Hahahahaha 😂

Rust is for bignner's , he will rewrite JDSL in JDSL

@@vaisakhkm783 Recursive JDSL lol

@@-rya1146 bootstrapping JDSL compiler

JVM in JDSL

@@vaisakhkm783 Rewrite Rust in JDSL then

Do you know how many common file formats are just zip with extra steps?

@@harrytsang1501 I even rename some zip files, so it can associated with a default program. It's a common technique.

@@thingsiplay OMG!

A lot of files are just zip files with some bits and bobs in them. And for those that aren't, they might be SQLite databases instead.

i think .exe files are also zip with extra steps... i might be wrong...

🤣🤣🤣

That would actually be pretty funny.

call it LPython

Lython

Lithium

guy is built different.

😂 😂 😂

Doesn't it technically bootstrap?

writing a VM is actually pretty simple! but its awesome still

I'm still extremely intrigued by macros in rust If the html macro can validate and provide validation, can we have a typeScript macro that creates an AST with type annotations and will throw errors if the typescript is invalid. Just a thought

@@harrytsang1501 why not? tokenize, parse, lint and return from macro. no rocket science, matklad could do that in a week, probably

@@lukaszoblakTom would do it in a day, because Tom is a genius.

I'm very suspicious of that @daltonyon is an experimental bot, testing whether they will get anyone noticing it's a bot.

HotSpot is already so supercharged that I doubt you can make a rewrite that is better than it is, even if you rewrite it in Rust

Can you link blog post please ? Would love to read about EVM implementation in rust.

@@shark950621 KZfaq doesn't let me share it, you'll have to figure out how to find it 🤣

Yes you can do any compiler in java.

​@@MrDejvidkitcan you write a JVM in Java?

@@BubbleBeeee Already exists. The one I know of is called SSVM.

​@@BubbleBeeeeGraal

And I assume that the JBoss devs couldn't set up a HornetQ JMS service? C# can speak to that quite easily over STOMP, after all.

@@HrHaakon That sounds interesting. I was doing Java a long time ago. I wasn't sure at the time if there were already ways to be compatible with RMI from C# (which is based on Java serialization). But the very first thing I noticed was a possibly huge vulnerability where you could invoke any zero arg constructor remotely(!!) as a side-effect of how it was done. I was making noise about it to Sun, but never head of a vulnerability based on it until a few years later.

“Java is pretty damn fast” 🥶🥶🥶🥶🥶🥶🥶🥶

The cooler thing is the languages that use the jvm as a compilation target like scalla and kotlin. The space is pretty interesting

Literally nothing about the Java world is exciting unless you are 80 years old

@@PanosPitsi "Java is already pretty damn fast for not running native code" There, fixed it. Java is one of the fastest interpreted languages you know?

@@user-kf7vy5yu2v Tell me you haven't heard about Kotlin without telling me you haven't heard about Kotlin.

The Lord's language Java is clearly my favorite

@@ThePrimeTimeagen cLeArlY

Wont happen

@@hilligans1 is there any specific reason that won't happen?

@@avalagum7957 they gain nothing from adding

​@@avalagum7957Adding a new language to HotSpot requires a new compiler, in this case rustc, which is a lot of work. Considering even just moving from C++14 to C++17 is already so difficult, I doubt Rust will ever be used in the JVM

I've been writing a GC for my Lua interpreter and I have had to use unsafe. I'm planning to abstract it all a bit to make optimisation easier but under the hood it will remain pointers.

GraalVM fan spotted! :p

I dont think so.

??? The primary JVM (HotSpot) is also written in C++, an also non garbage collected language

Half the speed in 1/100th the compile time is a great tradeoff if your test has half the load of production... ;)

​@@HrHaakon Well, it's more complicated than that. If it takes for example, 150ms to reach a stage in code where you need to check something or it fails then any code that runs after that is irrelevant. It does not matter how much things there is to test. Plus if you really wanted to, you could spawn an actual compilation task in the background and when it completes, transition the execution to actual compiled code. It is funny that's actually how java JIT works. It runs an interpreter that can instantly run code not that fast and starts compiling it in to an optimized machine code version. Once that's done, the interpreter hands off the runtime memory to it.

Definitely nothing newer than Clojure 1.9, as it is the last one that runs on Java 7.

Yea Generics are implemented primarily in the compiler. Type erasures are just that. The types are GONE at runtime. But reflection does have generic tie-ins which he said he didn't implement so maybe it would be just that like the Type interface, etc.

This is only true now, generics are being reworked in Valhalla

Recently we have evaluated Go and Java frameworks for writing Kubernetes operators. While Go is the main language in the kubernetes world, and the most supported, the java framwork resulted in better code. Less cluttered, more coherent, no magic comments for the framework etc. Java rules!

I've found this out too, from my limited experience with it, Java appears to be a solid and professionally wrought piece of engineering, if being perhaps somewhat of a product of its time (but gets better all the time!)

It's not so much Java that's the problem I don't think, more so the patterns that people are funneled into for using Java, like the much-maligned AbstractFactoryFactoryFactory. But really you don't have to do it that way, and you can do pretty ok with a bunch of static methods instead of OOP-the-world madness.

@@banatibor83sounds like java programmers not knowing how to use a language that doesn’t have classes

​@@isodoubIet Static methods are fine. You should probably start with static factory methods. Hell, that's how Java does it. UUID has fromString for example. You don't do new UUIDParser.parse(String). But there are reasons for factory objects etc. as well. And having a lot of static methods have led to absolute pain in my experience. It makes certain parts of testing things very hard and leads to people googling things like "how do I mock static methods that are hardwired?" instead of just injecting them in the first place as you're supposed to. For example a thing that takes a String and returns SomeThing is a Function. This actually does help a lot when the code base gets larger and older, because dependencies are now dragged out into the sunlight. Yeah, you now have to expose some wiring, but when you're later coming trying to find the faulty wiring, you'll be very happy that the wiring is visible... ;)

What is it written in if not Rust?

@@Cookiekeks 🥺 But but blazingly fast

Instead, just write a Rust compiler in Java.

It literally says it in the first paragraph or so. It's for fun and educational purposes, most of these things aren't useful by themselves, but doing them make us better developers.

For fun, to learn

@@anhta9001 I read that in the "programmers are human too" voice

It is fun bro

You think rust is superior? Likely so does the author. Guy implements a garbage-collected vm language and you’re hung up on the syntax. I would not be surprised if Rust ends up in the JDK after this, nor if the author scores a Rust job because of his skills.

@@conorx3 " and you’re hung up on the syntax." stop projecting your opinions onto others. The whole point of my comment is not about syntax, is precisely about the semantics, rust is superior to java because of the semantics, because there is no runtime. Rust syntax is made around the borrow checker, why implement a borrow checker language on a garbage collected? you wouldn't even have the same experience at all because by writing with a rust-ish syntax on the jvm will give you absolutely nothing of the flavor of rust, will hinder only your experience in real rust. Is just a waste of time in every sense. if you think is "just for fun" well, that is just silly anyways.

How is this related to the Rust ecosystem?

*reads video title and rushes to create drama*