OSCache is an open-source re-implementation of NSCache that behaves in a predictable, debuggable way. OSCache is an LRU (Least-Recently-Used) cache, meaning that objects will be discarded oldest-first based on the last time they were accessed. OSCache will automatically empty itself in the event of a memory warning.
OSCache inherits from NSCache for convenience (so it can be used more easily as a drop-in replacement), but does not rely on any inherited behaviour.
OSCache implements all of the NSCache methods, but does not currently support anything relating to NSDiscardableContent and will always return NO for evictsObjectsWithDiscardedContent regardless of the value you set it to.
- Supported build target - iOS 9.1 / tvOS 9.1 / Mac OS 10.11 (Xcode 7.1, Apple LLVM compiler 7.1)
- Earliest supported deployment target - iOS 7.0 / tvOS 9.0 / Mac OS 10.10
- Earliest compatible deployment target - iOS 4.3 / tvOS 9.0 / Mac OS 10.6
NOTE: 'Supported' means that the library has been tested with this version. 'Compatible' means that the library should work on this OS version (i.e. it doesn't rely on any unavailable SDK features) but is no longer being tested for compatibility and may require tweaking or bug fixes to run correctly.
OSCache requires ARC. If you wish to use OSCache in a non-ARC project, just add the -fobjc-arc compiler flag to the OSCache.m class. To do this, go to the Build Phases tab in your target settings, open the Compile Sources group, double-click OSCache.m in the list and type -fobjc-arc into the popover.
If you wish to convert your whole project to ARC, comment out the #error line in OSCache.m, then run the Edit > Refactor > Convert to Objective-C ARC... tool in Xcode and make sure all files that you wish to use ARC for (including OSCache.m) are checked.
To install OSCache into your app, drag the OSCache.h and .m files into your project. Create and use OSCache instances exactly as you would a normal NSCache.
In addition to all of the inherited NSCache methods, OSCache adds the following:
@property (nonatomic, readonly) NSUInteger count;
The total number of items currently stored in the cache;
@property (nonatomic, readonly) totalCost;
The total cost of all items currently stored in the cache;
- (void)enumerateKeysAndObjectsUsingBlock:(void (^)(id key, id obj, BOOL *stop))block;
Enumerates the keys and values stored in the cache.
The OSCache delegate now implements the OSCacheDelegate protocol, which is a superset of NSCacheDelegate. You can declare your delegate as supporting either the OSCacheDelegate or NSCacheDelegate protocols; either will work without warnings.
OSCacheDelegate adds the following optional method:
- (BOOL)cache:(OSCache *)cache shouldEvictObject:(id)entry;
This method is called before OSCache evicts an object from the cache, giving you an opportunity to veto the eviction. You can use this method to implement your own cache clearing criteria (e.g. you could decide to only empty the cache if another cache is already empty).
The method will only be called as the result of adding an item to the cache, or in the event of a memory warning; it is not called if you explicitly remove and object using -removeObjectForKey: or -removeAllObjects. Objects will always be evicted in order of least recently used.
When the cache still has space in it, reading, writing and removing entries has constant time (O(1)). When the cache is full, insertion time degrades to linear (O(n)), but reading and removal remain constant.
For this reason, you should ideally size your cache so that it will never get full, but if that isn't possible, it's better to select a smaller size, as very large sizes will degrade significantly in performance when they fill up.
Version 1.2.1
- Fixed bug where
enumerateKeysAndObjectsUsingBlock:returned internal wrapper instead of cached object - Added lightweight generics annotations
Version 1.2
- Substantially improved insertion performance when cache is full
- Fixed bug where private container objects were passed to delegate instead of cached object
- Added enumeration and subscripting access for cached objects and keys
Version 1.1.2
- Added nullability annotations
- Fixed nullability error in Xcode 7
- Fixed unit tests for Xcode 6.3 + 7
Version 1.1.1
- Fixed bug where sequence numbers would not be resequenced if an overflow happens when reading from cache
Version 1.1
- Exposed the
-countand-totalCostproperties - Added
OSCacheDelegateprotocol, which is a superset ofNSCacheDelegate - Added optional
-cache:shouldEvictObject:delegate method - If cache is cleared as the result of a memory warning,
-cache:shouldEvictObject:and-cache:willEvictObject:is now called for each item - Now uses sequence numbers instead of time for sorting cache items (more reliable)
- Now uses
NSLockinstead ofdispatch_semaphore(more appropriate) OSCachestill behaves as if it inherits fromNSCache, but no longer actually does so, avoiding possible breakage if Apple changes theNSCacheimplementation in future- Added unit tests
Version 1.0
- First Release