I’m not exactly sure what you mean. Doesn’t all dependency injection work the way I described?
Without being familiar with the framework, you can’t trace your way from the class getting injected into to the configuration, even if you’re experienced with the language.
I don’t think so. When I’ve seen it done it’s usually not been random values injected (except when those values are secret keys which should absolutely not be stored in code to begin with), it’s usually injecting a service. Another class, usually with a name that makes it easy to trace down. Like if you’re building an OrderService, that might take as a dependency an IProductService, which would have injected into it the class ProductService (personally, I don’t like the Hungarian notation that C# uses, but it’s a convention and my preference for sticking to strong conventions for the sake of consistency outweighs my distaste for Hungarian notation). That’s easy to find, and in fact your IDE can probably take you straight to it from the OrderService’s constructor.
Spring singleton beans are supposed to be stateless though, so they can’t be called variables. Maybe the DI aspect of Spring is less relevant today in the micro service era, but in the day Spring helped make layered monolith apps much cleaner.
Really? From my experience the opposite is the case. I work on a smallish team with 3 other developers and we also have a few spring services with < 100 classes and we constantly run into issues where making changes to a bean causes issues in another unrelated part of the codebase. I can’t imagine what a nightmare it would be with a larger codebase and more devs working on it.
My favourite take on DI is this set of articles from like 12 years ago, written by a guy who has written the first DI framework for Unity, on which are the currently popular ones, such as Zenject, based on.
The first two articles are pretty basic, explaining his reasoning and why it’s such a cool concept and way forward.
Followed by more articles about why he thinks it was a mistake, and he no longer recommends or uses DI in Unity in favor of manual dependency injection. And I kind of agree - his main reasoning is that it’s really easy for unnecessary dependencies to sneak up into your code-base, since it’s really easy to just write another [Inject] without a second thought and be done with it.
However, with manual dependency injection through constructor parameters, you will take a step back when you’re adding 11th parameter to the constructor, and will take a moment to think whether there’s really no other better way. Of course, this should not be an relevant issue with experienced programmers, but it’s not as inherently obvious you’re doing something potentially wrong, when you just add another [Inject], when compared to adding another constructor parameter.
Exactly. Dependency injection is good; if you need a framework to do it, you’re probably doing it wrong; if your framework is too magical, you’re probably not even doing it at all anymore.
Can we talk about annotations which are broken when you upgrade spring boot ? You are asked to upgrade some old application to the newest version of spring boot, application that you discover on the spot, the application does not work anymore after the upgrade, and you have to go through 10 intermediate upgrade guides to discover what could possibly be wrong ?
An application I’ve never heard or seen before that needs to be upgraded, and it breaks, so you now need to understand what the hell this application does so you can fix it properly.
Spring annotations in general. There’s a completely hidden bean context where every annotation seems to throw interceptors, filters, or some reflection crap into. Every stacktrace is 200 lines of garbage, every app somehow needs 500mb for just existing and if you add something with a very narrow scope, that suddenly causes something completely unrelated to stop working.
Realistically, DI and all the Spring crap does not add anything but complexity.
Our Spring service was so simple until we decided we needed annotations to handle the fetching of settings. Now we are corrupted with needless reflection.
Gradle, with it’s transitive dependency modifications is a huge pain in this area.
It used to be that if a library ended up having a flaw then it would be flagged and we would get the dependency updated. These days security block the “security risk” and you have to replace your dependencies dependency. Fingers crossed you can get it to actually test all the code paths.
If an second level project gets a flaw, and it’s used indirectly then we should really look at getting the import updated so that we know it works. If that import is abandoned then we should not be updating that second level dependency, either adopt and fix the first level dependency or look at an alternative.
Add comment