescription
[a-z]The lower case letters a through z
[abc123]The six characters a, b, c, 1, 2, and 3
[\p{Letter}]All characters with the Unicode General Category of Letter.
String Values

In addition to being a set of Unicode code point characters, a character class may also contain string values. Conceptually, a character class is always a set of strings, not a set of characters. Historically, regular expressions have treated [] character classes as being composed of single characters only, which is equivalent to a string that contains only a single character.

Character Class Patterns

Patterns are a series of characters bounded by square brackets that contain lists of characters and Unicode property sets. Lists are a sequence of characters that may have ranges indicated by a - between two characters, as in a-z. The sequence specifies the range of all characters from the left to the right, in Unicode order. For example, [a c d-f m] is equivalent to [a c d e f m]. Whitespace can be freely used for clarity as [a c d-f m] means the same as [acd-fm].

Unicode property sets are specified by a Unicode property, such as [:Letter:]. ICU version 2.0 supports General Category, Script, and Numeric Value properties (ICU will support additional properties in the future). For a list of the property names, see the end of this section. The syntax for specifying the property names is an extension of either POSIX or Perl syntax with the addition of =value. For example, you can match letters by using the POSIX syntax [:Letter:], or by using the Perl syntax \p{Letter}. The type can be omitted for the Category and Script properties, but is required for other properties.

The following table lists the standard and negated forms for specifying Unicode properties in both POSIX or Perl syntax. The negated form specifies a character class that includes everything but the specified property. For example, [:^Letter:] matches all characters that are not [:Letter:].

Syntax StyleStandardNegated
POSIX[:type=value:][:^type=value:]
Perl\p{type=value}\P{type=value}
See Also

Character classes can then be modified using standard set operations— Union, Inverse, Difference, and Intersection.

The binary operators & and - have equal precedence and bind left-to-right. Thus [[:letter:]-[a-z]-[\u0100-\u01FF]] is equivalent to [[[:letter:]-[a-z]]-[\u0100-\u01FF]]. Another example is the set [[ace][bdf] - [abc][def]] is not the empty set, but instead the set [def]. This only really matters for the difference operation, as the intersection operation is commutative.

Another caveat with the & and - operators is that they operate between sets. That is, they must be immediately preceded and immediately followed by a set. For example, the pattern [[:Lu:]-A] is illegal, since it is interpreted as the set [:Lu:] followed by the incomplete range -A. To specify the set of uppercase letters except for A, enclose the A in a set: [[:Lu:]-[A]].

PatternDescription
[a]The set containing a.
[a-z]The set containing a through z and all letters in between, in Unicode order.
[^a-z]The set containing all characters but a through z, that is, U+0000 through a-1 and z+1 through U+FFFF.
[[pat1][pat2]]The union of sets specified by pat1 and pat2.
[[pat1]&[pat2]]The intersection of sets specified by pat1 and pat2.
[[pat1]-[pat2]]The asymmetric difference of sets specified by pat1 and pat2.
[:Lu:]The set of characters belonging to the given Unicode category. In this case, Unicode uppercase letters. The long form for this is [:UppercaseLetter:].
[:L:]The set of characters belonging to all Unicode categories starting with L, that is, [[:Lu:][:Ll:][:Lt:][:Lm:][:Lo:]]. The long form for this is [:Letter:].
See Also
String Values in Character Classes

String values are enclosed in {curly brackets}. For example:

PatternDescription
[abc{def}]A set containing four members, the single characters a, b, and c and the string def
[{abc}{def}]A set containing two members, the string abc and the string def.
[{a}{b}{c}][abc]These two sets are equivalent. Each contains three items, the three individual characters a, b, and c. A {string} containing a single character is equivalent to that same character specified in any other way.

Character Quoting and Escaping in ICU Character Class Patterns

Single Quote

Two single quotes represent a single quote, either inside or outside single quotes. Text within single quotes is not interpreted in any way, except for two adjacent single quotes. It is taken as literal text— special characters become non-special. These quoting conventions for ICU character classes differ from those of Perl or Java. In those environments, single quotes have no special meaning, and are treated like any other literal character.

Backslash Escapes

Outside of single quotes, certain backslashed characters have special meaning:

PatternDescription
\uhhhhExactly 4 hex digits; h in [0-9A-Fa-f]
\UhhhhhhhhExactly 8 hex digits
\xhh1-2 hex digits
\ooo1-3 octal digits; o in [0-7]
\aU+0007 BELL
\bU+0008 BACKSPACE
\tU+0009 HORIZONTAL TAB
\nU+000A LINE FEED
\vU+000B VERTICAL TAB
\fU+000C FORM FEED
\rU+000D CARRIAGE RETURN
\\U+005C BACKSLASH

Anything else following a backslash is mapped to itself, except in an environment where it is defined to have some special meaning. For example, \p{Lu} is the set of uppercase letters. Any character formed as the result of a backslash escape loses any special meaning and is treated as a literal. In particular, note that \u and \U escapes create literal characters.

Whitespace

Whitespace, as defined by the ICU API, is ignored unless it is quoted or backslashed.

Property Values

The following property value styles are recognized:

StyleDescription
ShortOmits the =type argument. Used to prevent ambiguity and only allowed with the Category and Script properties.
MediumUses an abbreviated type and value.
LongUses a full type and value.

If the type or value is omitted, then the = equals sign is also omitted. The short style is only used for Category and Script properties because these properties are very common and their omission is unambiguous.

In actual practice, you can mix type names and values that are omitted, abbreviated, or full. For example, if Category=Unassigned you could use what is in the table explicitly, \p{gc=Unassigned}, \p{Category=Cn}, or \p{Unassigned}.

When these are processed, case and whitespace are ignored so you may use them for clarity, if desired. For example, \p{Category = Uppercase Letter} or \p{Category = uppercase letter}.

For a list of properties supported by ICU, see ICU User Guide - Unicode Properties.

See Also

Unicode Properties

The following tables list some of the commonly used Unicode Properties, which can be matched in a regular expression with \p{Property}. The tables were created from the Unicode 5.2 Unicode Character Database, which is the version used by ICU that ships with Mac OS X 10.6.

Category
LLetter
LCCasedLetter
LuUppercaseLetter
LlLowercaseLetter
LtTitlecaseLetter
LmModifierLetter
LoOtherLetter
 
PPunctuation
PcConnectorPunctuation
PdDashPunctuation
PsOpenPunctuation
PeClosePunctuation
PiInitialPunctuation
PfFinalPunctuation
PoOtherPunctuation
 
NNumber
NdDecimalNumber
NlLetterNumber
NoOtherNumber
 
MMark
MnNonspacingMark
McSpacingMark
MeEnclosingMark
 
SSymbol
SmMathSymbol
ScCurrencySymbol
SkModifierSymbol
SoOtherSymbol
 
ZSeparator
ZsSpaceSeparator
ZlLineSeparator
ZpParagraphSeparator
 
COther
CcControl
CfFormat
CsSurrogate
CoPrivateUse
CnUnassigned
Script
ArabicArmenianBalinese
BengaliBopomofoBraille
BugineseBuhidCanadian_​AboriginalCanadian_Aboriginal
CarianChamCherokee
CommonCopticCuneiform
CypriotCyrillicDeseret
DevanagariEthiopicGeorgian
GlagoliticGothicGreek
GujaratiGurmukhiHan
HangulHanunooHebrew
HiraganaInheritedKannada
KatakanaKayah_LiKharoshthi
KhmerLaoLatin
LepchaLimbuLinear_B
LycianLydianMalayalam
MongolianMyanmarNew_Tai_Lue
NkoOghamOl_Chiki
Old_ItalicOld_PersianOriya
OsmanyaPhags_PaPhoenician
RejangRunicSaurashtra
ShavianSinhalaSundanese
Syloti_NagriSyriacTagalog
TagbanwaTai_LeTamil
TeluguThaanaThai
TibetanTifinaghUgaritic
UnknownVaiYi
Extended Property Class
ASCII_Hex_DigitAlphabetic
Bidi_ControlDash
Default_​Ignorable_​Code_​PointDefault_Ignorable_Code_PointDeprecated
DiacriticExtender
Grapheme_BaseGrapheme_Extend
Grapheme_LinkHex_Digit
HyphenIDS_​Binary_​OperatorIDS_Binary_Operator
IDS_Trinary_OperatorID_Continue
ID_StartIdeographic
Join_ControlLogical_​Order_​ExceptionLogical_Order_Exception
LowercaseMath
Noncharacter_Code_PointOther_Alphabetic
Other_​Default_​Ignorable_​Code_​PointOther_Default_Ignorable_Code_PointOther_​Grapheme_​ExtendOther_Grapheme_Extend
Other_ID_ContinueOther_ID_Start
Other_LowercaseOther_Math
Other_UppercasePattern_Syntax
Pattern_White_SpaceQuotation_Mark
RadicalSTerm
Soft_DottedTerminal_​PunctuationTerminal_Punctuation
Unified_IdeographUppercase
Variation_SelectorWhite_Space
XID_ContinueXID_Start

Unicode Character Database

Unicode properties are defined in the Unicode Character Database, or UCD. From time to time the UCD is revised and updated. The properties available, and the definition of the characters they match, depend on the UCD that ICU was built with.

Note:

In general, the ICU and UCD versions change with each major operating system release.

See Also

ICU Replacement Text Syntax

Replacement Text Syntax
CharacterDescription
$n
The text of capture group n will be substituted for $n. n must be ≥ 0 and not greater than the number of capture groups. A $ not followed by a digit has no special meaning, and will appear in the substitution text as itself, a $.
Important:
Methods will raise a RKLICURegexException if n is greater than the number of capture groups in the regular expression.
\Treat the character following the backslash as a literal, suppressing any special meaning. Backslash escaping in substitution text is only required for $ and \, but may proceed any character. The backslash itself will not be copied to the substitution text.
See Also

RegexKitLite Cookbook

Some people, when confronted with a problem, think “I know, I'll use regular expressions.” Now they have two problems.
Jamie Zawinski

This section contains a collection of regular expressions and example code demonstrating how RegexKitLite makes some common programming choirs easier. RegexKitLite makes it easy to match part of a string and extract just that part, or even create an entirely new string using just a few pieces of the original string. A great example of this is a string that contains a URL and you need to extract just a part of it, perhaps the host or maybe just the port used. This example demonstrates how easy it is to extract the port used from a URL, which is then converted in to a NSInteger value:

searchString = @"http://www.example.com:8080/index.html"; regexString = @"\\bhttps?://[a-zA-Z0-9\\-.]+(?::(\\d+))?(?:(?:/[a-zA-Z0-9\\-._?,'+\\&%$=~*!():@\\\\]*)+)?"; NSInteger portInteger = [[searchString stringByMatching:regexString capture:1L] integerValue]; NSLog(@"portInteger: '%ld'", (long)portInteger); // 2008-10-15 08:52:52.500 host_port[8021:807] portInteger: '8080'

Inside you'll find more examples like this that you can use as the starting point for your own regular expression pattern matching solution. Keep in mind that these are meant to be examples to help get you started and not necessarily the ideal solution for every need. Trade‑offs are usually made when creating a regular expression, matching an email address is a perfect example of this. A regular expression that precisely matches the formal definition of email address is both complicated and usually unnecessary. Knowing which trade‑offs are acceptable requires that you understand what it is you're trying to match, the data that you're searching through, and the requirements and uses of the matched results. It won't take long until you gain an appreciation for Jamie Zawinski's infamous quote.

See Also

Enhanced Copy To Clipboard Functionality

Copy Regex to Clipboard Preferences
Copied Regex Escape Style:
Escape Style Options: Smart escape
  Use C99 \u character escapes
  Escaped Unicode in NSString literals

Note:
The escaped regular expression may be slightly different than the original because some escape sequences are parsed and rewritten.
Note:
Use of \u character escapes requires the C99 standard with the execution character set set to UTF-8 (the default for gcc).
Note:
Escaped Unicode in NSString literals requires Xcode ≥ 3.0 and gcc ≥ 4.0.

Escape Style Preview:

Regex Escape Tool

Preview:

 

This browser supports Copy To Clipboard features that can be used by this document. Using the Copy Regex to Clipboard Preferences interface, you can configure the behavior of Edit Copy so that when you select and copy a regular expression in this document, the clipboard will contain the selected regular expression that has been escaped according to your preference choices. The resulting text in the clipboard can be pasted directly in to your source code, no further modification of the regular expression is required.

Your preference choices are stored in the browser using a cookie. Provided that cookies are enabled in your browser, your preferences should persist across browser sessions. The preference panel also contains a preview of what would be copied to the clipboard for some example regular expressions using the current settings. When you change a setting in the preferences, the examples will transition between the previous settings and the new settings, while briefly highlighting the differences between the two settings, so you can determine the effect the change has caused.

Some of the problems of using regular expressions unmodified in C and Objective-C are:

There are several different ways to escape Unicode characters. When using Unicode characters in NSString literals, the best method depends on the version of Xcode that is used to compile the source and what C standard you are using (i.e., gcc -std=(c|gnu)99).

When the Copied Regex Escape Style is set to None, then the selected regular expression is copied unmodified to the clipboard.

When the Copied Regex Escape Style is set to Escape Only, C String, or NSString then these details are automatically dealt with for you. This allows you to paste the selected regular expression directly in to your source code without having to escape every use of \ by hand, or having to convert the Unicode characters in to a acceptable format.

Escape Option Details

Smart escape

If this option is disabled then all occurrences of \ and " are escaped with a \, resulting in \\ or \". No other processing is performed. Since this prevents the C compiler from interpreting any special meaning of \ sequences, it is the safest option. The ICU regular expression engine is then responsible for interpreting the meaning of any \ escaped character sequences.

If this option is enabled then certain escape sequences are interpreted and rewritten using the current preference options.

Use C99 \u character escapes

Normally, Unicode characters are embedded in string literals as the characters UTF-8 byte sequence using \ddd octal escapes. When this option is enabled, the C99 \u and \U character escape sequences are used instead. gcc will issue a warning if \u character escape sequences are present and the compiler is not configured to use the C99 (or later) standard (i.e., gcc -std=(c|gnu)99).

Under the C99 standard, \u and \U are used to specify a universal character name, which is a character encoded in the ISO/IEC 10646 character set (essentially identical to Unicode in this context). Ultimately, a universal character name is translated in to a sequence of bytes needed to represent the designated character in the C environments execution character set. Usually, although certainly not always, a string literal should be encoded as UTF-8, which happens to be the default execution character set for gcc. This is an important point to remember because the more convenient and easier to use \u escape sequences are not guaranteed to convert in to a specific sequence of bytes, unlike an octal \ddd or hex \xhh escape sequence. There is currently no way to specify that a particular string literal should always be translated using a specific character set encoding. This may result in undefined behavior if the \u universal character name is not translated in to the expected character set, which in this case must be UTF-8.

Escaped Unicode in NSString literals

Prior to Xcode 3.0, gcc only supported the use of ASCII characters (i.e., characters ≤ 127) in constant NSString literals. If one needed to include Unicode characters in an NSString, one would typically convert the string in to UTF-8, and then create a NSString at run time using the stringWithUTF8String: method, with the UTF-8 encoded C string passed as the argument. For example, "€1.99", which contains the € euro symbol, would be created using the following:

NSString *euroString = [NSString stringWithUTF8String:"\342\202\2541.99"]; // or with C99 \u character escapes: NSString *euroString = [NSString stringWithUTF8String:"\u20ac1.99"];

One of the obvious disadvantages of this approach is that it instantiates a new, autoreleased NSString each time it's used, unlike a constant NSString literal like @"$1.99". Beginning with Xcode 3.0 and gcc 4.0, constant NSString literals that contain Unicode characters can be specified directly in source-code using the standard @"" syntax. For example:

NSString *euroString = @"\342\202\2541.99"; // or with C99 \u character escapes: NSString *euroString = @"\u20ac1.99";

The compiler converts these strings to UTF-16 using the endianness of the target architecture. Since Mach-O object file format allows for multiple architectures, this allows each architecture to encode the string as native UTF-16 byte ordering for that architecture, so there are no issues with proper byte ordering. Within the object file itself, these strings are essentially identical to their ASCII-only counterparts: effectively they are pre-instantiated objects. The only real difference is that the compiler sets some internal CFString bits differently so that the CFString object knows that the strings data is encoded as UTF-16 and not simple 8-bit data.

Although this functionality has been present since the release of 10.5, it has only recently been documented in The Objective-C 2.0 Programming Language - Compiler Directives, under the @"string" entry. A copy of the relevant text is provided below:

On Mac OS X v10.4 and earlier, the string must be 7-bit ASCII-encoded. On Mac OS X v10.5 and later (with Xcode 3.0 and later), you can also use UTF-16 encoded strings. (The runtime from Mac OS X v10.2 and later supports UTF-16 encoded strings, so if you use Mac OS X v10.5 to compile an application for Mac OS X v10.2 and later, you can use UTF-16 encoded strings.)

Visual Indication that the Selected Text has been Escaped

This document will briefly add a red outline around the selected text as a visual aid in determining whether or not the selected text was modified before placing it in the clipboard. In addition to this, a HUD display will drop down briefly at the top of the documents window and will display the escaped text that was placed in to the clipboard. Here is an example of what would be displayed if the NSString escape style was selected:

The outline that is display when a regular expression is selected and copied to the clipboard: (\d+):\s*(.*)
The drop down HUD that would be displayed:
Copied to Clipboard as:
@"(\\d+):\\s*(.*)"

The selected text will only be escaped if it can be determined that it is a regular expression, and the selection only contains a regular expression. If the regular expression is part of an overall larger selection then the text that is copied to the clipboard is not modified.

Note:

The method used to escape regular expressions is only a simple set of heuristics. While these heuristics seem to work correctly for a variety of inputs, it is by no means error-free. You should always double check the escaped output for errors.

Regex Escape Tool

Also available is the Regex Escape Tool which allows you to enter a regular expression and have it immediately escaped using the current preference settings. This can be useful if you have a complex regular expression that needs to be escaped before you can use it in a constant NSString literal. It can also be used to easily create a constant NSString literal that contains several Unicode characters that would otherwise have to be manually converted by hand.

Pattern Matching Recipes

Numbers

DescriptionRegexExamples
Integer[+\-]?[0-9]+123-42+23
Hex Number0[xX][0-9a-fA-F]+0x00xdeadbeef0xF3
Floating Point[+\-]?(?:[0-9]*\.[0-9]+|[0-9]+\.)123..123+.42
Floating Point with Exponent[+\-]?(?:[0-9]*\.[0-9]+|[0-9]+\.)(?:[eE][+\-]?[0-9]+)?123..12310.0E131.23e-7
Comma Separated Number[0-9]{1,3}(?:,[0-9]{3})*421,2341,234,567
Comma Separated Number[0-9]{1,3}(?:,[0-9]{3})*(?:\.[0-9]+)?421,2341,234,567.89
Extracting and Converting Numbers

NSString includes several methods for converting the contents of the string in to a numeric value in the various C primitive types. The following demonstrates the matching of an int and double in a NSString, and then converting the matched string in to its base type.

Integer conversion…
NSString *searchString = @"The int 5542 to convert"; NSString *regexString = @"([+\\-]?[0-9]+)"; int matchedInt = [[searchString stringByMatching:regexString capture:1L] intValue];

The variable matchedInt now contains the value of 5542.

Floating Point conversion…
NSString *searchString = @"The double 4321.9876 to convert"; NSString *regexString = @"([+\\-]?(?:[0-9]*\\.[0-9]+|[0-9]+\\.))"; double matchedDouble = [[searchString stringByMatching:regexString capture:1L] doubleValue];

The variable matchedDouble now contains the value of 4321.9876. doubleValue can even convert numbers that are in scientific notation, which represent numbers as n × 10exp:

Floating Point conversion…
NSString *searchString = @"The double 1.010489e5 to convert"; NSString *regexString = @"([+\\-]?(?:[0-9]*\\.[0-9]+|[0-9]+\\.)(?:[eE][+\\-]?[0-9]+)?)"; double matchedDouble = [[searchString stringByMatching:regexString capture:1L] doubleValue];

The variable matchedDouble now contains the value of 101048.9.

Extracting and Converting Hex Numbers

Converting a string that contains a hex number in to a more basic type, such as an int, takes a little more work. Unfortunately, Foundation does not provide an easy way to convert a hex value in a string in to a more basic type as it does with intValue or doubleValue. Thankfully the standard C library provides a set of functions for performing such a conversion. For this example we will use the strtol() (string to long) function to convert the hex value we've extracted from searchString. We can not pass the pointer to the NSString object that contains the matched hex value since strtol() is part of the standard C library which can only work on pointers to C strings. We use the UTF8String method to get a pointer to a compatible C string of the matched hex value.

Hex conversion…
NSString *searchString = @"A hex value: 0x0badf00d"; NSString *regexString = @"\\b(0[xX][0-9a-fA-F]+)\\b"; NSString *hexString = [searchString stringByMatching:regexString capture:1L]; // Use strtol() to convert the string to a long. long hexLong = strtol([hexString UTF8String], NULL, 16); NSLog(@"hexLong: 0x%lx / %ld", (u_long)hexLong, hexLong); // 2008-09-01 09:40:44.848 hex_example[30583:807] hexLong: 0xbadf00d / 195948557

The full set of string to… functions are: strtol(), strtoll(), strtoul(), and strtoull(). These convert a string value, from base 2 to base 36, in to a long, long long, unsigned long, and unsigned long long respectively.

Adding Hex Value Conversions to NSString

Since it seems to be a frequently asked question, and a common search engine query for RegexKit web site visitors, here is a NSString category addition that converts the receivers text in to a NSInteger value. This is the same functionality as intValue or doubleValue, except that it converts hexadecimal text values instead of decimal text values.

Note:

The following code can also be found in the RegexKitLite distributions examples/ directory.

The example conversion code is fairly quick since it uses Core Foundation directly along with the stack to hold any temporary string conversions. Any whitespace at the beginning of the string will be skipped and the hexadecimal text to be converted may be optionally prefixed with either 0x or 0X. Returns 0 if the receiver does not begin with a valid hexadecimal text representation. Refer to strtol(3) for additional conversion details.

Important:

If the receiver needs to be converted in to an encoding that is compatible with strtol(), only the first sixty characters of the receiver are converted.

File name:NSString-HexConversion.h
#import <Foundation/NSString.h> @interface NSString (HexConversion) -(NSInteger)hexValue; @end
File name:NSString-HexConversion.m
#import "NSString-HexConversion.h" #import <CoreFoundation/CFString.h> #include <stdlib.h> @implementation NSString (HexConversion) -(NSInteger)hexValue { CFStringRef cfSelf = (CFStringRef)self; UInt8 buffer[64]; const char *cptr; if((cptr = CFStringGetCStringPtr(cfSelf, kCFStringEncodingMacRoman)) == NULL) { CFRange range = CFRangeMake(0L, CFStringGetLength(cfSelf)); CFIndex usedBytes = 0L; CFStringGetBytes(cfSelf, range, kCFStringEncodingUTF8, '?', false, buffer, 60L, &usedBytes); buffer[usedBytes] = 0; cptr = (const char *)buffer; } return((NSInteger)strtol(cptr, NULL, 16)); } @end
See Also

Text Files

DescriptionRegex
Empty Line(?m:^$)
Empty or Whitespace Only Line(?m-s:^\s*$)
Strip Leading Whitespace(?m-s:^\s*(.*?)$)
Strip Trailing Whitespace(?m-s:^(.*?)\s*$)
Strip Leading and Trailing Whitespace(?m-s:^\s*(.*?)\s*$)
Quoted String, Can Span Multiple Lines, May Contain \""(?:[^"\\]*+|\\.)*"
Quoted String, Single Line Only, May Contain \""(?:[^"\\\r\n]*+|\\[^\r\n])*"
HTML Comment(?s:<--.*?-->)
Perl / Shell Comment(?m-s:#.*$)
C, C++, or ObjC Comment(?m-s://.*$)
C, C++, or ObjC Comment and Leading Whitespace(?m-s:\s*//.*$)
C, C++, or ObjC Comment(?s:/\*.*?\*/)
The Newline Debacle

Unfortunately, when processing text files, there is no standard 'newline' character or character sequence. Today this most commonly surfaces when converting text between Microsoft Windows / MS-DOS and Unix / Mac OS X. The reason for the proliferation of newline standards is largely historical and goes back many decades. Below is a table of the dominant newline character sequence 'standards':

DescriptionSequenceC StringControlCommon Uses
Line Feed\u000A\n^JUnix, Amiga, Mac OS X
Vertical Tab\u000B\v^K 
Form Feed\u000C\f^L 
Carriage Return\u000D\r^MApple ][, Mac OS ≤ 9
Next Line (NEL)\u0085 IBM / EBCDIC
Line Separator\u2028 Unicode
Paragraph Separator\u2029 Unicode
Carriage Return + Line Feed\u000D\u000A\r\n^M^JMS-DOS, Windows

Ideally, one should be flexible enough to accept any of these character sequences if one has to process text files, especially if the origin of those text files is not known. Thankfully, regular expressions excel at just such a task. Below is a regular expression pattern that will match any of the above character sequences. This is also the character sequence that the metacharacter $ matches.

DescriptionRegexNotes
Any newline(?:\r\n|[\n\v\f\r\x85\p{Zl}\p{Zp}])UTS #18 recommended. Character sequence that $ matches.
See Also
Matching the Beginning and End of a Line

It is often necessary to work with the individual lines of a file. There are two regular expression metacharacters, ^ and $, that match the beginning and end of a line, respectively. However, exactly what is matched by ^ and $ depends on whether or not the multi-line option is enabled for the regular expression, which by default is disabled. It can be enabled for the entire regular expression by passing RKLMultiline via the options: method argument, or within the regular expression using the options syntax— (?m:).

If multi-line is disabled, then ^ and $ match the beginning and end of the entire string. If there is a newline character sequence at the very end of the string, then $ will match the character just before the newline character sequence. Any newline character sequences in the middle of the string will not be matched.

If multi-line is enabled, then ^ and $ match the beginning and end of a line, where the end of a line is the newline character sequence. The metacharacter ^ matches either the first character in the string, or the first character following a newline character sequence. The metacharacter $ matches either the last character in the string, or the character just before a newline character sequence.

Creating a NSArray Containing Every Line in a String

A common text processing pattern is to process a file one line at a time. Using the recommended regular expression for matching any newline and the componentsSeparatedByRegex: method, you can easily create a NSArray containing every line in a file and process it one line at a time:

Process every line…
NSString *fileNameString = @"example"; NSString *regexString = @"(?:\r\n|[\n\v\f\r\302\205\\p{Zl}\\p{Zp}])"; NSError *error = NULL; NSString *fileString = [NSString stringWithContentsOfFile:fileNameString usedEncoding:NULL error:&error]; if(fileString) { NSArray *linesArray = [fileString componentsSeparatedByRegex:regexString]; for(NSString *lineString in linesArray) { // ObjC 2.0 for…in loop. // Per line processing. } } else { NSLog(@"Error reading file '%@'", fileNameString); if(error) { NSLog(@"Error: %@", error); } }

The componentsSeparatedByRegex: method effectively 'chops off' the matched regular expression, or in this case any newline character. In the example above, within the for…in loop, lineString will not have a newline character at the end of the string.

Parsing CSV Data
DescriptionRegex
Split CSV line ,(?=(?:(?:[^"\\]*+|\\")*"(?:[^"\\]*+|\\")*")​*(?!(?:[^"\\]*+|\\")*"(?:[^"\\]*+|\\")*$)) ,(?=(?:(?:[^"\\]*+|\\")*"(?:[^"\\]*+|\\")*")*(?!(?:[^"\\]*+|\\")*"(?:[^"\\]*+|\\")*$))

This regular expression essentially works by ensuring that there are an even number of unescaped " quotes following a , comma. This is done by using look-head assertions. The first look-head assertion, (?=, is a pattern that matches zero or more strings that contain two " characters. Then, a negative look-head assertion matches a single, unpaired " quote character remaining at the $ end of the line. It also uses possessive matches in the form of *+ for speed, which prevents the regular expression engine from backtracking excessively. It's certainly not a beginners regular expression.

The following is used as a substitute for a CSV data file in the example below.

Example CSV data…

NSString *csvFileString = @"RegexKitLite,1.0,\"Mar 23, 2008\",27004\n" @"RegexKitLite,1.1,\"Mar 28, 2008\",28081\n" @"RegexKitLite,1.2,\"Apr 01, 2008\",28765\n" @"RegexKitLite,2.0,\"Jul 07, 2008\",40569\n" @"RegexKitLite,2.1,\"Jul 12, 2008\",40660\n";

This example really highlights the power of regular expressions when it comes to processing text. It takes just 17 lines, which includes comments, to parse a CSV data file of any newline type and create a row by column of NSArray values of the results while correctly handling " quoted values, including escaped \" quotes.

Parse CSV data…

NSString *newlineRegex = @"(?:\r\n|[\n\v\f\r\\x85\\p{Zl}\\p{Zp}])"; NSString *splitCSVLineRegex = @",(?=(?:(?:[^\"\\\\]*+|\\\\\")*​\"(?:[^\"\\\\]*+|\\\\\")*\")*​(?!(?:[^\"\\\\]*+|\\\\\")*​\"(?:[^\"\\\\]*+|\\\\\")*$)),(?=(?:(?:[^\"\\\\]*+|\\\\\")*\"(?:[^\"\\\\]*+|\\\\\")*\")*(?!(?:[^\"\\\\]*+|\\\\\")*\"(?:[^\"\\\\]*+|\\\\\")*$))"; // Create a NSArray of every line in csvFileString. NSArray *csvLinesArray = [csvFileString componentsSeparatedByRegex:newlineRegex]; // Create an id array to hold the comma split line results. id splitLines[[csvLinesArray count]]; // C99 variable length array. NSUInteger splitLinesIndex = 0UL;   // Index of next splitLines[] member. for(NSString *csvLineString in csvLinesArray) {   // ObjC 2 for…in loop. if([csvLineString isMatchedByRegex:@"^\\s*$"]) { continue; } // Skip empty lines. splitLines[splitLinesIndex++] = [csvLineString componentsSeparatedByRegex:splitCSVLineRegex]; } // Gather up all the individual comma split results in to a single NSArray. NSArray *splitLinesArray = [NSArray arrayWithObjects:&splitLines[0] count:splitLinesIndex];

Network and URL

DescriptionRegex
HTTP\bhttps?://[a-zA-Z0-9\-.]+(?:(?:/[a-zA-Z0-9\-._?,'+\&%$=~*!():@\\]*)+)?
HTTP\b(https?)://([a-zA-Z0-9\-.]+)((?:/[a-zA-Z0-9\-._?,'+\&%$=~*!():@\\]*)+)?
HTTP\b(https?)://(?:(\S+?)(?::(\S+?))?@)?([a-zA-Z0-9\-.]+)(?::(\d+))?((?:/[a-zA-Z0-9\-._?,'+\&%$=~*!():@\\]*)+)?
E-Mail\b([a-zA-Z0-9%_.+\-]+)@([a-zA-Z0-9.\-]+?\.[a-zA-Z]{2,6})\b
Hostname\b(?:[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}?[a-zA-Z0-9]\.)+[a-zA-Z]{2,6}\b
IP\b(?:\d{1,3}\.){3}\d{1,3}\b
IP with Optional Netmask\b((?:\d{1,3}\.){3}\d{1,3})(?:/(\d{1,2}))?\b
IP or Hostname\b(?:(?:\d{1,3}\.){3}\d{1,3}|(?:[a-zA-Z0-9][a-zA-Z0-9\-]{0,61}?[a-zA-Z0-9]\.)+[a-zA-Z]{2,6})\b
Creating a NSDictionary of URL Information

The following example demonstrates how to match several fields in a URL and create a NSDictionary with the extracted results. Only the capture groups that result in a successful match will create a corresponding key in the dictionary.

HTTP URL…
NSString *searchString = @"http://johndoe:secret@www.example.com:8080/private/mail/index.html"; NSString *regexString = @"\\b(https?)://(?:(\\S+?)(?::(\\S+?))?@)?([a-zA-Z0-9\\-.]+)(?::(\\d+))?((?:/[a-zA-Z0-9\\-._?,'+\\&%$=~*!():@\\\\]*)+)?"; if([searchString isMatchedByRegex:regexString]) { NSString *protocolString = [searchString stringByMatching:regexString capture:1L]; NSString *userString = [searchString stringByMatching:regexString capture:2L]; NSString *passwordString = [searchString stringByMatching:regexString capture:3L]; NSString *hostString = [searchString stringByMatching:regexString capture:4L]; NSString *portString = [searchString stringByMatching:regexString capture:5L]; NSString *pathString = [searchString stringByMatching:regexString capture:6L]; NSMutableDictionary *urlDictionary = [NSMutableDictionary dictionary]; if(protocolString) { [urlDictionary setObject:protocolString forKey:@"protocol"]; } if(userString) { [urlDictionary setObject:userString forKey:@"user"]; } if(passwordString) { [urlDictionary setObject:passwordString forKey:@"password"]; } if(hostString) { [urlDictionary setObject:hostString forKey:@"host"]; } if(portString) { [urlDictionary setObject:portString forKey:@"port"]; } if(pathString) { [urlDictionary setObject:pathString forKey:@"path"]; } NSLog(@"urlDictionary: %@", urlDictionary); }

RegexKitLite 4.0 adds a new method, dictionaryByMatchingRegex:…, that makes the creation of NSDictionary objects like this much easier, as the following example demonstrates:

RegexKitLite ≥ 4.0 example…
NSString *searchString = @"http://johndoe:secret@www.example.com:8080/private/mail/index.html"; NSString *regexString = @"\\b(https?)://(?:(\\S+?)(?::(\\S+?))?@)?([a-zA-Z0-9\\-.]+)(?::(\\d+))?((?:/[a-zA-Z0-9\\-._?,'+\\&%$=~*!():@\\\\]*)+)?"; NSDictionary *urlDictionary = [searchString dictionaryByMatchingRegex:regexString withKeysAndCaptures:@"protocol", 1, @"user", 2, @"password", 3, @"host", 4, @"port", 5, @"path", 6, NULL]; if(urlDictionary != NULL) { NSLog(@"urlDictionary: %@", urlDictionary); }
Note:

Other than the difference in mutability for the dictionary containing the result, the RegexKitLite 4.0 dictionaryByMatchingRegex:… example produces the same result as the more verbose, pre-4.0 example.

These examples can form the basis of a function or method that takes a NSString as an argument and returns a NSDictionary as a result, maybe even as a category addition to NSString. The following is the output when the examples above are compiled and run:

shell% ./http_example 2008-09-01 10:57:55.245 test_nsstring[31306:807] urlDictionary: { host = "www.example.com"; password = secret; path = "/private/mail/index.html"; port = 8080; protocol = http; user = johndoe; } shell%

Adding RegexKitLite to your Project

Note:

The following outlines a typical set of steps that one would perform. This is not the only way, nor the required way to add RegexKitLite to your application. They may not be correct for your project as each project is unique. They are an overview for those unfamiliar with adding additional shared libraries to the list of libraries your application links against.

Outline of Required Steps

The following outlines the steps required to use RegexKitLite in your project.

See Also

Adding RegexKitLite using Xcode

Important:
These instructions apply to Xcode versions 2.4.1 and 3.0. Other versions should be similar, but may vary for specific details.

Unfortunately, adding additional dynamic shared libraries that your application links to is not a straightforward process in Xcode, nor is there any recommended standard way. Two options are presented below— the first is the 'easy' way that alters your applications Xcode build settings to pass an additional command line argument directly to the linker. The second option attempts to add the ICU dynamic shared library to the list of resources for your project and configuring your executable to link against the added resource.

The 'easy' way is the recommended way to link against the ICU dynamic shared library.

The Easy Way To Add The ICU Library
  1. First, determine the build settings layer of your project that should have altered linking configuration change applied to. The build settings in Xcode are divided in to layers and each layer inherits the build settings from the layer above it. The top, global layer is Project Settings, followed by Target Settings, and finally the most specific layer Executable Settings. If your project is large enough to have multiple targets and executables, you probably have an idea which layer is appropriate. If you are unsure or unfamiliar with the different layers, Target Settings is recommended.

  2. Select the appropriate layer from the Project menu. If you are unsure, Project Edit Active Target is recommended.

  3. Select Build from the tab near the top of the Target Info window. Find the Other Linker Flags build setting from the many build settings available and edit it. Add -licucore [dash ell icucore as a single word, without spaces]. If there are already other flags present, it is recommended that you add -licucore to the end of the existing flags.

    Important:
    If other linker flags are present, there must be at least one space separating -licucore from the other linker flags. For example, -flag1 -licucore -flag2
    Note:
    The Configuration drop down menu controls which build configuration the changes you make are applied to. All Configurations should be selected if this is the first time your are making these changes.
  4. Follow the Add The RegexKitLite Source Files To Your Project steps below.
See Also
The Hard Way To Add The ICU Library
  1. First, add the ICU dynamic shared library to your Xcode project. You may choose to add the library to any group in your project, and which groups are created by default is dependent on the template type you chose when you created your project. For a typical Cocoa application project, a good choice is the Frameworks group. To add the ICU dynamic shared library, control/right-click on the Framework group and choose Add Existing Files…

  2. Next, you will need to choose the ICU dynamic shared library file to add. Exactly which file to choose depends on your project, but a fairly safe choice is to select /Developer/SDKs/MacOSX10.6.sdk/usr/lib/libicucore.dylib. You may have installed your developer tools in a different location than the default /Developer directory, and the Mac OS X SDK version should be the one your project is targeting, typically the latest one available.

  3. Then, in the dialog that follows, make sure that Copy items into… is unselected. Select the targets you will be using RegexKitLite in and then click Add to add the ICU dynamic shared library to your project.

  4. Once the ICU dynamic shared library is added to your project, you will need to add it to the libraries that your executable is linked with. To do so, expand the Targets group, and then expand the executable targets you will be using RegexKitLite in. You will then need to select the libicucore.dylib file that you added in the previous step and drag it in to the Link Binary With Libraries group for each executable target that you will be using RegexKitLite in. The order of the files within the Link Binary With Libraries group is not important, and for a typical Cocoa application the group will contain the Cocoa.framework file.

Add The RegexKitLite Source Files To Your Project
  1. Next, add the RegexKitLite source files to your Xcode project. In the Groups & Files outline view on the left, control/right-click on the group that would like to add the files to, then select Add Existing Files…

    Note:

    You can perform the following steps once for each file (RegexKitLite.h and RegexKitLite.m), or once by selecting both files from the file dialog.

  2. Select the RegexKitLite.h and / or RegexKitLite.m file from the file chooser dialog.

  3. The next dialog will present you with several options. If you have not already copied the RegexKitLite files in to your projects directory, you may want to click on the Copy items into… option. Select the targets that you would like add the RegexKitLite functionality to.

  4. Finally, you will need to include the RegexKitLite.h header file. The best way to do this is very dependent on your project. If your project consists of only half a dozen source files, you can add:

    #import "RegexKitLite.h"

    manually to each source file that makes uses of RegexKitLites features. If your project has grown beyond this, you've probably already organized a common "master" header to include to capture headers that are required by nearly all source files already.

Adding RegexKitLite using the Shell

Using RegexKitLite from the shell is also easy. Again, you need to add the header #import to the appropriate source files. Then, to link to the ICU library, you typically only need to add -licucore, just as you would any other library. Consider the following example:

File name:link_example.m
#import <Foundation/NSObjCRuntime.h> #import <Foundation/NSAutoreleasePool.h> #import "RegexKitLite.h" int main(int argc, char *argv[]) { NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; // Copyright COPYRIGHT_SIGN APPROXIMATELY_EQUAL_TO 2008 // Copyright \u00a9 \u2245 2008 char *utf8CString = "Copyright \xC2\xA9 \xE2\x89\x85 2008"; NSString *regexString = @"Copyright (.*) (\\d+)"; NSString *subjectString = [NSString stringWithUTF8String:utf8CString]; NSString *matchedString = [subjectString stringByMatching:regexString capture:1L]; NSLog(@"subject: \"%@\"", subjectString); NSLog(@"matched: \"%@\"", matchedString); [pool release]; return(0); }

Compiled and run from the shell:

shell% cd examples shell% gcc -g -I.. -o link_example link_example.m ../RegexKitLite.m -framework Foundation -licucore shell% ./link_example 2008-03-14 03:52:51.187 test[15283:807] subject: "Copyright © ≅ 2008" 2008-03-14 03:52:51.269 test[15283:807] matched: "© ≅" shell%

RegexKitLite NSString Additions Reference

Extends by categoryNSString, NSMutableString
RegexKitLite4.0
Declared in
  • RegexKitLite.h

Overview

RegexKitLite is not meant to be a full featured regular expression framework. Because of this, it provides only the basic primitives needed to create additional functionality. It is ideal for developers who:

RegexKitLite consists of only two files, the header file RegexKitLite.h and RegexKitLite.m. The only other requirement is to link with the ICU library that comes with Mac OS X. No new classes are created, all functionality is provided as a category extension to the NSString and NSMutableString classes.

See Also

Compile Time Preprocessor Tunables

The settings listed below are implemented using the C Preprocessor. Some of the setting are simple boolean enabled or disabled settings, while others specify a value, such as the number of cached compiled regular expressions. There are several ways to alter these settings, but if you are not familiar with this style of compile time configuration settings and how to alter them using the C Preprocessor, it is recommended that you use the default values provided.

SettingDefaultDescription
NS_​BLOCK_​ASSERTIONSNS_BLOCK_ASSERTIONSn/aRegexKitLite contains a number of extra run-time assertion checks that can be disabled with this flag. The standard NSException.h assertion macros are not used because of the multithreading lock. This flag is typically set for Release style builds where the additional error checking is no longer necessary.
RKL_​APPEND_​TO_​ICU_​FUNCTIONSRKL_APPEND_TO_ICU_FUNCTIONSNoneThis flag is useful if you are supplying your own version of the ICU library. When set, this preprocessor define causes the ICU functions used by RegexKitLite to have the value of RKL_​APPEND_​TO_​ICU_​FUNCTIONSRKL_APPEND_TO_ICU_FUNCTIONS appended to them. For example, if RKL_​APPEND_​TO_​ICU_​FUNCTIONSRKL_APPEND_TO_ICU_FUNCTIONS is set to _4_0 (i.e., -DRKL_​APPEND_​TO_​ICU_​FUNCTIONS=_4_0-DRKL_APPEND_TO_ICU_FUNCTIONS=_4_0), it would cause uregex_find() to become uregex_find_4_0().
RKL_​BLOCKSRKL_BLOCKSAutomatic Enables blocks support. This feature is automatically enabled if NS_​BLOCKS_​AVAILABLENS_BLOCKS_AVAILABLE is set to 1, which is typically set if support for blocks is appropriate. At the time of this writing, this typically means that the Xcode setting for the minimum version of Mac OS X supported must be 10.6. This feature may be explicitly disabled under all circumstances by setting its value to 0, or alternatively it can be explicitly enabled under all circumstances by setting its value to 1. The behavior is undefined if RKL_​BLOCKSRKL_BLOCKS is set to 1 and the compiler does not support the blocks language extension or if the run-time does not support blocks.
RKL_​CACHE_​SIZERKL_CACHE_SIZE13 RegexKitLite uses a 4-way set associative cache and RKL_​CACHE_​SIZERKL_CACHE_SIZE controls the number of sets in the cache. The total number of compiled regular expressions that can be cached is RKL_​CACHE_​SIZERKL_CACHE_SIZE * 4, for a default value of 52. RKL_​CACHE_​SIZERKL_CACHE_SIZE should always be a prime number to maximize the use of the cache.
RKL_​DTRACERKL_DTRACEDisabledThis preprocessor define controls whether or not RegexKitLite provider DTrace probe points are enabled. This feature may be explicitly disabled under all circumstances by setting its value to 0.
RKL_​FAST_​MUTABLE_​CHECKRKL_FAST_MUTABLE_CHECKDisabledEnables the use of the undocumented, private Core Foundation __CFStringIsMutable() function to determine if the string to be searched is immutable. This can significantly increase the number of matches per second that can be performed on immutable strings since a number of mutation checks can be safely skipped.
RKL_​FIXED_​LENGTHRKL_FIXED_LENGTH2048Sets the size of the fixed length UTF-16 conversion cache buffer. Strings that need to be converted to UTF-16 that have a length less than this size will use the fixed length conversion cache. Using a fixed sized buffer for all small strings means less malloc() overhead, heap fragmentation, and reduces the chances of a memory leak occurring.
RKL_​METHOD_​PREPENDRKL_METHOD_PREPENDNoneWhen set, this preprocessor define causes the RegexKitLite methods defined in RegexKitLite.h to have the value of RKL_​METHOD_​PREPENDRKL_METHOD_PREPEND prepended to them. For example, if RKL_​METHOD_​PREPENDRKL_METHOD_PREPEND is set to xyz_ (i.e., -DRKL_​METHOD_​PREPEND=xyz_-DRKL_METHOD_PREPEND=xyz_), it would cause clearStringCache to become xyz_clearStringCache.
RKL_​REGISTER_​FOR_​IPHONE_​LOWMEM_​NOTIFICATIONSRKL_REGISTER_FOR_IPHONE_LOWMEM_NOTIFICATIONSAutomaticThis preprocessor define controls whether or not extra code is included that attempts to automatically register with the NSNotificationCenter for the UIApplication​DidReceive​MemoryWarning​NotificationUIApplicationDidReceiveMemoryWarningNotification notification. This feature is automatically enabled if it can be determined at compile time that the iPhone is being targeted. This feature may be explicitly disabled under all circumstances by setting its value to 0.
RKL_​STACK_​LIMITRKL_STACK_LIMIT131072The maximum amount of stack space that will be used before switching to heap based allocations. This can be useful for multithreading programs where the stack size of secondary threads is much smaller than the main thread.
See Also

Fast Mutable Checks

Setting RKL_FAST_MUTABLE_CHECK allows RegexKitLite to quickly check if a string to search is immutable or not. Every call to RegexKitLite requires checking a strings hash and length values to guard against a string mutating and using invalid cached data. If the same string is searched repeatedly and it is immutable, these checks aren't necessary since the string can never change while in use. While these checks are fairly quick, it can add approximately 15 to 20 percent of extra overhead, and not performing the checks is always faster.

Since checking a strings mutability requires calling an undocumented, private Core Foundation function, RegexKitLite takes extra precautions and does not use the function directly. Instead, an internal, local stub function is created and called to determine if a string is mutable. The first time this function is called, RegexKitLite uses dlsym() to look up the address of the __CFStringIsMutable() function. If the function is found, RegexKitLite will use it from that point on to determine if a string is immutable. However, if the function is not found, RegexKitLite has no way to determine if a string is mutable or not, so it assumes the worst case that all strings are potentially mutable. This means that the private Core Foundation __CFStringIsMutable() function can go away at any time and RegexKitLite will continue to work, although with slightly less performance.

This feature is disabled by default, but should be fairly safe to enable due to the extra precautions that are taken. If this feature is enabled and the __CFStringIsMutable() function is not found for some reason, RegexKitLite falls back to its default behavior which is the same as if this feature was not enabled.

iPhone Low Memory Notifications

The RKL_REGISTER_FOR_IPHONE_LOWMEM_NOTIFICATIONS preprocessor define controls whether or not extra code is compiled in that automatically registers for the iPhone UIKit UIApplicationDidReceiveMemoryWarningNotification notification. When enabled, an initialization function tagged with __attribute__((constructor)) is executed by the linker at load time which causes RegexKitLite to check if the low memory notification symbol is available. If the symbol is present then RegexKitLite registers to receive the notification. When the notification is received, RegexKitLite will automatically call clearStringCache to flush the caches and return the memory used to hold any cached compiled regular expressions.

This feature is normally automatically enabled if it can be determined at compile time that the iPhone is being targeted. This feature is safe to enable even if the target is Mac OS X for the desktop. It can also be explicitly disabled, even when targeting the iPhone, by setting RKL_REGISTER_FOR_IPHONE_LOWMEM_NOTIFICATIONS to 0 (i.e., -DRKL_REGISTER_FOR_IPHONE_LOWMEM_NOTIFICATIONS=0).

See Also

Using RegexKitLite with a Custom ICU Build

The details of building and linking to a custom build of ICU will not be covered here. ICU is a very large and complex library that can be configured and packaged in countless ways. Building and linking your application to a custom build of ICU is non‑trivial. Apple provides the full source to the version of ICU that they supply with Mac OS X. At the time of this writing, the latest version available was for Mac OS X 10.6.2— ICU-400.38.tar.gz.

RegexKitLite provides the RKL_APPEND_TO_ICU_FUNCTIONS pre-processor define if you would like to use RegexKitLite with a custom ICU build that you supply. A custom version of ICU will typically have the ICU version appended to all of its functions, and RKL_APPEND_TO_ICU_FUNCTIONS allows you to append that version to the ICU functions that RegexKitLite calls. For example, passing -DRKL_APPEND_TO_ICU_FUNCTIONS=_4_0 to gcc would cause the ICU function uregex_find() used by RegexKitLite to be called as uregex_find_4_0().

Xcode 3 Integrated Documentation

This documentation is available in the Xcode DocSet format at the following URL:

feed://regexkit.sourceforge.net/RegexKitLiteDocSets.atom

For Xcode < 3.2, select Help Documentation. Then, in the lower left hand corner of the documentation window, there should be a gear icon with a drop down menu indicator which you should select and choose New Subscription… and enter the DocSet URL.

For Xcode ≥ 3.2, select Xcode Preferences…. Then select the Documentation preference group, typically the right most group, and press Add Documentation Set Publisher… and enter the DocSet URL.

Once you have added the URL, a new group should appear, inside which will be the RegexKitLite documentation with a Get button. Click on the Get button and follow the prompts.

Note:

Xcode will ask you to enter an administrators password to install the documentation, which is explained here.

Cached Information and Mutable Strings

While RegexKitLite takes steps to ensure that the information it has cached is valid for the strings it searches, there exists the possibility that out of date cached information may be used when searching mutable strings. For each compiled regular expression, RegexKitLite caches the following information about the last NSString that was searched:

An ICU compiled regular expression must be "set" to the text to be searched. Before a compiled regular expression is used, the pointer to the string object to search, its hash, length, and the pointer to the UTF-16 buffer is compared with the values that the compiled regular expression was last "set" to. If any of these values are different, the compiled regular expression is reset and "set" to the new string.

If a NSMutableString is mutated between two uses of the same compiled regular expression and its hash, length, or UTF-16 buffer changes between uses, RegexKitLite will automatically reset the compiled regular expression with the new values of the mutated string. The results returned will correctly reflect the mutations that have taken place between searches.

It is possible that the mutations to a string can go undetected, however. If the mutation keeps the length the same, then the only way a change can be detected is if the strings hash value changes. For most mutations the hash value will change, but it is possible for two different strings to share the same hash. This is known as a hash collision. Should this happen, the results returned by RegexKitLite may not be correct.

Therefore, if you are using RegexKitLite to search NSMutableString objects, and those strings may have mutated in such a way that RegexKitLite is unable to detect that the string has changed, you must manually clear the internal cache to ensure that the results accurately reflect the mutations. To clear the cached information for a specific string you send the instance a flushCachedRegexData message:

NSMutableString *aMutableString; // Assumed to be valid. [aMutableString flushCachedRegexData];

To clear all of the cached information in RegexKitLite, which includes all the cached compiled regular expressions along with any cached information and UTF-16 conversions for strings that have been searched, you use the following class method:

[NSString clearStringCache];
Warning:

When searching NSMutableString objects that have mutated between searches, failure to clear the cache may result in undefined behavior. Use flushCachedRegexData to selectively clear the cached information about a NSMutableString object.

See Also

Block-based Enumeration Methods

The RegexKitLite Block-based enumeration methods are modeled after their NSString counterparts. There are a few differences, however.

While RegexKitLite may not support mutating a NSMutableString during Block-based enumeration, it does provide the means to create a new string from the NSString object returned by the block used to enumerate the matches in a string, and in the case of NSMutableString, to replace the contents of that NSMutableString with the modified string at the end of the enumeration. This functionality is available via the following methods:

Exception to the Cocoa Memory Management Rules for Block-based Enumeration

The standard Cocoa Memory Management Rules specify that objects returned by a method, or in this case the objects passed to a Block, remain valid throughout the scope of the calling method. Due to the potentially large volume of temporary strings that are created during a Block-based enumeration, RegexKitLite makes an exception to this rule– the strings passed to a Block via capturedStrings[] are valid only until the closing brace of the Block:

[searchString enumerateStringsMatchedByRegex:regex usingBlock: ^(NSInteger captureCount, NSString * const capturedStrings, const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { // Block code. } /* <- capturedStrings[] is valid only up to this point. */ ];

If you need to refer to a string past the closing brace of the Block, you need to send that string a retain message. Of course, it is not always necessary to explicitly send a capturedStrings[] string a retain message when you need it to exist past the closing brace of a Block– adding a capturedStrings[] string to a NSMutableDictionary will send the string a retain as a side effect of adding it to the dictionary.

Memory management during RegexKitLite Block-based enumeration is conceptually similar to the following pseudo-code:

NSInteger captureCount = [regex captureCount]; while(moreMatches) { NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init]; BOOL stop = NO; NSRange capturedRanges[captureCount]; NSString *capturedStrings[captureCount]; for(capture = 0L; capture < captureCount; capture++) { capturedRanges[capture] = [searchString rangeOfRegex:regex capture:capture]; capturedStrings[capture] = [searchString stringByMatching:regex capture:capture]; } // The following line represents the execution of the user supplied Block. enumerationBlock(captureCount, capturedStrings, capturedRanges, &stop); // ... and this represents when the user supplied Block has finished / returned. [pool release]; // capturedStrings[] are sent a release at this point. if(stop != NO) { break; } }

While conceptually and behaviorally similar, it is important to note that RegexKitLite does not actually use or create autorelease pools when performing Block-based enumeration. Instead, a CFMutableArray object is used to accumulate the temporary string objects during an iteration, and at the start of an iteration, any previously accumulated temporary string objects are removed from the array.

See Also

Usage Notes

Convenience Methods

For convenience methods where an argument is not present, the default value used is given below.

ArgumentDefault Value
capture:0
options:RKLNoOptions
range:The entire range of the receiver.
enumerationOptions:RKLRegexEnumerationNoOptions

Exceptions Raised

Methods will raise an exception if their arguments are invalid, such as passing NULL for a required parameter. An invalid regular expression or RKLRegexOptions parameter will not raise an exception. Instead, a NSError object with information about the error will be created and returned via the address given with the optional error argument. If information about the problem is not required, error may be NULL. For convenience methods that do not have an error argument, the primary method is invoked with NULL passed as the argument for error.

Important:
Methods raise NSInvalidArgumentException if regex is NULL, or if capture < 0 or is not valid for regex.
Important:
Methods raise NSRangeException if range exceeds the bounds of the receiver.
Important:
Methods raise NSRangeException if the receivers length exceeds the maximum value that can be represented by a signed 32-bit integer, even on 64-bit architectures.
Important:
Search and replace methods raise RKLICURegexException if replacement contains $n capture references where n is greater than the number of capture groups in the regular expression regex.
See Also

Tasks

DTrace Probe Points
Clearing Cached Information
Determining the Number of Captures
Finding all Captures of all Matches
Getting all the Captures of a Match
Finding all Matches
Dividing Strings
Identifying Matches
Determining if a Regular Expression is Valid
Determining the Range of a Match
Modifying Mutable Strings
Creating Temporary Strings from a Match
Replacing Substrings
Creating a Dictionary from a Match
Creating a Dictionary from every Match

DTrace Probe Points

RegexKitLite:::compiledRegexCache
This probe point fires each time the compiled regular expression cache is accessed.
RegexKitLite:::compiledRegexCache(unsigned long eventID, const char *regexUTF8, int options, int captures, int hitMiss, int icuStatusCode, const char *icuErrorMessage, double *hitRate);
Arguments
  • arg0, eventID
    The unique ID for this mutex lock acquisition.
  • arg1, regexUTF8
    Up to 64 characters of the regular expression encoded in UTF-8. Must be copied with copyinstr(arg1).
  • arg2, options
    The RKLRegexOptions options used.
  • arg3, captures
    The number of captures present in the regular expression, or -1 if there was an error.
  • arg4, hitMiss
    A boolean value that indicates whether or not this event was a cache hit or not, or -1 if there was an error.
  • arg5, icuStatusCode
    If an error occurs, this contains the error number returned by ICU.
  • arg6, icuErrorMessage
    If an error occurs, this contains a UTF-8 encoded string of the ICU error. Must be copied with copyinstr(arg6).
  • arg7, hitRate
    A pointer to a floating point value, between 0.0 and 100.0, that represents the effectiveness of cache. Higher is better. Must be copied with copyin(arg7, sizeof(double)).
Discussion

An example of how to copy the double value pointed to by hitRate:

RegexKitLite*:::compiledRegexCache { this->hitRate = (double *)copyin(arg7, sizeof(double)); printf("compiledRegexCache hitRate: %6.2f%%\n", this->hitRate); }
RegexKitLite:::utf16ConversionCache
This probe point fires each time the UTF-16 conversion cache is accessed.
RegexKitLite:::utf16ConversionCache(unsigned long eventID, unsigned int lookupResultFlags, double *hitRate, const void *string, unsigned long NSRange_location, unsigned long NSRange_length, long length);
Arguments
  • arg0, eventID
    The unique ID for this mutex lock acquisition.
  • arg1, lookupResultFlags
    A set of status flags about the result of the conversion cache lookup.
  • arg2, hitRate
    A pointer to a floating point value, between 0.0 and 100.0, that represents the effectiveness of cache. Higher is better. Must be copied with copyin(arg2, sizeof(double)).
  • arg3, string
    A pointer to the NSString that this UTF-16 conversion cache check is being performed on.
  • arg4, NSRange_location
    The location value of the range argument from the invoking RegexKitLite method.
  • arg5, NSRange_length
    The length value of the range argument from the invoking RegexKitLite method.
  • arg6, length
    The length of the string.
Discussion

Only strings that require a UTF-16 conversion count towards the value calculated for hitRate.

An example of how to copy the double value pointed to by hitRate:

RegexKitLite*:::utf16ConversionCache { this->hitRate = (double *)copyin(arg2, sizeof(double)); printf("utf16ConversionCache hitRate: %6.2f%%\n", this->hitRate); }

Class Methods

captureCountForRegex:
Returns the number of captures that regex contains. Deprecated in RegexKitLite 3.0. Use captureCount instead.
+ (NSInteger)captureCountForRegex:(NSString *)regex;
Discussion

Since the capture count of a regular expression does not depend on the string to be searched, this is a NSString class method. For example:

NSInteger regexCaptureCount = [NSString captureCountForRegex:@"(\\d+)\.(\\d+)"]; // regexCaptureCount would be set to 2.
Availability

Deprecated in RegexKitLite 3.0

Return Value

Returns -1 if an error occurs. Otherwise the number of captures in regex is returned, or 0 if regex does not contain any captures.

captureCountForRegex:options:error:
Returns the number of captures that regex contains. Deprecated in RegexKitLite 3.0. Use captureCountWithOptions:error: instead.
+ (NSInteger)captureCountForRegex:(NSString *)regex options:(RKLRegexOptions)options error:(NSError **)error;
Discussion

The optional error parameter, if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.

Since the capture count of a regular expression does not depend on the string to be searched, this is a NSString class method. For example:

NSInteger regexCaptureCount = [NSString captureCountForRegex:@"(\\d+)\.(\\d+)" options:RKLNoOptions error:NULL]; // regexCaptureCount would be set to 2.
Availability

Deprecated in RegexKitLite 3.0

Return Value

Returns -1 if an error occurs. Otherwise the number of captures in regex is returned, or 0 if regex does not contain any captures.

clearStringCache
Clears the cached information about strings and cached compiled regular expressions.
+ (void)clearStringCache;
Discussion

This method clears all the the cached state maintained by RegexKitLite. This includes all the cached compiled regular expressions and any cached UTF-16 conversions.

An example of clearing the cache:

[NSString clearStringCache]; // Clears all RegexKitLite cache state.
Warning:

When searching NSMutableString objects that have mutated between searches, failure to clear the cache may result in undefined behavior. Use flushCachedRegexData to selectively clear the cached information about a NSMutableString object.

Note:

You do not need to call clearStringCache or flushCachedRegexData when using the NSMutableString replaceOccurrencesOfRegex:withString: methods. The cache entry for that regular expression and NSMutableString is automatically cleared as necessary.

Availability

Available in RegexKitLite 1.1 and later.

Instance Methods

arrayOfCaptureComponentsMatchedByRegex:
Returns an array containing all the matches from the receiver that were matched by the regular expression regex. Each match result consists of an array of the substrings matched by all the capture groups present in the regular expression.
- (NSArray *)arrayOfCaptureComponentsMatchedByRegex:(NSString *)regex;
Return Value

A NSArray object containing all the matches from the receiver by regex. Each match result consists of a NSArray which contains all the capture groups present in regex. Array index 0 represents all of the text matched by regex and subsequent array indexes contain the text matched by their respective capture group.

Discussion

A match result array index will contain an empty string, or @"", if a capture group did not match any text.

Availability

Available in RegexKitLite 3.0 and later.

arrayOfCaptureComponentsMatchedByRegex:range:
Returns an array containing all the matches from the receiver that were matched by the regular expression regex within range. Each match result consists of an array of the substrings matched by all the capture groups present in the regular expression.
- (NSArray *)arrayOfCaptureComponentsMatchedByRegex:(NSString *)regex range:(NSRange)range;
Return Value

A NSArray object containing all the matches from the receiver by regex. Each match result consists of a NSArray which contains all the capture groups present in regex. Array index 0 represents all of the text matched by regex and subsequent array indexes contain the text matched by their respective capture group.

Discussion

A match result array index will contain an empty string, or @"", if a capture group did not match any text.

Availability

Available in RegexKitLite 3.0 and later.

arrayOfCaptureComponentsMatchedByRegex:options:range:error:
Returns an array containing all the matches from the receiver that were matched by the regular expression regex within range using options. Each match result consists of an array of the substrings matched by all the capture groups present in the regular expression.
- (NSArray *)arrayOfCaptureComponentsMatchedByRegex:(NSString *)regex options:(RKLRegexOptions)options range:(NSRange)range error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Return Value

A NSArray object containing all the matches from the receiver by regex. Each match result consists of a NSArray which contains all the capture groups present in regex. Array index 0 represents all of the text matched by regex and subsequent array indexes contain the text matched by their respective capture group.

Discussion

If the receiver is not matched by regex then the returned value is a NSArray that contains no items.

A match result array index will contain an empty string, or @"", if a capture group did not match any text.

The match results in the array appear in the order they did in the receiver. For example, this code fragment:

NSString *list = @"$10.23, $1024.42, $3099"; NSArray *listItems = [list arrayOfCaptureComponentsMatchedByRegex:@"\\$((\\d+)(?:\\.(\\d+)|\\.?))"];

produces a NSArray equivalent to:

[NSArray arrayWithObjects: [NSArray arrayWithObjects: @"$10.23", @"10.23", @"10", @"23", NULL], // Index 0 [NSArray arrayWithObjects:@"$1024.42", @"1024.42", @"1024", @"42", NULL], // Index 1 [NSArray arrayWithObjects: @"$3099", @"3099", @"3099", @"", NULL], // Index 2 NULL];
Availability

Available in RegexKitLite 3.0 and later.

arrayOfDictionariesByMatchingRegex:withKeysAndCaptures:
Returns an array containing all the matches in the receiver that were matched by the regular expression regex. Each match result consists of a dictionary containing that matches substrings constructed from the specified set of keys and captures.
- (NSArray *)arrayOfDictionariesByMatchingRegex:(NSString *)regex withKeysAndCaptures:(id)firstKey, ...;
Parameters
  • regex
    A NSString containing a regular expression.
  • firstKey
    The first key to add to the new dictionary.
  • ...
    First the capture for firstKey, then a NULL-terminated list of alternating keys and captures. Captures are specified using int values.
    Important:

    Use of non-int sized capture arguments will result in undefined behavior. Do not append capture arguments with a L suffix.

    Important:

    Failure to NULL-terminate the keys and captures list will result in undefined behavior.

Return Value
A NSArray object containing all the matches from the receiver by regex. Each match result consists of a NSDictionary containing that matches substrings constructed from the specified set of keys and captures.
Discussion

If the receiver is not matched by regex then the returned value is a NSArray that contains no items.

A dictionary will not contain a given key if its corresponding capture group did not match any text.

Availability

Available in RegexKitLite 4.0 and later.

arrayOfDictionariesByMatchingRegex:range:withKeysAndCaptures:
Returns an array containing all the matches in the receiver that were matched by the regular expression regex within range. Each match result consists of a dictionary containing that matches substrings constructed from the specified set of keys and captures.
- (NSArray *)arrayOfDictionariesByMatchingRegex:(NSString *)regex range:(NSRange)range withKeysAndCaptures:(id)firstKey, ...;
Parameters
  • regex
    A NSString containing a regular expression.
  • range
    The range of the receiver to search.
  • firstKey
    The first key to add to the new dictionary.
  • ...
    First the capture for firstKey, then a NULL-terminated list of alternating keys and captures. Captures are specified using int values.
    Important:

    Use of non-int sized capture arguments will result in undefined behavior. Do not append capture arguments with a L suffix.

    Important:

    Failure to NULL-terminate the keys and captures list will result in undefined behavior.

Return Value
A NSArray object containing all the matches from the receiver by regex. Each match result consists of a NSDictionary containing that matches substrings constructed from the specified set of keys and captures.
Discussion

If the receiver is not matched by regex then the returned value is a NSArray that contains no items.

A dictionary will not contain a given key if its corresponding capture group did not match any text.

Availability

Available in RegexKitLite 4.0 and later.

arrayOfDictionariesByMatchingRegex:options:range:error:withKeysAndCaptures:
Returns an array containing all the matches in the receiver that were matched by the regular expression regex within range using options. Each match result consists of a dictionary containing that matches substrings constructed from the specified set of keys and captures.
- (NSArray *)arrayOfDictionariesByMatchingRegex:(NSString *)regex options:(RKLRegexOptions)options range:(NSRange)range error:(NSError **)error withKeysAndCaptures:(id)firstKey, ...;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
  • firstKey
    The first key to add to the new dictionary.
  • ...
    First the capture for firstKey, then a NULL-terminated list of alternating keys and captures. Captures are specified using int values.
    Important:

    Use of non-int sized capture arguments will result in undefined behavior. Do not append capture arguments with a L suffix.

    Important:

    Failure to NULL-terminate the keys and captures list will result in undefined behavior.

Return Value
A NSArray object containing all the matches from the receiver by regex. Each match result consists of a NSDictionary containing that matches substrings constructed from the specified set of keys and captures.
Discussion

If the receiver is not matched by regex then the returned value is a NSArray that contains no items.

A dictionary will not contain a given key if its corresponding capture group did not match any text. It is important to note that a regular expression can successfully match zero characters:

NSString *name = @"Name: Bob\n" @"Name: John Smith"; NSString *regex = @"(?m)^Name:\\s*(\\w*)\\s*(\\w*)$"; NSArray *nameArray = [name arrayOfDictionariesByMatchingRegex:regex options:RKLNoOptions range:NSMakeRange(0UL,) error:NULL withKeysAndCaptures:@"first", 1, @"last", 2, NULL]; // 2010-04-16 01:15:30.061 RegexKitLite[42984:a0f] nameArray: ( // { first = Bob, last = "" }, // { first = John, last = Smith } // )

Compared to this example, where the second capture group does not match any characters:

NSString *name = @"Name: Bob\n" @"Name: John Smith"; NSString *regex = @"(?m)^Name:\\s*(\\w*)(?:\\s*|\\s+(\\w+))$"; NSArray *nameArray = [name arrayOfDictionariesByMatchingRegex:regex options:RKLNoOptions range:NSMakeRange(0UL,) error:NULL withKeysAndCaptures:@"first", 1, @"last", 2, NULL]; // 2010-04-16 01:15:30.061 RegexKitLite[42984:a0f] nameArray: ( // { first = Bob }, // { first = John, last = Smith } // )
Availability

Available in RegexKitLite 4.0 and later.

captureComponentsMatchedByRegex:
Returns an array containing the substrings matched by each capture group present in regex for the first match of regex in the receiver.
- (NSArray *)captureComponentsMatchedByRegex:(NSString *)regex;
Return Value
A NSArray containing the substrings matched by each capture group present in regex for the first match of regex in the receiver. Array index 0 represents all of the text matched by regex and subsequent array indexes contain the text matched by their respective capture group.
Discussion

A match result array index will contain an empty string, or @"", if a capture group did not match any text.

Availability

Available in RegexKitLite 3.0 and later.

captureComponentsMatchedByRegex:range:
Returns an array containing the substrings matched by each capture group present in regex for the first match of regex within range of the receiver.
- (NSArray *)captureComponentsMatchedByRegex:(NSString *)regex range:(NSRange)range;
Return Value
A NSArray containing the substrings matched by each capture group present in regex for the first match of regex within range of the receiver. Array index 0 represents all of the text matched by regex and subsequent array indexes contain the text matched by their respective capture group.
Discussion

A match result array index will contain an empty string, or @"", if a capture group did not match any text.

Availability

Available in RegexKitLite 3.0 and later.

captureComponentsMatchedByRegex:options:range:error:
Returns an array containing the substrings matched by each capture group present in regex for the first match of regex within range of the receiver using options.
- (NSArray *)captureComponentsMatchedByRegex:(NSString *)regex options:(RKLRegexOptions)options range:(NSRange)range error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Return Value
A NSArray containing the substrings matched by each capture group present in regex for the first match of regex within range of the receiver using options. Array index 0 represents all of the text matched by regex and subsequent array indexes contain the text matched by their respective capture group.
Discussion

If the receiver is not matched by regex then the returned value is a NSArray that contains no items.

A match result array index will contain an empty string, or @"", if a capture group did not match any text.

The returned value is for the first match of regex in the receiver. For example, this code fragment:

NSString *list = @"$10.23, $1024.42, $3099"; NSArray *listItems = [list captureComponentsMatchedByRegex:@"\\$((\\d+)(?:\\.(\\d+)|\\.?))"];

produces a NSArray equivalent to:

[NSArray arrayWithObjects: @"$10.23", @"10.23", @"10", @"23", NULL];
Availability

Available in RegexKitLite 3.0 and later.

captureCount
Returns the number of captures that regex contains.
- (NSInteger)captureCount;
Discussion

Returns the capture count of the receiver, which should be a valid regular expression. For example:

NSInteger regexCaptureCount = [@"(\\d+)\.(\\d+)" captureCount]; // regexCaptureCount would be set to 2.
Return Value

Returns -1 if an error occurs. Otherwise the number of captures in regex is returned, or 0 if regex does not contain any captures.

Availability

Available in RegexKitLite 3.0 and later.

captureCountWithOptions:error:
Returns the number of captures that regex contains.
- (NSInteger)captureCountWithOptions:(RKLRegexOptions)options error:(NSError **)error;
Discussion

The optional error parameter, if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.

Returns the capture count of the receiver, which should be a valid regular expression. For example:

NSInteger regexCaptureCount = [@"(\\d+)\.(\\d+)" captureCountWithOptions:RKLNoOptions error:NULL]; // regexCaptureCount would be set to 2.
Return Value

Returns -1 if an error occurs. Otherwise the number of captures in regex is returned, or 0 if regex does not contain any captures.

Availability

Available in RegexKitLite 3.0 and later.

componentsMatchedByRegex:
Returns an array containing all the substrings from the receiver that were matched by the regular expression regex.
- (NSArray *)componentsMatchedByRegex:(NSString *)regex;
Return Value
A NSArray object containing all the substrings from the receiver that were matched by regex.
Availability

Available in RegexKitLite 3.0 and later.

componentsMatchedByRegex:capture:
Returns an array containing all the substrings from the receiver that were matched by capture number capture from the regular expression regex.
- (NSArray *)componentsMatchedByRegex:(NSString *)regex capture:(NSInteger)capture;
Return Value
A NSArray object containing all the substrings for capture group capture from the receiver that were matched by regex.
Discussion

An array index will contain an empty string, or @"", if the capture group did not match any text.

Availability

Available in RegexKitLite 3.0 and later.

componentsMatchedByRegex:range:
Returns an array containing all the substrings from the receiver that were matched by the regular expression regex within range.
- (NSArray *)componentsMatchedByRegex:(NSString *)regex range:(NSRange)range;
Return Value
A NSArray object containing all the substrings from the receiver that were matched by regex within range.
Availability

Available in RegexKitLite 3.0 and later.

componentsMatchedByRegex:options:range:capture:error:
Returns an array containing all the substrings from the receiver that were matched by capture number capture from the regular expression regex within range using options.
- (NSArray *)componentsMatchedByRegex:(NSString *)regex options:(RKLRegexOptions)options range:(NSRange)range capture:(NSInteger)capture error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • capture
    The string matched by capture from regex to return. Use 0 for the entire string that regex matched.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Return Value
A NSArray object containing all the substrings from the receiver that were matched by capture number capture from regex within range using options.
Discussion

If the receiver is not matched by regex then the returned value is a NSArray that contains no items.

An array index will contain an empty string, or @"", if a capture group did not match any text.

The match results in the array appear in the order they did in the receiver.

Example:

NSString *list = @"$10.23, $1024.42, $3099"; NSArray *listItems = [list componentsMatchedByRegex:@"\\$((\\d+)(?:\\.(\\d+)|\\.?))"]; // listItems == [NSArray arrayWithObjects:@"$10.23", @"$1024.42", @"$3099", NULL];

Example of extracting a specific capture group:

NSString *list = @"$10.23, $1024.42, $3099"; NSRange listRange = NSMakeRange(0UL, [list length]); NSArray *listItems = [list componentsMatchedByRegex:@"\\$((\\d+)(?:\\.(\\d+)|\\.?))" options:RKLNoOptions range:listRange capture:3L error:NULL]; // listItems == [NSArray arrayWithObjects:@"23", @"42", @"", NULL];
Availability

Available in RegexKitLite 3.0 and later.

componentsSeparatedByRegex:
Returns an array containing substrings from the receiver that have been divided by the regular expression regex.
- (NSArray *)componentsSeparatedByRegex:(NSString *)regex;
Return Value
A NSArray object containing the substrings from the receiver that have been divided by regex.
Availability

Available in RegexKitLite 2.0 and later.

componentsSeparatedByRegex:range:
Returns an array containing substrings within range of the receiver that have been divided by the regular expression regex.
- (NSArray *)componentsSeparatedByRegex:(NSString *)regex range:(NSRange)range;
Return Value
A NSArray object containing the substrings from the receiver that have been divided by regex.
Availability

Available in RegexKitLite 2.0 and later.

componentsSeparatedByRegex:options:range:error:
Returns an array containing substrings within range of the receiver that have been divided by the regular expression regex using options.
- (NSArray *)componentsSeparatedByRegex:(NSString *)regex options:(RKLRegexOptions)options range:(NSRange)range error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Return Value
A NSArray object containing the substrings from the receiver that have been divided by regex.
Discussion

The substrings in the array appear in the order they did in the receiver. For example, this code fragment:

NSString *list = @"Norman, Stanley, Fletcher"; NSArray *listItems = [list componentsSeparatedByRegex:@",\\s*"];

produces an array { @"Norman", @"Stanley", @"Fletcher" }.

If the receiver begins or ends with regex, then the first or last substring is, respectively, empty. For example, the string ", Norman, Stanley, Fletcher" creates an array that has these contents: { @"", @"Norman", @"Stanley", @"Fletcher" }.

If the receiver has no separators that are matched by regex—for example, "Norman"—the array contains the string itself, in this case { @"Norman" }.

If regex contains capture groups—for example, @",(\\s*)"—the array will contain the text matched by each capture group as a separate element appended to the normal result. An additional element will be created for each capture group. If an individual capture group does not match any text the result in the array will be a zero length string—@"". As an example—the regular expression @",(\\s*)" would produce the array { @"Norman", @" ", @"Stanley", @" ", @"Fletcher" }.

Availability

Available in RegexKitLite 2.0 and later.

dictionaryByMatchingRegex:withKeysAndCaptures:
Creates and returns a dictionary containing the matches constructed from the specified set of keys and captures for the first match of regex in the receiver.
- (NSDictionary *)dictionaryByMatchingRegex:(NSString *)regex withKeysAndCaptures:(id)firstKey, ...;
Parameters
  • regex
    A NSString containing a regular expression.
  • firstKey
    The first key to add to the new dictionary.
  • ...
    First the capture for firstKey, then a NULL-terminated list of alternating keys and captures. Captures are specified using int values.
    Important:

    Use of non-int sized capture arguments will result in undefined behavior. Do not append capture arguments with a L suffix.

    Important:

    Failure to NULL-terminate the keys and captures list will result in undefined behavior.

Return Value
A NSDictionary containing the matched substrings constructed from the specified set of keys and captures.
Discussion

The returned value is for the first match of regex in the receiver.

If the receiver is not matched by regex then the returned value is a NSDictionary that contains no items.

A dictionary will not contain a given key if its corresponding capture group did not match any text.

Availability

Available in RegexKitLite 4.0 and later.

dictionaryByMatchingRegex:range:withKeysAndCaptures:
Creates and returns a dictionary containing the matches constructed from the specified set of keys and captures for the first match of regex within range of the receiver.
- (NSDictionary *)dictionaryByMatchingRegex:(NSString *)regex range:(NSRange)range withKeysAndCaptures:(id)firstKey, ...;
Parameters
  • regex
    A NSString containing a regular expression.
  • range
    The range of the receiver to search.
  • firstKey
    The first key to add to the new dictionary.
  • ...
    First the capture for firstKey, then a NULL-terminated list of alternating keys and captures. Captures are specified using int values.
    Important:

    Use of non-int sized capture arguments will result in undefined behavior. Do not append capture arguments with a L suffix.

    Important:

    Failure to NULL-terminate the keys and captures list will result in undefined behavior.

Return Value
A NSDictionary containing the matched substrings constructed from the specified set of keys and captures.
Discussion

The returned value is for the first match of regex in the receiver.

If the receiver is not matched by regex then the returned value is a NSDictionary that contains no items.

A dictionary will not contain a given key if its corresponding capture group did not match any text.

Availability

Available in RegexKitLite 4.0 and later.

dictionaryByMatchingRegex:options:range:error:withKeysAndCaptures:
Creates and returns a dictionary containing the matches constructed from the specified set of keys and captures for the first match of regex within range of the receiver using options.
- (NSDictionary *)dictionaryByMatchingRegex:(NSString *)regex options:(RKLRegexOptions)options range:(NSRange)range error:(NSError **)error withKeysAndCaptures:(id)firstKey, ...;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
  • firstKey
    The first key to add to the new dictionary.
  • ...
    First the capture for firstKey, then a NULL-terminated list of alternating keys and captures. Captures are specified using int values.
    Important:

    Use of non-int sized capture arguments will result in undefined behavior. Do not append capture arguments with a L suffix.

    Important:

    Failure to NULL-terminate the keys and captures list will result in undefined behavior.

Return Value
A NSDictionary containing the matched substrings constructed from the specified set of keys and captures.
Discussion

The returned value is for the first match of regex in the receiver.

If the receiver is not matched by regex then the returned value is a NSDictionary that contains no items.

A dictionary will not contain a given key if its corresponding capture group did not match any text. It is important to note that a regular expression can successfully match zero characters:

NSString *name = @"Name: Joe"; NSString *regex = @"Name:\\s*(\\w*)\\s*(\\w*)"; NSDictionary *nameDictionary = [name dictionaryByMatchingRegex:regex options:RKLNoOptions range:NSMakeRange(0UL,) error:NULL withKeysAndCaptures:@"first", 1, @"last", 2, NULL]; // 2010-01-29 12:19:54.559 RegexKitLite[64944:a0f] nameDictionary: { // first = Joe; // last = ""; // }

Compared to this example, where the second capture group does not match any characters:

NSString *name = @"Name: Joe"; NSString *regex = @"Name:\\s*(\\w*)\\s*(\\w+)?"; NSDictionary *nameDictionary = [name dictionaryByMatchingRegex:regex options:RKLNoOptions range:NSMakeRange(0UL, [name length]) error:NULL withKeysAndCaptures:@"first", 1, @"last", 2, NULL]; // 2010-01-29 12:12:52.177 RegexKitLite[64893:a0f] nameDictionary: { // first = Joe; // }
Availability

Available in RegexKitLite 4.0 and later.

enumerateStringsMatchedByRegex:usingBlock:
Enumerates the matches in the receiver by the regular expression regex and executes block for each match found.
- (BOOL)enumerateStringsMatchedByRegex:(NSString *)regex usingBlock:(void (^)(NSInteger captureCount,
                         NSString * const capturedStrings[captureCount],
                         const NSRange capturedRanges[captureCount],
                         volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Return Value
Returns YES if there was no error, otherwise returns NO.
Availability

Available in RegexKitLite 4.0 and later.

enumerateStringsMatchedByRegex:options:inRange:error:enumerationOptions:usingBlock:
Enumerates the matches in the receiver by the regular expression regex within range using options and executes block using enumerationOptions for each match found.
- (BOOL)enumerateStringsMatchedByRegex:(NSString *)regex options:(RKLRegexOptions)options inRange:(NSRange)range error:(NSError **)error enumerationOptions:(RKLRegexEnumerationOptions)enumerationOptions usingBlock:(void (^)(NSInteger captureCount,
                         NSString * const capturedStrings[captureCount],
                         const NSRange capturedRanges[captureCount],
                         volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
  • enumerationOptions
    A mask of options specified by combining RKLRegexEnumerationOptions flags with the C bitwise OR operator. Either 0 or RKLRegexEnumerationNoOptions may be used if no options are required.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Return Value
Returns YES if there was no error, otherwise returns NO and indirectly returns a NSError object if error is not NULL.
Availability

Available in RegexKitLite 4.0 and later.

enumerateStringsSeparatedByRegex:usingBlock:
Enumerates the strings of the receiver that have been divided by the regular expression regex and executes block for each divided string.
- (BOOL)enumerateStringsSeparatedByRegex:(NSString *)regex usingBlock:(void (^)(NSInteger captureCount,
                         NSString * const capturedStrings[captureCount],
                         const NSRange capturedRanges[captureCount],
                         volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Return Value
Returns YES if there was no error, otherwise returns NO.
Availability

Available in RegexKitLite 4.0 and later.

enumerateStringsSeparatedByRegex:options:inRange:error:enumerationOptions:usingBlock:
Enumerates the strings of the receiver that have been divided by the regular expression regex within range using options and executes block using enumerationOptions for each divided string.
- (BOOL)enumerateStringsSeparatedByRegex:(NSString *)regex options:(RKLRegexOptions)options inRange:(NSRange)range error:(NSError **)error enumerationOptions:(RKLRegexEnumerationOptions)enumerationOptions usingBlock:(void (^)(NSInteger captureCount,
                         NSString * const capturedStrings[captureCount],
                         const NSRange capturedRanges[captureCount],
                         volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
  • enumerationOptions
    A mask of options specified by combining RKLRegexEnumerationOptions flags with the C bitwise OR operator. Either 0 or RKLRegexEnumerationNoOptions may be used if no options are required.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Return Value
Returns YES if there was no error, otherwise returns NO and indirectly returns a NSError object if error is not NULL.
Availability

Available in RegexKitLite 4.0 and later.

flushCachedRegexData
Clears any cached information about the receiver.
- (void)flushCachedRegexData;
Discussion

This method should be used when performing searches on NSMutableString objects and there is the possibility that the string has mutated in between calls to RegexKitLite.

This method clears the cached information for the receiver only. This is more selective than clearStringCache, which clears all the cached information from RegexKitLite, including all the cached compiled regular expressions.

RegexKitLite automatically detects the vast majority of string mutations and clears any cached information for the mutated string. To detect mutations, RegexKitLite records a strings length and hash value at the point in time when it caches data for a string. Cached data for a string is invalidated if either of these values change between calls to RegexKitLite. The problem case is when a string is mutated but the strings length remains the same and the hash value for the mutated string is identical to the hash value of the string before it was mutated. This is known as a hash collision. Since RegexKitLite is unable to detect that a string has mutated when this happens, the programmer needs to explicitly inform RegexKitLite that any cached data about the receiver needs to be cleared by sending flushCachedRegexData to the mutated string.

While it is possible to have "perfect mutation detection", and therefore guarantee that only valid cached data is used, it has a significant performance penalty. The first problem is that when caching information about a string, an immutable copy of that string needs to be made. The second problem is that determining that two strings are not identical is usually very fast and cheap— if their lengths are not the same, no further checks are required. The most expensive case is when two strings are identical because it requires a character by character comparison of the entire string to guarantee that they are equal. The most expensive case also happens to be the most common case, by far. To make matters worst, Cocoa provides no public way to determine if an instance is a mutable NSMutableString or an immutable NSString object. Therefore RegexKitLite must assume the worst case that all strings are mutable and have potentially mutated between calls to RegexKitLite.

RegexKitLite is optimized for the common case which is when regular expression operations are performed on strings that are not mutating. The majority of mutations to a string can be quickly and cheaply detected by RegexKitLite automatically. Since the programmer has the context of the string that is to be matched, and whether or not the string is being mutated, RegexKitLite relies on the programmer to inform it whether or not the possibility exists that the string could have mutated in a way that is undetectable.

An example of clearing a strings cached information:

NSMutableString *mutableSearchString; // Assumed to be valid. NSString *foundString = [mutableSearchString stringByMatching:@"\\d+"]; // Searched.. [mutableSearchString replaceCharactersInRange:NSMakeRange(5UL, 10UL) withString:@"[replaced]"]; // Mutated.. [mutableSearchString flushCachedRegexData]; // Clear cached information about mutableSearchString.
Warning:

Failure to clear the cached information for a NSMutableString object that has mutated between searches may result in undefined behavior.

Note:

You do not need to call clearStringCache or flushCachedRegexData when using the NSMutableString replaceOccurrencesOfRegex:withString: methods. The cached information for that NSMutableString is automatically cleared as necessary.

Availability

Available in RegexKitLite 3.0 and later.

isMatchedByRegex:
Returns a Boolean value that indicates whether the receiver is matched by regex.
- (BOOL)isMatchedByRegex:(NSString *)regex;
Availability

Available in RegexKitLite 1.0 and later.

isMatchedByRegex:inRange:
Returns a Boolean value that indicates whether the receiver is matched by regex within range.
- (BOOL)isMatchedByRegex:(NSString *)regex inRange:(NSRange)range;
Availability

Available in RegexKitLite 1.0 and later.

isMatchedByRegex:options:inRange:error:
Returns a Boolean value that indicates whether the receiver is matched by regex within range.
- (BOOL)isMatchedByRegex:(NSString *)regex options:(RKLRegexOptions)options inRange:(NSRange)range error:(NSError **)error;
Discussion

The optional error parameter, if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.

Availability

Available in RegexKitLite 1.0 and later.

isRegexValid
Returns a Boolean value that indicates whether the regular expression contained in the receiver is valid.
- (BOOL)isRegexValid;
Availability

Available in RegexKitLite 3.0 and later.

isRegexValidWithOptions:error:
Returns a Boolean value that indicates whether the regular expression contained in the receiver is valid using options.
- (BOOL)isRegexValidWithOptions:(RKLRegexOptions)options error:(NSError **)error;
Parameters
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Discussion

This method can be used to determine if a regular expression is valid. For example:

NSError *error = NULL; NSString *regexString = @"[a-z"; // Missing the closing ] if([regexString isRegexValidWithOptions:RKLNoOptions error:&error] == NO) { NSLog(@"The regular expression is invalid. Error: %@", error); }
Availability

Available in RegexKitLite 3.0 and later.

rangeOfRegex:
Returns the range for the first match of regex in the receiver.
- (NSRange)rangeOfRegex:(NSString *)regex;
Return Value

A NSRange structure giving the location and length of the first match of regex in the receiver. Returns {NSNotFound, 0} if the receiver is not matched by regex or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

rangeOfRegex:capture:
Returns the range of capture number capture for the first match of regex in the receiver.
- (NSRange)rangeOfRegex:(NSString *)regex capture:(NSInteger)capture;
Return Value

A NSRange structure giving the location and length of capture number capture for the first match of regex in the receiver. Returns {NSNotFound, 0} if the receiver is not matched by regex or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

rangeOfRegex:inRange:
Returns the range for the first match of regex within range of the receiver.
- (NSRange)rangeOfRegex:(NSString *)regex inRange:(NSRange)range;
Return Value

A NSRange structure giving the location and length of the first match of regex within range of the receiver. Returns {NSNotFound, 0} if the receiver is not matched by regex within range or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

rangeOfRegex:options:inRange:capture:error:
Returns the range of capture number capture for the first match of regex within range of the receiver.
- (NSRange)rangeOfRegex:(NSString *)regex options:(RKLRegexOptions)options inRange:(NSRange)range capture:(NSInteger)capture error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • capture
    The matching range of the capture number from regex to return. Use 0 for the entire range that regex matched.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Return Value

A NSRange structure giving the location and length of capture number capture for the first match of regex within range of the receiver. Returns {NSNotFound, 0} if the receiver is not matched by regex within range or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

replaceOccurrencesOfRegex:usingBlock:
Enumerates the matches in the receiver by the regular expression regex and executes block for each match found. Replaces the characters that were matched with the contents of the string returned by block, returning the number of replacements made.
- (NSInteger)replaceOccurrencesOfRegex:(NSString *)regex usingBlock:(NSString *(^)(NSInteger captureCount,
                              NSString * const capturedStrings[captureCount],
                              const NSRange capturedRanges[captureCount],
                              volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Discussion

This method modifies the receivers contents. An exception will be raised if it is sent to an immutable object.

Return Value
Returns -1 if there was an error, otherwise returns the number of replacements performed.
Availability

Available in RegexKitLite 4.0 and later.

replaceOccurrencesOfRegex:options:inRange:error:enumerationOptions:usingBlock:
Enumerates the matches in the receiver by the regular expression regex within range using options and executes block using enumerationOptions for each match found. Replaces the characters that were matched with the contents of the string returned by block, returning the number of replacements made.
- (NSInteger)replaceOccurrencesOfRegex:(NSString *)regex options:(RKLRegexOptions)options inRange:(NSRange)range error:(NSError **)error enumerationOptions:(RKLRegexEnumerationOptions)enumerationOptions usingBlock:(NSString *(^)(NSInteger captureCount,
                              NSString * const capturedStrings[captureCount],
                              const NSRange capturedRanges[captureCount],
                              volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
  • enumerationOptions
    A mask of options specified by combining RKLRegexEnumerationOptions flags with the C bitwise OR operator. Either 0 or RKLRegexEnumerationNoOptions may be used if no options are required.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Discussion

This method modifies the receivers contents. An exception will be raised if it is sent to an immutable object.

Return Value
Returns -1 if there was an error and indirectly returns a NSError object if error is not NULL, otherwise returns the number of replacements performed.
Availability

Available in RegexKitLite 4.0 and later.

replaceOccurrencesOfRegex:withString:
Replaces all occurrences of the regular expression regex with the contents of replacement string after performing capture group substitutions, returning the number of replacements made.
- (NSInteger)replaceOccurrencesOfRegex:(NSString *)regex withString:(NSString *)replacement;
Important:
Raises RKLICURegexException if replacement contains $n capture references where n is greater than the number of capture groups in the regular expression regex.
Discussion

This method modifies the receivers contents. An exception will be raised if it is sent to an immutable object.

Return Value
Returns -1 if there was an error, otherwise returns the number of replacements performed.
Availability

Available in RegexKitLite 2.0 and later.

replaceOccurrencesOfRegex:withString:range:
Replaces all occurrences of the regular expression regex within range with the contents of replacement string after performing capture group substitutions, returning the number of replacements made.
- (NSInteger)replaceOccurrencesOfRegex:(NSString *)regex withString:(NSString *)replacement range:(NSRange)range;
Important:
Raises RKLICURegexException if replacement contains $n capture references where n is greater than the number of capture groups in the regular expression regex.
Discussion

This method modifies the receivers contents. An exception will be raised if it is sent to an immutable object.

Return Value
Returns -1 if there was an error, otherwise returns the number of replacements performed.
Availability

Available in RegexKitLite 2.0 and later.

replaceOccurrencesOfRegex:withString:options:range:error:
Replaces all occurrences of the regular expression regex using options within range with the contents of replacement string after performing capture group substitutions, returning the number of replacements made.
- (NSInteger)replaceOccurrencesOfRegex:(NSString *)regex options:(RKLRegexOptions)options withString:(NSString *)replacement range:(NSRange)range error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • replacement
    The string to use as the replacement text for matches by regex. See ICU Replacement Text Syntax for more information.
    Important:
    Raises RKLICURegexException if replacement contains $n capture references where n is greater than the number of capture groups in the regular expression regex.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Discussion

This method modifies the receivers contents. An exception will be raised if it is sent to an immutable object.

Return Value
Returns -1 if there was an error and indirectly returns a NSError object if error is not NULL, otherwise returns the number of replacements performed.
Availability

Available in RegexKitLite 2.0 and later.

stringByMatching:
Returns a string created from the characters of the receiver that are in the range of the first match of regex.
- (NSString *)stringByMatching:(NSString *)regex;
Return Value

A NSString containing the substring of the receiver matched by regex. Returns NULL if the receiver is not matched by regex or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

stringByMatching:capture:
Returns a string created from the characters of the receiver that are in the range of the first match of regex for capture.
- (NSString *)stringByMatching:(NSString *)regex capture:(NSInteger)capture;
Return Value

A NSString containing the substring of the receiver matched by capture number capture of regex. Returns NULL if the receiver is not matched by regex or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

stringByMatching:inRange:
Returns a string created from the characters of the receiver that are in the range of the first match of regex within range of the receiver.
- (NSString *)stringByMatching:(NSString *)regex inRange:(NSRange)range;
Return Value

A NSString containing the substring of the receiver matched by regex within range of the receiver. Returns NULL if the receiver is not matched by regex within range or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

stringByMatching:options:inRange:capture:error:
Returns a string created from the characters of the receiver that are in the range of the first match of regex using options within range of the receiver for capture.
- (NSString *)stringByMatching:(NSString *)regex options:(RKLRegexOptions)options inRange:(NSRange)range capture:(NSInteger)capture error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • capture
    The string matched by capture from regex to return. Use 0 for the entire string that regex matched.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Return Value

A NSString containing the substring of the receiver matched by capture number capture of regex within range of the receiver. Returns NULL if the receiver is not matched by regex within range or an error occurs.

Availability

Available in RegexKitLite 1.0 and later.

stringByReplacingOccurrencesOfRegex:usingBlock:
Enumerates the matches in the receiver by the regular expression regex and executes block for each match found. Returns a string created by replacing the characters that were matched in the receiver with the contents of the string returned by block.
- (NSString *)stringByReplacingOccurrencesOfRegex:(NSString *)regex usingBlock:(NSString *(^)(NSInteger captureCount,
                              NSString * const capturedStrings[captureCount],
                              const NSRange capturedRanges[captureCount],
                              volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Return Value

A NSString created from the characters of the receiver in which all matches of the regular expression regex are replaced with the contents of the NSString returned by block. If the receiver is not matched by regex then the string that is returned is a copy of the receiver as if stringWithString: had been sent to it.

Returns NULL if there was an error.

Availability

Available in RegexKitLite 4.0 and later.

stringByReplacingOccurrencesOfRegex:options:inRange:error:enumerationOptions:usingBlock:
Enumerates the matches in the receiver by the regular expression regex within range using options and executes block using enumerationOptions for each match found. Returns a string created by replacing the characters that were matched in the receiver with the contents of the string returned by block.
- (NSString *)stringByReplacingOccurrencesOfRegex:(NSString *)regex options:(RKLRegexOptions)options inRange:(NSRange)range error:(NSError **)error enumerationOptions:(RKLRegexEnumerationOptions)enumerationOptions usingBlock:(NSString *(^)(NSInteger captureCount,
                              NSString * const capturedStrings[captureCount],
                              const NSRange capturedRanges[captureCount],
                              volatile BOOL * const stop))block;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
  • enumerationOptions
    A mask of options specified by combining RKLRegexEnumerationOptions flags with the C bitwise OR operator. Either 0 or RKLRegexEnumerationNoOptions may be used if no options are required.
  • block
    The block that is executed for each match of regex in the receiver. The block takes four arguments:
    • captureCount
      The number of strings that regex captured. captureCount is always at least 1.
    • capturedStrings
      An array containing the substrings matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a pointer to a string that is equal to @"". This argument may be NULL if enumerationOptions had RKLRegexEnumerationCapturedStringsNotRequired set.
    • capturedRanges
      An array containing the ranges matched by each capture group present in regex. The size of the array is captureCount. If a capture group did not match anything, it will contain a NSRange equal to {NSNotFound, 0}.
    • stop
      A reference to a BOOL value that the block can use to stop the enumeration by setting *stop = YES;, otherwise it should not touch *stop.
Return Value

A NSString created from the characters within range of the receiver in which all matches of the regular expression regex using options are replaced with the contents of the NSString returned by block. Returns the characters within range as if substringWithRange: had been sent to the receiver if the substring is not matched by regex.

Returns NULL if there was an error and indirectly returns a NSError object if error is not NULL.

Availability

Available in RegexKitLite 4.0 and later.

stringByReplacingOccurrencesOfRegex:withString:
Returns a string created from the characters of the receiver in which all matches of the regular expression regex are replaced with the contents of the replacement string after performing capture group substitutions.
- (NSString *)stringByReplacingOccurrencesOfRegex:(NSString *)regex withString:(NSString *)replacement;
Important:
Raises RKLICURegexException if replacement contains $n capture references where n is greater than the number of capture groups in the regular expression regex.
Return Value

A NSString created from the characters of the receiver in which all matches of the regular expression regex are replaced with the contents of the replacement string after performing capture group substitutions. If the receiver is not matched by regex then the string that is returned is a copy of the receiver as if stringWithString: had been sent to it.

Availability

Available in RegexKitLite 2.0 and later.

stringByReplacingOccurrencesOfRegex:withString:range:
Returns a string created from the characters within range of the receiver in which all matches of the regular expression regex are replaced with the contents of the replacement string after performing capture group substitutions.
- (NSString *)stringByReplacingOccurrencesOfRegex:(NSString *)regex withString:(NSString *)replacement range:(NSRange)range;
Important:
Raises RKLICURegexException if replacement contains $n capture references where n is greater than the number of capture groups in the regular expression regex.
Return Value

A NSString created from the characters within range of the receiver in which all matches of the regular expression regex are replaced with the contents of the replacement string after performing capture group substitutions. Returns the characters within range as if substringWithRange: had been sent to the receiver if the substring is not matched by regex.

Availability

Available in RegexKitLite 2.0 and later.

stringByReplacingOccurrencesOfRegex:​withString:options:range:error:
Returns a string created from the characters within range of the receiver in which all matches of the regular expression regex using options are replaced with the contents of the replacement string after performing capture group substitutions.
- (NSString *)stringByReplacingOccurrencesOfRegex:(NSString *)regex options:(RKLRegexOptions)options withString:(NSString *)replacement range:(NSRange)range error:(NSError **)error;
Parameters
  • regex
    A NSString containing a regular expression.
  • options
    A mask of options specified by combining RKLRegexOptions flags with the C bitwise OR operator. Either 0 or RKLNoOptions may be used if no options are required.
  • replacement
    The string to use as the replacement text for matches by regex. See ICU Replacement Text Syntax for more information.
    Important:
    Raises RKLICURegexException if replacement contains $n capture references where n is greater than the number of capture groups in the regular expression regex.
  • range
    The range of the receiver to search.
  • error
    An optional parameter that if set and an error occurs, will contain a NSError object that describes the problem. This may be set to NULL if information about any errors is not required.
Return Value

A NSString created from the characters within range of the receiver in which all matches of the regular expression regex using options are replaced with the contents of the replacement string after performing capture group substitutions. Returns the characters within range as if substringWithRange: had been sent to the receiver if the substring is not matched by regex.

Availability

Available in RegexKitLite 2.0 and later.

Constants

RKLRegexOptions
Type for regular expression options.
typedef uint32_t RKLRegexOptions;
Discussion

See Regular Expression Options for possible values.

Availability

Available in RegexKitLite 1.0 and later.

Declared In
RegexKitLite.h
Regular Expression Options
The following flags control various aspects of regular expression matching. The flag values may be specified at the time that a regular expression is used, or they may be specified within the pattern itself using the (?ismwx-ismwx) pattern options.
Constants
RKLNoOptions
No regular expression options specified.
Available in RegexKitLite 1.0 and later.
RKLCaseless
If set, matching will take place in a case-insensitive manner.
Available in RegexKitLite 1.0 and later.
RKLComments
If set, allow use of white space and #comments within patterns.
Available in RegexKitLite 1.0 and later.
RKLDotAll
If set, a . in a pattern will match a line terminator in the input text. By default, it will not. Note that a carriage-return / line-feed pair in text behave as a single line terminator, and will match a single . (period) in a regular expression pattern.
Available in RegexKitLite 1.0 and later.
RKLMultiline
Control the behavior of ^ and $ in a pattern. By default these will only match at the start and end, respectively, of the input text. If this flag is set, ^ and $ will also match at the start and end of each line within the input text.
Available in RegexKitLite 1.0 and later.
RKLUnicodeWordBoundaries
Controls the behavior of \b in a pattern. If set, word boundaries are found according to the definitions of word found in Unicode UAX 29 - Text Boundaries. By default, word boundaries are identified by means of a simple classification of characters as either word or non-word, which approximates traditional regular expression behavior. The results obtained with the two options can be quite different in runs of spaces and other non-word characters.
Available in RegexKitLite 1.0 and later.
Discussion

Options for controlling the behavior of a regular expression pattern can be controlled in two ways. When the method supports it, options may specified by combining RKLRegexOptions flags with the C bitwise OR operator. For example:

matchedRange = [aString rangeOfRegex:@"^(blue|red)$" options:(RKLCaseless | RKLMultiline) inRange:range error:NULL];

The other way is to specify the options within the regular expression itself, of which there are two ways. The first specifies the options for everything following it, and the other sets the options on a per capture group basis. Options are either enabled, or following a -, disabled. The syntax for both is nearly identical:

OptionExampleDescription
(?ixsmw-ixsmw)(?i)Enables the RKLCaseless option for everything that follows it. Useful at the beginning of a regular expression to set the desired options.
(?ixsmw-ixsmw:)(?iw-m:)Enables the RKLCaseless and RKLUnicodeWordBoundaries options and disables RKLMultiline for the capture group enclosed by the parenthesis.

The following table lists the regular expression pattern option character and its corresponding RKLRegexOptions flag:

CharacterOption
iRKLCaseless
xRKLComments
sRKLDotAll
mRKLMultiline
wRKLUnicodeWordBoundaries
Availability

Available in RegexKitLite 1.0 and later.

Declared In
RegexKitLite.h
RKLRegexEnumerationOptions
Type for regular expression enumeration options.
typedef NSUInteger RKLRegexEnumerationOptions;
Discussion

See Regular Expression Enumeration Options for possible values.

Important:
RKLRegexEnumerationOptions is only available if RKL_BLOCKS is set.
Availability

Available in RegexKitLite 4.0 and later.

Declared In
RegexKitLite.h
Regular Expression Enumeration Options
The following flags control various aspects of regular expression enumeration.
Constants
RKLRegexEnumerationNoOptions
No regular expression enumeration options specified.
Available in RegexKitLite 4.0 and later.
RKLRegexEnumerationCapturedStringsNotRequired

If set, this indicates that the block does not require the captured strings and NULL will be passed to the blocks capturedStrings[] argument.

Available in RegexKitLite 4.0 and later.
RKLRegexEnumerationReleaseStringReturnedByReplacementBlock
If set, RegexKitLite will send the string returned by the block a release message. If garbage collection is active, this flag does nothing. This flag is only valid for the stringByReplacingOccurrencesOfRegex: and replaceOccurrencesOfRegex: groups of methods. See Regular Expression Enumeration Options - Using RKLRegexEnumerationReleaseStringReturnedByReplacementBlock for more information.
Important:
Methods will raise NSInvalidArgumentException if RKLRegexEnumerationReleaseStringReturnedByReplacementBlock is set and the block enumeration method is not a 'Search and Replace' method.
Available in RegexKitLite 4.0 and later.
RKLRegexEnumerationFastCapturedStringsXXX

If set, RegexKitLite will create the captured strings that are passed to the blocks capturedStrings[] argument in a special way. The enumeration identifier ends in XXX to act as a visual reminder that special care must be taken. See Regular Expression Enumeration Options - Using RKLRegexEnumerationFastCapturedStringsXXX for more information.

Caution:

RKLRegexEnumerationFastCapturedStringsXXX is an advanced feature option. Do not use this option unless you fully understand the consequences. Incorrect usage of this option will result in undefined behavior and will probably result in your program crashing.

Available in RegexKitLite 4.0 and later.
Discussion
Important:
RKLRegexEnumerationOptions is only available if RKL_BLOCKS is set.

When the method supports it, enumeration options may specified by combining RKLRegexEnumerationOptions flags with the C bitwise OR operator.For example:

RKLRegexEnumerationOptions enumerationOptions = (RKLRegexEnumerationCapturedStringsNotRequired | RKLRegexEnumerationReleaseStringReturnedByReplacementBlock);

A typical use of stringByReplacingOccurrencesOfRegex: might be something like the following:

replacedString = [aString stringByReplacingOccurrencesOfRegex:@"\\w+" options:RKLNoOptions inRange:NSMakeRange(0UL, [aString length]) error:NULL enumerationOptions:RKLRegexEnumerationNoOptions usingBlock:^(NSInteger captureCount, NSString * const capturedStrings[captureCount], const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { return((NSString *)[NSString stringWithFormat:@"<%@>", capturedStrings[0]]); }];
Note:
The above example casts the value returned by stringWithFormat:, which has a return type of id, to NSString * because of a bug in gcc 4.2.1-5646. Without the cast, the compiler gives an error that the block does not return a compatible type. clang does not require a cast in order to compile.

An alternate way to work around the gcc bug is to explicitly declare the return type of the block. The example below shows how this is done with the changes highlighted:

stringByReplacingOccurrencesOfRegex:… usingBlock:^NSString *(NSInteger captureCount, NSString * const capturedStrings[captureCount], const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { return(/*(NSString *)*/[NSString stringWithFormat:@"<%@>", capturedStrings[0]]); }];
Using RKLRegexEnumerationReleaseStringReturnedByReplacementBlock

The use of convenience methods like stringWithFormat: can create a lot of temporary strings that are autoreleased. The memory used to hold these temporary strings won't be reclaimed until the autorelease pool that they are in is released. When performing a replace on a large string, this can end up being a considerable about of memory. If this is becomes a problem in your application, the RKLRegexEnumerationReleaseStringReturnedByReplacementBlock enumeration option can be used to reclaim the memory used by these temporary strings immediately. RKLRegexEnumerationReleaseStringReturnedByReplacementBlock essentially transfers ownership of the string that is returned by the block to RegexKitLite, which then becomes responsible for sending the string a release message when it is no longer needed. Typically this is right after RegexKitLite has appended the contents of the returned string to the temporary internal string of all the accumulated replacement strings returned by the block. For example:

replacedString = [aString stringByReplacingOccurrencesOfRegex:@"\\w+" options:RKLNoOptions inRange:NSMakeRange(0UL, [aString length]) error:NULL enumerationOptions:RKLRegexEnumerationReleaseStringReturnedByReplacementBlock usingBlock:^NSString *(NSInteger captureCount, NSString * const capturedStrings[captureCount], const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { return([[NSString alloc] initWithFormat:@"<%@>", capturedStrings[0]]); }];

When using RKLRegexEnumerationReleaseStringReturnedByReplacementBlock, it is also possible to directly return the strings that were passed via capturedStrings[]. RegexKitLite automatically checks if the pointer of the string returned by the block is equal to the pointer of any of the strings it created and passed via capturedStrings[], in which case RegexKitLite does not send the string that was returned a release message since this would result in over releasing the string object. For example:

replacedString = [aString stringByReplacingOccurrencesOfRegex:@"\\w+" options:RKLNoOptions inRange:NSMakeRange(0UL, [aString length]) error:NULL enumerationOptions:RKLRegexEnumerationReleaseStringReturnedByReplacementBlock usingBlock:^NSString *(NSInteger captureCount, NSString * const capturedStrings[captureCount], const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { if([capturedStrings[0] isEqual:@"with"]) { return(capturedStrings[0]); } else { return([[NSString alloc] initWithFormat:@"<%@>", capturedStrings[0]]); } }];
Using RKLRegexEnumerationFastCapturedStringsXXX

Normally, RegexKitLite instantiates a new string object for each string passed via capturedStrings[]. When RKLRegexEnumerationFastCapturedStringsXXX is set, RegexKitLite creates up to captureCount strings using the Core Foundation function CFStringCreateMutableWithExternalCharactersNoCopy and passes these strings via capturedStrings[].

Caution:

RKLRegexEnumerationFastCapturedStringsXXX is an advanced feature option. Do not use this option unless you fully understand the consequences. Incorrect usage of this option will result in undefined behavior and will probably result in your program crashing.

The reason for doing this is strings created with CFStringCreateMutableWithExternalCharactersNoCopy can have the strings buffer set directly with CFStringSetExternalCharactersNoCopy. This allows RegexKitLite to quickly update the strings that were captured by a match and is much faster than instantiating new strings for each match iteration since the only thing that needs to be updated is the captured strings pointer and length. In fact, the speed of RKLRegexEnumerationFastCapturedStringsXXX is nearly has fast as RKLRegexEnumerationCapturedStringsNotRequired.

Special care must be taken when using the strings passed via capturedStrings[] when RKLRegexEnumerationFastCapturedStringsXXX has been set. The following rules must be followed:

  • The lifetime of the string is the duration of the blocks execution. The string can not be used or referenced past the closing brace of the block.
  • Do not mutate any of capturedStrings[].
  • Do not retain or release any of the capturedStrings[].
  • If you need the string to exist past the end of the block, a -copy of the string must be made:

    NSString *copiedString = [[capturedStrings[0] copy] autorelease];

Example usages:

  • When using capturedStrings[] as an argument to stringWithFormat:, you do not need to make a copy of the captured string:

    [searchString stringByReplacingOccurrencesOfRegex:regex usingBlock:^NSString *(NSInteger captureCount, NSString * const capturedStrings[captureCount], const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { NSString *replacedString = NULL; replacedString = [NSString stringWithFormat:@"{1: '%@'}", capturedStrings[1]]; // OK to use directly. return(replacedString); }];
  • Adding to a collection:

    [aDictionary setObject:capturedStrings[0] forKey:@"aKey"]; // Do not add a captured string directly to a collection. [aDictionary setObject:[[capturedStrings[0] copy] autorelease] forKey:@"aKey"]; // Do add a copy of the captured string to a collection.
  • Assigning a captured string to a variable whose scope extends past the blocks:

    __block NSString *foundString = NULL; [searchString enumerateStringsMatchedByRegex:regex usingBlock:^(NSInteger captureCount, NSString * const capturedStrings[captureCount], const NSRange capturedRanges[captureCount], volatile BOOL * const stop) { foundString = capturedStrings[0]; // Do not assign a captured string to a variable whose scope extends past the blocks. foundString = [[capturedStrings[0] copy] autorelease]; // Do assign a copy of the captured string to a variable whose scope extends past the blocks. }];
Availability

Available in RegexKitLite 4.0 and later.

Declared In
RegexKitLite.h
RegexKitLite:::utf16ConversionCache arg1 Flags
The following flags are used to indicate the status of the result from arg1 lookupResultFlags of the RegexKitLite:::utf16ConversionCache DTrace probe point.
Constants
RKLCacheHitLookupFlag
If set, there was a successful lookup hit. This flag is only set if RKLConversionRequiredLookupFlag is also set.
Available in RegexKitLite 3.0 and later.
RKLConversionRequiredLookupFlag
If set, direct access to a strings UTF-16 buffer was not available and a conversion was required.
Available in RegexKitLite 3.0 and later.
RKLSetTextLookupFlag
If set, the ICU uregex_setText() function was called in order to set the compiled regular expression to a buffer.
Available in RegexKitLite 3.0 and later.
RKLDynamicBufferLookupFlag
If set, the strings size was large enough to require the use of the dynamically sized conversion buffer.
Available in RegexKitLite 3.0 and later.
RKLErrorLookupFlag
If set, there was some kind of error during the UTF-16 conversion process.
Available in RegexKitLite 3.0 and later.
RKLEnumerationBufferLookupFlag
If set, the probe was fired due to a blocks enumeration. The conversion result is one time use only and is only available to the block enumeration that required it.
Available in RegexKitLite 4.0 and later.

RegexKitLite NSError Error Domains

The following NSError error domains are defined.
Constants
RKLICURegexErrorDomain
ICU Regular Expression Errors.
Availability

Available in RegexKitLite 1.0 and later.

Declared In
RegexKitLite.h

RegexKitLite NSException Exception Names

The following NSException exception names are defined.
Constants
RKLICURegexException
ICU Regular Expression Exceptions
Discussion

Returns a user info dictionary populated with keys as defined in RegexKitLite NSError and NSException User Info Dictionary Keys.

Availability

Available in RegexKitLite 2.0 and later.

Declared In
RegexKitLite.h

RegexKitLite NSError and NSException User Info Dictionary Keys

Constants
RKLICURegexEnumerationOptionsErrorKey
The RKLRegexEnumerationOptions enumeration options specified.
Available in RegexKitLite 4.0 and later.
RKLICURegexErrorCodeErrorKey
The error code returned by the ICU library function calls status argument.
Available in RegexKitLite 2.0 and later.
RKLICURegexErrorNameErrorKey
The string returned by invoking u_errorName with the error code RKLICURegexErrorCodeErrorKey. Example: U_REGEX_RULE_SYNTAX.
Available in RegexKitLite 1.0 and later.
RKLICURegexLineErrorKey
The line number where the error occurred in the regular expression.
Available in RegexKitLite 1.0 and later.
RKLICURegexOffsetErrorKey
The offset from the beginning of the line where the error occurred in the regular expression.
Available in RegexKitLite 1.0 and later.
RKLICURegexPreContextErrorKey
Up to 16 characters leading up to the cause error in the regular expression.
Available in RegexKitLite 1.0 and later.
RKLICURegexPostContextErrorKey
Up to 16 characters after the cause error in the regular expression.
Available in RegexKitLite 1.0 and later.
RKLICURegexRegexErrorKey
The regular expression that caused the error.
Available in RegexKitLite 1.0 and later.
RKLICURegexRegexOptionsErrorKey
The RKLRegexOptions regular expression options specified.
Available in RegexKitLite 1.0 and later.
RKLICURegexReplacedCountErrorKey
The number of times a replacement was performed.
Available in RegexKitLite 4.0 and later.
RKLICURegexReplacedStringErrorKey
The replaced string result after performing the search and replace operations.
Available in RegexKitLite 4.0 and later.
RKLICURegexReplacementStringErrorKey
The replacement template string used to perform search and replace operations.
Available in RegexKitLite 4.0 and later.
RKLICURegexSubjectRangeErrorKey
The range of the subject string that was being matched by the regular expression.
Available in RegexKitLite 4.0 and later.
RKLICURegexSubjectStringErrorKey
The string that was being matched by the regular expression.
Available in RegexKitLite 4.0 and later.
Discussion

Different error keys will be present depending on the type of error encountered.

Declared In
RegexKitLite.h

Release Information

Initial release.

Changes:

Bug fixes:

  • Fixed a bug that for strings that required UTF-16 conversion, the conversion from the previous string that required conversion may have been re-used for the current string, even though the two strings are different and the new string requires conversion.
  • Updated the internal inconsistency exception macro to correctly handle non-ASCII file names.

Changes:

  • Updated and clarified the documentation regarding adding RegexKitLite to an Xcode project.
  • Created Xcode 3 DocSet documentation.

New features:

  • The ability to split a string in to a NSArray with a regular expression.
  • Search and replace using common $n capture substitution in the replacement string.
  • Support for Leopards Garbage Collection

Changes:

  • A new compile time configuration option, RKL_FAST_MUTABLE_CHECK, enables the use of a private, non-public Core Foundation function to determine if the string that will be searched is immutable. This allows many of the mutation safety checks to be skipped for that string and can significantly improve the performance of repeated matches on the same immutable NSString. For safety, this option is disabled by default.
  • Gives a warning at compile time if the minimum Mac OS X target is < 10.3, which is the first version that shipped with the ICU library.
  • Compile time tunable setting of how much of the stack can be used before switching to heap based storage. Changed by setting RKL_STACK_LIMIT to the maximum number of bytes to use from the stack. Defaults to 131072, or 128Kbytes. By default in Cocoa the main thread has a stack size of 8Mbytes and additional threads have a stack size of 512Kbytes.
  • New NSException name for RegexKitLite specific exceptions, RKLICURegexException.
  • Garbage Collection is safe to use in "mixed-mode" Garbage Collection applications. When compiled with Garbage Collection support (-fobjc-gc), RegexKitLite will dynamically select either Garbage Collection or manual memory management at run-time depending on whether or not Garbage Collection is enabled and active.

New NSString Methods:

New NSMutableString Methods:

Bug fixes:

  • Fixed a bug dealing with the UTF-16 conversion cache. A bug could cause the text in the conversion buffer to be incorrectly used for the wrong string.

This release contains several large documentation additions and a few bug fixes. No new major functionality was added.

Documentation Changes:

  • Added a section covering ICU Regular Expression Character Classes
  • Added a section covering Unicode Properties, which are specified in a regular expression with \p{Unicode Property Name} and \P{Unicode Property Name}.
  • Added the RegexKitLite Cookbook section. Special Copy to Clipboard functionality is enabled for Safari users that allows a regular expression in this document that is selected and Copied to the Clipboard to be automatically escaped so the regular expression can be pasted directly in to your source code, ready to be used, with no additional work required.
  • Added the Epilogue section. This section is meant to be less formal, where I can place bits and pieces that don't fit elsewhere or editorialize on some topic.
  • A number of small additions, deletions, and edits through out the documentation.

Changes:

  • A new compile time configuration option, RKL_METHOD_PREPEND, can be used to add custom prefix to all the RegexKitLite methods. For example, setting RKL_METHOD_PREPEND to xyz_ would cause stringByReplacingOccurrencesOfRegex:withString: to become xyz_stringByReplacingOccurrencesOfRegex:withString:. This is for users who prefer that any category additions to the core classes, such as NSString, have a user defined prefix to prevent the possibility of user defined method names interfering with later system defined method names.
  • Support for iPhone low memory notifications. When it can be determined at compile time that the iPhone is being targeted, the preprocessor flag RKL_REGISTER_FOR_IPHONE_LOWMEM_NOTIFICATIONS is enabled to add code that automatically registers for low memory notifications on the iPhone. When a low memory notification is received, RegexKitLite will flush all its caches, freeing the memory used to hold the cached data.

Bug fixes:

  • Bug 2105213

    Fixed a bug in stringByReplacingOccurrencesOfRegex:withString: and replaceOccurrencesOfRegex:withString: where if the receiver was an empty string (i.e., @""), then RegexKitLite would throw an exception because the ICU library reported an error. This turned out to be a bug in the ICU library itself in the uregex_reset() function (and the methods it calls to perform its work). A bug was opened with the ICU project: http://bugs.icu-project.org/trac/ticket/6545. A work-around for the buggy behavior was put in place so that if the ICU library reports a U_INDEX_OUTOFBOUNDS_ERROR error for a string with a length of zero, that error is ignored since it is spurious. Thanks go to Andy Kim for reporting this.

  • Bug 2050825

    Fixed a bug with NSScannedOption when targeting the iPhone. NSScannedOption is not available on the iPhone, so C Pre-Processor statements were added to ensure that NSScannedOption is not referenced when Objective-C Garbage Collection is not enabled. Thanks go to Shaun Inman for reporting this first.

  • Fixed a bug in stringByReplacingOccurrencesOfRegex:withString: and replaceOccurrencesOfRegex:withString: where an internal assertion exception would be thrown if the string to search and the replacement string were both zero length strings (i.e., @"").

This release of RegexKitLite is a major release that includes new features, new APIs, and a number of bug fixes.

New Features:

  • DTrace support. This release includes RegexKitLite specific DTrace provider probe points that allow you to monitor the effectiveness of RegexKitLite's caches. DTrace support can be enabled with the compile time configuration option RKL_DTRACE.
  • Easy support for custom built ICU libraries. A new compile time configuration option, RKL_APPEND_TO_ICU_FUNCTIONS, can be used to append a string to the ICU functions called by RegexKitLite. Custom builds of the ICU library include the ICU version appended to all functions by default and this feature allows you to easily retarget RegexKitLite to your custom build. Example usage: -DRKL_APPEND_TO_ICU_FUNCTIONS=_4_0

New NSString Methods:

Deprecated Functionality:

Changes that effect the results that are returned:

  • When the regular expression matched the tail end of the string to search, componentsSeparatedByRegex: would incorrectly add a zero length string to the results returned.

    NSArray *splitArray = [@"Bob, Tom, Sam," componentsSeparatedByRegex:@",\\s*"]; // { @"Bob", @"Tom", @"Sam", @"" } <- RegexKitLite ≤ v2.2. // { @"Bob", @"Tom", @"Sam" } <- RegexKitLite ≥ v3.0.
  • The results returned by componentsSeparatedByRegex: were found to differ from the expected results when zero-width assertions, such as \b, were used. Prior to v3.0, regular expressions that matched zero characters would be considered a 'match', and a zero length string would be added to the results array. The expected results are the results that are returned by the perl split() function, which does not create a zero length string in such cases. ICU ticket #6826.

    NSArray *splitArray = [@"I|at|ice I eat rice" componentsSeparatedByRegex:@"\\b\\s*"]; // ICU : { @"", @"I", @"|", @"at", @"|", @"ice", @"", @"I", @"", @"eat", @"", @"rice" } <- RegexKitLite ≤ v2.2. // perl: { @"I", @"|", @"at", @"|", @"ice", @"I", @"eat", @"rice" } <- RegexKitLite ≥ v3.0.
  • As part of the 64-bit tidy, a check of the length of a string that is passed to RegexKitLite, and therefore ICU, was added to ensure that the length is less than INT_MAX. If the length of the string is ≥ INT_MAX, then RegexKitLite will raise NSRangeException. The value of INT_MAX is 231-1 (0x7fffffff). This was done because ICU uses the int type for representing offset values.

Other Changes:

  • Ticket #2027975 - Request for captureCount like functionality.
  • Ticket #2779301 - Request for componentsMatchedByRegex: functionality.
  • Ticket #2779965 - Request that documentation be updated with how to match a literal \ with a regex specified using a string literal: @"\\\\";.
  • Ticket #2786878 - Request for isRegexValid functionality.
  • The GCC variable cleanup __attribute__() is now used to provide an extra safety net around the use of the RegexKitLite spin lock. The cleanup function ensures that if the spin lock was locked by a function that it was also unlocked by the function. If a function obtains the lock but does not unlock it, the cleanup function forcibly unlocks the spin lock.
  • Minor GC changes. Some minor changes in the way heap buffers were allocated was required to support the new methods that return a NSArray.
  • 64-bit cleanup. Various literal numeric constant values had either L or UL appended to them if they were used with NSInteger, NSUInteger, or various other 'long' data types.
  • 64-bit cleanup. Added checks to verify that a strings length is < INT_MAX and throw an exception if it isn't. This check was added because ICU uses signed 32-bit int values to represent offsets.
  • Changed some macros to static inline functions for compile-time prototype checking.
  • Updated examples/NSString-HexConversion.m. 64-bit tidies.
  • Updated examples/RKLMatchEnumerator.h. Added the deprecated attribute to matchEnumeratorWithRegex:.
  • Updated examples/RKLMatchEnumerator.m. 64-bit tidies. Added a preprocessor warning that RKLMatchEnumerator has been deprecated in favor of componentsSeparatedByRegex:.
  • Updated the documentations visual style to better match the style currently used by Apple.
  • Many small DocSet tweaks and improvements.

Bug fixes:

  • Bug #2319200 - The documentation for stringByReplacingOccurrencesOfRegex:withString: (and related methods) was updated. The text for the Returned Value section was obviously copy and pasted from somewhere else and never updated. This has been fixed.
  • Bug #2408447 - stringByReplacingOccurrencesOfRegex: bug fixed. There was a bug in ICU that would cause a search and replace operation to fail with U_BUFFER_OVERFLOW_ERROR, which would then cause RegexKitLite to throw RKLICURegexException, if the length of the replaced string was significantly longer than the original string. ICU ticket #6656
  • componentsSeparatedByRegex: bug fixed. When the regular expression matched the tail end of the string, an extra zero length string was incorrectly added to the results.
  • componentsSeparatedByRegex: bug fixed. The results returned by the ICU function uregex_split() was found to be different that the results returned by the perl split() function. ICU ticket #6826
  • rangeOfRegex, stringByMatching: bug fixed. A bug was fixed where the inRange: parameter was not correctly honored.
  • If RegexKitLite was able to get direct access to a strings UTF-16 buffer, there was a very remote chance that RegexKitLite would continue to use a pointer to a strings older UTF-16 buffer if the string mutated and allocated a new buffer. The pointer to a strings direct buffer is now verified before each use.
  • isMatchedByRegex:inRange: was never documented. Fixed.

This release of RegexKitLite is a bug fix release.

Bug Fixes:

  • Bug #2790480 - If a regular expression had an error, as detected by ICU, then under some circumstances the lastCacheSlot variable would not be properly cleared. As a result, this would cause the pointer to the compiled regular expression to be NULL. The NULL value would be caught by the run‑time assertion checks if they were not disabled via NS_BLOCK_ASSERTIONS, otherwise it would most likely lead to a crash.

This release of RegexKitLite is a bug fix release.

Bug Fixes:

  • Bug #2828966 - Fixed a minor bug that caused compiling on Mac OS X 10.4 to fail. Changed the header include line from #include <objc/runtime.h>, which is available only on Mac OS X ≥ 10.5, to #include <objc/objc-runtime.h>, which is available on Mac OS X ≥ 10.4.
  • Bug #2862398 - Modified the logic for determining if the cached UTF-16 conversion is valid for the string to be searched. Previously, RegexKitLite used a combination of the string to be searched hash and length to determine if the cached UTF-16 conversion was a match. For speed, NSString only uses a fixed number of characters to calculate the hash. This means that as the length of a string grows, the hash function becomes less effective at detecting differences between strings. The original intent behind the UTF-16 conversion cache logic was to be able to re-use the UTF-16 conversion for strings that were "the same" (i.e., hash equality), but were different instantiations (i.e., pointer inequality). On reflection, it was decided that the more likely real-world case would be strings that would share the same hash value, but actually be slightly different. RegexKitLite now requires that the NSString pointer, as well as the hash and length, be the same as the string used for the cached UTF-16 conversion.
  • Bug #2879356 - arrayOfCaptureComponentsMatchedByRegex:, captureComponentsMatchedByRegex:, componentsMatchedByRegex:, and componentsSeparatedByRegex: bug fixed. Originally reported by Jesse Grosjean. Duplicate bug #2890342. If a regex matched a part of the string that included the very end of the string, subsequent matches would incorrectly return that the regex did not match the string. Clearing the cache, either manually or because another regex match replaced the affected cache slot, would clear the condition.
  • Bug #2890810 - stringByReplacingOccurrencesOfRegex: bug fixed. If the size of the buffer needed to hold the replaced string was larger than the initial estimated buffer size, ICU should return an error of U_BUFFER_OVERFLOW_ERROR. However, due to a bug in ICU (see bug #2408447, ICU ticket #6656), this error was not reported correctly and a workaround was implemented. A rare corner case could cause ICU to return U_STRING_NOT_TERMINATED_WARNING during the time that RegexKitLite was calculating the size of the buffer needed to hold all of the replaced string. When this happened it would cause RegexKitLite to miss the U_BUFFER_OVERFLOW_ERROR error needed to resize the buffer.

This release of RegexKitLite is a bug fix release.

Bug Fixes:

  • Bug #2893824 - Fixed a minor bug that caused compiling with the iPhone device SDK to fail. This is related to the RegexKitLite 3.2 fix for bug #2828966. It turns out that the iPhone simulator SDK has the header file objc/objc-runtime.h, but the iPhone device SDK only has the header file objc/runtime.h. RegexKitLite will now conditionally use #include <objc/runtime.h> when targeting the iPhone or Mac OS X ≥ 10.5, and #include <objc/objc-runtime.h> when targeting Mac OS X ≤ 10.4.

This release of RegexKitLite is a major release that includes new features, new APIs, and bug fixes.

API Compatibility Changes:

  • The NSMutableString replaceOccurrencesOfRegex: methods now return a result of type NSInteger, whereas RegexKitLite < 4.0 returned a result of type NSUInteger. This change, while technically breaking API compatibility, is likely to be completely backwards compatible for most users. The change was made so that errors, such as an invalid regular expression, can be distinguished from a search and replace operation that replaced zero occurrences. An error condition that is not an exception now returns a value of -1. Prior to RegexKitLite 4.0, the only way to distinguish between a search and replace operation that replaced zero occurrences and an error that prevented a search and replace from taking place was to use a replaceOccurrencesOfRegex: method with an error: parameter.

Changes that effect the results that are returned:

  • The results returned by componentsSeparatedByRegex: were found to differ from the expected results for some regex patterns. For example, prior to v4.0, when the regular expression . (dot, match any character) was used to split a string, RegexKitLite returned an array of seven zero length strings. The expected results are the results that are returned by the perl split() function– the equivalent of a NSArray with zero items in it.

    NSArray *splitArray = [@"abc.def" componentsSeparatedByRegex:@"."]; // { @"", @"", @"", @"", @"", @"", @"" } <- RegexKitLite ≤ v3.3. // { } <- RegexKitLite ≥ v4.0 & perl.

New features:

  • Documentation now available in PDF format.

    The RegexKitLite documentation is now available in PDF format. The PDF version is ideal for making printed copies of the documentation.

  • Improved compiled regular expression cache.

    RegexKitLite versions < 4.0 used a direct mapped, or 1-way set associative, cache. RegexKitLite 4.0 uses a 4-way set associative cache. In general, a n-way set associative cache, where n > 1, has fewer misses than a direct mapped cache. Entries in a set are managed on a least recently used, or LRU, basis. This means that when a compiled regular expression is added to the cache, the entry in a set that is the least recently used is the one chosen to hold the new compiled regular expression.

    In benchmark tests, the old compiled regular expression cache required a dramatically larger cache size to achieve the same effectiveness (i.e., the likelihood that the regular expression was already in the cache) as the new compiled regular expression cache. Retrieving a cached compiled regular expression is 541.3 times faster than actually compiling the regular expression:

    RetrieveTimeRate (per second)
    Cached51 ns19,680,762
    Compile27,560 ns36,285

    Timing done a MacBook Pro 2.66GHz running Mac OS X 10.6.2.

  • Improved UTF-16 conversion cache.

    The UTF-16 conversion cache now uses the same 4-way set associative cache system as the compiled regular expression cache. For strings that required a UTF-16 conversion, RegexKitLite versions < 4.0 used two buffers for caching UTF-16 conversions— a small, fixed sized buffer for 'small' strings (≤ 2048 characters by default), and a dynamically sized buffer for 'large' strings. RegexKitLite 4.0 keeps the same fixed and dynamic size dichotomy, but now each type contains four buffers which are reused on a least recently used basis.

  • Blocks support.

    Although support for blocks was introduced with Mac OS X 10.6, you can still use blocks on Mac OS X 10.5 and iPhone ≥ 2.2 by using Plausible Blocks. This solution provides a compiler and runtime that has been back-ported from the Mac OS X 10.6 Snow Leopard sources.

    Note:
    RegexKitLite checks if the C preprocessor define NS_BLOCKS_AVAILABLE is set to 1 to automatically determine if blocks functionality should be enabled. Typically NS_BLOCKS_AVAILABLE is only set to 1 when using the standard developer tools and the minimum version of Mac OS X supported is set to 10.6. RegexKitLite blocks support can be explicitly enabled by setting RKL_BLOCKS to 1 (i.e., -DRKL_BLOCKS=1). The behavior is undefined if RKL_BLOCKS is set to 1 and the compiler does not support the blocks language extension or if the run-time does not support blocks.

New NSString Methods:

New Block-based NSString Methods:

New Block-based NSMutableString Methods:

Other changes:

  • Improved performance. Many small performance tweaks and optimizations were made to frequently used code paths.
  • The meaning of RKL_CACHE_SIZE changed. This compile-time preprocessor tunable now controls the number of "sets" in the compiled regular expression cache, and each "set" contains four entries. This means that the maximum number of potential compiled regular expressions that can be cached is RKL_CACHE_SIZE times four. Like previous versions, RKL_CACHE_SIZE should always be a prime number to maximize the use of the cache.
  • RKL_CACHE_SIZE changed to 13. The total number of compiled regular expressions that can be cached is 13 * 4, or 52. This is a 126% increase from previous versions of RegexKitLite which had a default value of 23 for RKL_CACHE_SIZE.
  • Added the flag RKLEnumerationBufferLookupFlag to the DTrace RegexKitLite:::utf16ConversionCache flags.
  • Added a number of new keys to RegexKitLite NSError and NSException User Info Dictionary Keys.

Bug fixes:

  • If the ICU library returned a status of U_STRING_NOT_TERMINATED_WARNING during a search and replace operation, RegexKitLite erred on the side of caution and treated it as a failure due to the numerous search and replace bugs in the past. RegexKitLite 4.0 continues to err on the side of caution, but now treats U_STRING_NOT_TERMINATED_WARNING as if the ICU library had returned a status of U_BUFFER_OVERFLOW_ERROR.
  • The NSMutableString replaceOccurrencesOfRegex: methods could have returned a value > 0, the replaced count, even if no actual replacements were performed. This would occur if an error was encountered during the search and replace, in which case the replaced count was not correctly reset to 0. RegexKitLite 4.0 changed the replaceOccurrencesOfRegex: API so that a value of -1 could be returned to indicate that an error was encountered.
  • componentsSeparatedByRegex: bug fixed. The results returned were found to be different that the results returned by the perl split() function.

Epilogue

Coding Style

One noticeable break in style conventions is in line lengths. There was a time when 80 column limits made a lot of sense as it was the lowest common denominator. Today, a modern computer screen can display much more than just 80 columns. Even an iPhone, which has a screen size of 320x480, can display 96 columns by 24 rows of the usual Terminal.app Monaco 10pt font (5x13 pixels) in landscape mode. Because of this, my personal style is not to have an arbitrary limit on line lengths. This allows for much more code to fit on the screen at once, which I've heard referred to as "Man—Machine Interface Bandwidth". While you can always page up and down, the simple movement of your eye is almost always an order of magnitude faster. Paging through code also tends to break your concentration as you briefly try to mentally orientate yourself with the freshly displayed text and where the section of code is that you're looking for.

I try to group a line around relevancy so that based on the start of the line you can quickly determine if the rest of the line is applicable. Clearly the number of spaces used to indent a block plays a similar role, it allows you to quickly visually establish the logical boundaries of what lines of code are applicable. I also try to horizontally align related statements since your eye tends to be extremely sensitive to such visual patterns. For example, in variable declaration and initialization, I try to align the type declaration and the = (equal sign) across multiple lines. This tends to cause the declaration type, the variable name, and the value assigned to visually pop. Without the alignment, you typically have to scan back and forth along a line to separate and find a variable name and its initialization value. Sometimes line breaks and horizontal alignment are done purely on what's subjectively aesthetically pleasing and allows the eye to quickly flow over the code.

The source code of RegexKitLite isn't exactly what you'd call clean, there's more than a few crufty C barnacles in there. There is usually a choice between two polar opposites and in this case it's between elegant, easy to maintain and comprehend code, and speed. If you use regular expressions for very long, you will undoubtedly encounter a situation where you need to scan through tens of megabytes of text and the speed of your regular expression matching loop needs to be faster. A lot faster. RegexKitLite was written to go fast, and the source code style reflects this choice.

The Need for Speed

A significant amount of time was spent using Shark to optimize the critical sections of RegexKitLite. This included tweaking even the most insignificant details, such as the order of boolean expressions in if() statements to minimize the number of branches that would have to be evaluated to determine if the statement is true or false. Wherever possible, Core Foundation is used directly. This avoids the overhead of an Objective-C message dispatch, which would invariably end up calling the exact same Core Foundation function anyways.

Even the cache has what is essentially a cache— the last regular expression used. Each time a regular expression is used, the compiled ICU regular expression must be retrieved from the cache, or if it does not exist in the cache, instantiated. Checking the cache involves calculating the remainder of the regular expression strings hash modulo the cache size prime, which is a moderately expensive division and multiplication operation. However, checking if the regular expression being retrieved this time is exactly the same as the last regular expression retrieved is just a fast and simple comparison check. As it turns out, this is very often the case. Even the functions are arranged in such a way that the compiler will often inline everything in to one large aggregate function, eliminating the overhead of a function call in many places.

Normally, this kind of micro-optimization is completely unjustified. A rough rule of thumb is that 80% of your programs execution time is spent in 20% of your programs code. The only code worth optimizing is the 20% that is heavily executed. If your program makes heavy use of regular expressions, such as a loop that scans megabytes of text using regular expressions, RegexKitLite is almost guaranteed to be a part of the 20% of code where most of the execution time is spent. The release notes for CotEditor would seem to indicate that all this effort has paid off for at least one user:

More than 10.4 when running on the color process to review the details of the definition of a regular expression search RegexKitLite adopted, 0.9.3, as compared to the speed of color from 1.5 to 4 times as much improved.

Note:

This is an automatic machine translation from Japanese to English.

Documentation

Clearly documentation has been a high priority for this project. Documentation is always hard to write and good documentation is exponentially harder still. The vast majority of 'development effort and time' is spent on the documentation. It's always hard to judge the quality and effectiveness of something you wrote, so hopefully the extra effort is worth it and appreciated.

As an aside and a small rant, I have no idea how anyone manages to build so-called 'web applications'. I waste a truly unbelievable amount of time trying to accomplish the simplest of things in HTML, which is then multiplied when I check for 'compatibility' with different browsers and their various popular versions. What a joke, and that's just for this 'simple' documentation. Though I do have to give kudos to the Safari / WebKit guys, it always seems to be a lot easier to get the result you're looking for with the WebKit engine. The little things add so much: shadows, round rects, gradients, CSS animations, the canvas element, etc.

Questions about licensing, including in your software, etc

What do I need to do to include RegexKitLite in my software?
Not much. The BSD License is very permissive. In short, you just need to 'acknowledge' your use of it. Safari's Help Acknowledgements is a good example of this.
I am selling a 'closed-source' commercial application. I would like to use RegexKitLite in it, how much will it cost?
RegexKitLite is free for any use.
Do I have to distribute the source with my application?
No. Unlike the GNU GPL License, there is no requirement that you include or make the source code available, even if you release a "compiled, binaries only" end product.
What about modifications to RegexKitLite? Do I have to make any changes that I make available? Or contribute them back to the author of RegexKitLite?
You may make any modifications you want and are under no obligation to release those changes to anyone.
Why the BSD License?

It's a well known license. If you are part of a Large Corporate Organization, chances are the Corporate Lawyers have Decided whether or not the use of source code licensed under the BSD License is Acceptable or not. This can be a godsend for anyone who has to deal with such situations.

It also expresses a few things I think are perfectly reasonable:

  1. Give credit where credit is due.
  2. There is no warranty, so use it at your own risk.
  3. You can't hold the author responsible.

The first point is prescribed by most professional ethics already. Plus, it's always nice to see where your stuff ends up and how it's being used. The other points make explicit what should already be obvious. After all, you get what you pay for.

License Information

RegexKitLite is distributed under the terms of the BSD License, as specified below.

See Also

License

Copyright © 2008-2010, John Engelhart

All rights reserved.

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

 
RegexKit project hosted by: