Am 15.02.2015 21:05 schrieb "Rowan Collins" <rowan.coll...@gmail.com>: > > This sounds to me like you should just be using the Singleton pattern,
Of course this is singleton under the hood. > // Now wherever in the code you want the default instance, just use this: > $value = MyRedis::singleton()->get($key); You can surely see how this is more readable / easier to write: $value = MyRedir::get($key); > The nice thing about an explicit Singleton is you can migrate to Dependency Injection (call "$redis = MyRedis::singleton()" and start passing the instance around, I _can_ do that with my redis class, where I need it, by calling a method MyRedis::link() which just returns $this. But the places where I need that are rare. > ... or to a cleverer factory / service locator (store more than one instance in different static variables or a keyed array, to connect to different stores for different purposes). This i do with subclasses, where the subclass define suitable static protected properties that guide the connect logic (which sits in the baseclass). I have a global redis, a local redis (local to the host the code runs on), a volatile redis (global, but does not make prersistent dumps), and so on. Calling code ready red_global::this(), red_volatile::that(), and so on. You don't get it better readable (and greppable) as that. > Making static calls implicitly access a singleton seems like it would make code harder to read, and harder to extend, I don't understand. It's the best to read to me. And the best to extend - I just make another subclass. Of course, there are places (in other classes) where stuff needs to use such a redis instance without caring which one it is - i.e. an instances is passed to a constructor there - in which case I can use the MyRedis::link() method to get at the singleton instance. But in most (I'd guess 90%) of the calling places, it's just the static calls I need, and writing each of them as MyRedis::singleton()->xxx() is more tedious. best regards Patrick
