MapStruct VS runtimelab

Currently using .NET and ASP.NET for the API of a project that I'm working on, overall it's a pretty lovely experience, especially with the JetBrains Rider as my IDE. Having worked with Java in the enterprise in the past, it feels like a more focused and sometimes more coherent experience (vs the more decentralized nature of Java), that now thankfully runs fine on Linux as well, don't even need to hope that Mono is good enough anymore.

Entity Framework Core is great, the way how they fetch related data is straightforwards: https://learn.microsoft.com/en-us/ef/core/querying/related-d...

I also really like the ability to use split queries, when needed: https://learn.microsoft.com/en-us/ef/core/querying/single-sp...

You can even scaffold your entity mappings from an existing database and manage database migrations separately (which I feel is a must unless you want to pour hours upon hours into learning yet another JPA-like), it's nice when something like that exists in a given ecosystem: https://blog.jetbrains.com/dotnet/2022/01/31/entity-framewor...

Some of the stuff in the ecosystem is also really nice, like using ActionBlocks from Dataflow for simple queues and task scheduling is very easy to get started with and pretty pleasant: https://learn.microsoft.com/en-us/dotnet/api/system.threadin...

Honestly, most of the docs that I've come across are rather pleasant and there's plenty of code examples to be had and even LLMs have enough codebases out there to be trained on and give decent output, in addition to the regular autocomplete in the IDE.

Even some third party stuff like CsvHelper is really well put together: https://joshclose.github.io/CsvHelper/getting-started/#insta...

That said, I've had issues where trying to deserialize incoming JSON requests ([FromBody]) lead to the whole object being null, due to it not matching the fields 1:1 or violating some nullability constraint, however that happened without a clear error message or as much as a warning in the logs, which felt insane and wasn't very discoverable. Now I just use JObject as the parameter and deserialize at the top of the controller method.

I've also run into some ecosystem issues where the OpenAPI codegen was a bit lacking and wouldn't actually generate code that works, but maybe that's because the spec was bad: https://openapi-generator.tech/docs/generators/csharp

I could also not get AutoMapper (https://docs.automapper.org/en/stable/index.html) working satisfactorily, so for now I'm stuck writing mappers by hand with said autocomplete sometimes doing it for me. There were issues with the runtime being unable to register the mappers or something like that, it was a while ago. Overall it felt like the MapStruct library in Java did everything better: https://mapstruct.org/

On the other hand, I haven't had any of the Spring Boot DI related headaches (needing @Lazy sometimes) with ASP.NET, things there seem way more straightforwards and you clearly define whether something should be a Singleton or Scoped and the rest... just gets out of your way and works.

Sadly, there have been very hard to track down cases where I try to do a _dbContext.Add(someEntity) and then when I do a save with just the fields in that one entity changed, the performance absolutely seems to slow down to a crawl. Adding _dbContext.ChangeTracker.AutoDetectChangesEnabled = false seemed to help, but I am still not entirely sure what lead to that behavior, the temp directory on my Linux box also started filling up very rapidly, yet the logs had nothing of use, definitely seems like something I'd need to investigate properly.

Overall, .NET and its ecosystem is pretty good in most cases. And .NET being used in gamedev is also cool, for those who care about that sort of thing, I wonder why jMonkeyEngine never really took off.

In Java (woof..) we have MapStruct (https://mapstruct.org/). Anything like that in Python? I think maybe the Sqlalchemy mapping in Litestar2 is the closest I've seen.

Done right i can look something like mapstruct for example. But like any other feature you need to get a feeling for when it's a good idea to use it.

We are using mapstruct but it sucks when you have protobuf :/

You all beat me to MapStruct and Testcontainers. Honorable mention to RxJava, which I use in Desktop apps.

For the mapping, I'm using Mapstruct, I got the instance example from their examples they use INSTANCE as constant.

Now update your code so that no entities are passed to/from the domain layer and that no objects are passed to/from the domain layer. You'll look at me as though I'm nuts, but you will thank me later. Look at things like http://immutables.github.io/ and https://mapstruct.org/

When my pairing partner and I ran into this use case, we were initially stumped. Even our combined Google Fu efforts were not turning up a good solution. However, we eventually discovered a PR for MapStruct that seemed to be the solution to our issues! But, of course, if it were that simple, I would not be writing this now. The SubClassMapping annotation provided by that PR, while quite helpful, requires specific Mapping annotations for every field on the subclass. We wanted a more elegant solution.

Just use MapStruct...

There was a "green thread" experiment for dotnet a while ago, here is the conclusion: https://github.com/dotnet/runtimelab/issues/2398

Experiment result write-up: https://github.com/dotnet/runtimelab/blob/e69dda51c7d796b812...

TLDR: The green threads experiment was a failure as it found (expected and obvious) issues that the Java applications are now getting to enjoy, joining their Go colleagues, while also requiring breaking changes. It, however, gave inspiration to subsequent re-examination of current async/await implementation and whether it can be improved by moving state machine generation and execution away from IL completely to runtime. It was a massive success as evidenced by preliminary overhead estimations in the results.

Yeah, it kind of is. There are quite a few of experiments that are conducted to see if they show promise in the prototype form and then are taken further for proper integration if they do.

Unfortunately, object stack allocation was not one of them even though DOTNET_JitObjectStackAllocation configuration knob exists today, enabling it makes zero impact as it almost never kicks in. By the end of the experiment[0], it was concluded that before investing effort in this kind of feature becomes profitable given how a lot of C# code is written, there are many other lower hanging fruits.

To contrast this, in continuation to green threads experiment, a runtime handled tasks experiment[1] which moves async state machine handling from IL emitted by Roslyn to special-cased methods and then handling purely in runtime code has been a massive success and is now being worked on to be integrated in one of the future version of .NET (hopefully 10?)

[0] https://github.com/dotnet/runtime/issues/11192

[1] https://github.com/dotnet/runtimelab/blob/feature/async2-exp...

Unlike these folks from dotnet, which tested directly on ASP for real workload

  https://github.com/dotnet/runtimelab/issues/2398?darkschemeovr=1

[1] https://github.com/dotnet/runtimelab/issues/2398

The only way to avoid it is to not build on top of Java or not adding any features on top of Java.

> To give another example with C#, there has been a lot of recent discussion about finding potential alternatives to their async-await concurrency model. They cite the level of effort it takes to maintain the async await style code and the costs that come from this.

I had a very different take-away. They did PoC with virtual threads and decided it's not worth the switch now and async-await that they have is good enough.

https://github.com/dotnet/runtimelab/issues/2398

> Some of the languages it gets compared too aren't even that old yet.

C# is old enough to drink and Scala just had its 20th birthday this week :)

It was tried and the dotnet team decided to drop it: https://github.com/dotnet/runtimelab/issues/2398

Technical details here: https://github.com/dotnet/runtimelab/blob/feature/green-thre...

Just last month .NET ended a green threading experiment, mainly because the overhead it adds to FFI was too high:

https://github.com/dotnet/runtimelab/issues/2398

Rust had green threads until late 2014, and they were removed because of their impact on performance.

Everyone has done the basic research: green threading is a convenient abstraction that comes with certain performance trade offs. It doesn't work for the kind of profile that Rust is trying to target.