Anyone know what's the minimum size of the compiled JS now? Is there any way Scala.js can produce code as compact as BuckleScript/ReasonML in the future?
I like Scala's syntax but the last time I tried Scala.js, the output was several hundred kb for something slightly more than a Hello World -- it used some List and Seq functions. I'd bet the comparable ReasonML output would likely be in tens of kb (including List and Array functions from Belt).
The bare minimum for a hello world is 6 KB (before gzip). But of course as soon as you start using Scala collections, it's getting bigger. The minimum for a realistic application is 90 KB (before gzip).
That's less than your typical JavaScript framework, so for any realistic application it shouldn't be a concern.
For refernce- Jquery unzipped is similar size- 87K and while it adds much more functionality, Scala collections easily trump jquery's in usability and power.
SJRD, Would you also add a bit more info about how this size increases? I haven't cared much but it seems from your comments that the slope should be much less once you reach 90kb.
ClojureScript ends up somewhere around the same size. Maybe it's some magic number that a full-featured environment converges to.
Since Scala.js (like CLJS) is backed by Google Closure Compiler, I bet you'll also reach for the Closure Library for a lot of stuff where one might reach for JQuery or equivalent otherwise. The stuff in there is of course optimized for Closure Compiler, and slims down quite well.
There is a significant difference between CLJS and Scala.js regarding their use of GCC (Google Closure Compiler).
In ClojureScript, GCC is a user-visible feature of the toolchain, so users are encouraged to use GCC-compatible libraries and the Closure Library in particular.
In Scala.js, GCC is really an implementation detail of the toolchain, and its constraints are completely abstracted away from the user. This does not really encourage using GCC-oriented libraries like the Closure Library.
The two approaches each have pros and cons:
* CLJS' approach allows tree shaking across the CLJS code and the GCC-compatible libraries that are used. However, it makes it more difficult to use libraries that do not comply with GCC's Advanced Optimizations requirements, as one needs to declare the proper `externs` files.
* Scala.js' approach means that you can use any JavaScript library out of the box, whether or not it was designed for GCC, and without having to declare any `externs` file. However, it means that GCC's tree shaking will not be applied across the Scala.js code and the JS libraries. It only applies to Scala.js' own code.
I think the externs file maneuver might be outdated.
Admittedly, I'm using shadow-cljs as my go-to compiler (so I'm not sure what the state of vanilla is), but all I do is yarn add <package>, and then require it in the file where I want to use it, and interact with it as if it were any other library.
Though it is true of shadow-cljs as well that it doesn't try to tree-shake external (non-CLJS/Closure) libs by default, though it will do it for project code and Clojure libraries.
Indeed, once you end up pulling in most common Scala classes (collections, futures, etc.), adding new functionality to your app increases app size by very little. Our app has ~600KB gzipped js (attributable to Scala.js compiled code) + ~200KB of React and other libraries. The overall size has barely moved beyond +-100KB in last few years despite adding new functionality all the time.
Once Scala.js supports lazy loading of modules (hopefully soon after 1.0), the size issue will have significantly less practical impact.
Scala.js is being fed to the Google Closure compiler [1] for minification via tree-shaking, after applying optimizations on its own. Output can be unreadable, but debugging works via source maps.
It's not perfect, but you should not get hundreds of KB for a Hello World. And indeed, the code cannot be as minimal as hand crafted code either.
Note that this is often a false problem, because people often use a lot of JavaScript libraries without any tree-shaking in their build process, so if you do care about tens of KB, then you don't use JQuery, React, etc. A valid use case indeed, I've been there, but not for your typical app.
While i understand your concern, i wouldn't use scala.js in a context where several hundred kb are an issue. I've tried it and it works great, but i think it only pays of if you implemented a complicated web-app with complicated server-side code, e.g. if you replace a desktop-application to manage some system with a web-based one. Something where showing a loading-screen on the first visit is totally reasonable.
> the output was several hundred kb for something slightly more than a Hello World -- it used some List and Seq functions
That's just nonsense, maybe 6 or 7 years ago that could have been the case. There's an approximate 100KB "tax" for pulling in Scala collections (i.e. as soon as you depend on List, Map, Seq, etc.), which is the price to pay for being able to use the power of Scala in the browser.
> Is there any way Scala.js can produce code as compact as BuckleScript/ReasonML in the future?
No, iirc, not without a redeisgn.
There's a longstanding issue wrt lazy loading, which is slated to be worked on at some point post-1.0 -- I'd say that's probably the biggest missing feature, the ability to break out one's app into dynamic chunks, so for example a lightweight frontend/public facing area of an app can be isolated from the backend where all the heavy lifting is done.
Perhaps with Dotty (AKA Scala 3) next year there will be further reductions in Scala.js code size, but no, nothing like BuckleScript.
I can't speak for everyone, but while I loved programming in Scala, I can't stand sbt[0] as a build tool. If I could use Scala.js without all this (e.g. Webpack or Rollup) could see myself using.
sbt is not the only build tool that supports Scala.js. I at least know of support for Maven [1], Pants [2], Gradle [3] (a bit outdated that one, but an interested person could revive it), and there seems to be a path forward for Bazel [4].
That said, I would like to see a Webpack plugin for Scala.js as well.
It's exactly like TypeScript. You write type definitions or you find them online, like in TypeScript, then you can call any JavaScript API using the relevant types.
Always been curious about workflow of people who use Scala.js. I am a complete newbie when it comes to frontend. How do you do frontendish interactive development considering scalac's slowness?
scalac is only slow when it needs to compile an entire project. But the Scala tooling ecosystem has invested a lot in incremental compilation, which means that most of the time when you make a change, only 1 or 2 files are recompiled, which is quite snappy. In turn, the Scala.js linker also performs its linking and optimizations incrementally, so that in an interactive development setting, the whole cycle is quite fast.
fine, but what tooling/workflow would one use for the type of development javascripters are accustomed to? Like make a change and immediately observe result on the page?
As someone who has dropped Scala on the backend (not doing FUD, just not happy with the JVM memory reqs)- what keeps me still using scala is Scala.js.
While some argue that there is more readable code with Bloomberg's OCaml.js, I like Scala.js because of Scala itself, which is
1) Very well designed
2) Lots of "functional" constructs in accompanying libraries
3) Good typechecking
4) Fun language to use.
and because Scala.js interop with JS is very seamless.
Do check out ScalablyTyped repo on Github, a Scala.js analogue of DefinitelyTyped.
Being typed- one of the best things I like about Scala is that I can do something like this (provided by a couple of frameworks- Laminar, Scalatags et al)
No more mismatching tags, wrongly spelled tags, unclosed tags. Typechecked HTML in some sense.
This would also be a good place to mention some alternatives where a mainstream language can also compile to JS so hopefully people can comment with their experience:
1) PureScript
2) ELM
3) Clojure script (child comment)
... and of course... TypeScript! :)
It should be noted that while Scala.js isn't as popular as some of it's alternatives in "popular" programmer culture, the author is very focussed on correctness and I don't know of anybody who has ever used Scala.js for a project but dropped it. (Happy to be corrected)
I'm curious if you did any measurements and what you prefer instead.
The managed heap is very efficient. I have a web server (built in Scala) configured on my own VPS to run with a 10-100 MB heap. To this you add the JVM's memory used for managing OS threads, opened sockets, JNI, GC overhead and I end up with a process using 190 MB (RES), which is reasonable for a server-side process, because you can actually do a lot in it, Java processes being able to be fat.
In my experience a well grown Java process will be much more efficient and predictable in terms of memory used than Node.js, PHP, Ruby, Python, at the very least.
The only problem I've experienced is the startup time, plus depending on use case I might prefer native binaries. GraalVM looks promising.
Scala.js is useful on the sever for example if you need to work with "serverless" tech, e.g. AWS Lambda, for which startup time is important.
I have some very simple lua processes that consume 2MB. They basically do nothing but listen to a websocket and run some shell comands. Lets say I have 3 of those. With the JVM I would have blown away more than half a gigabyte. Even with micronaut you're looking at 100MB per process. This just isn't sustainable. I personally think the JVM is a piece of crap because of this.
There is GraalVM but compilation takes a relatively long time even compared to a C++ project and as soon as you use features that it doesn't support it will fail to work which means you are stuck with Java because you don't want to risk using a language that doesn't officially support GraalVM (groovy kotlin etc).
True, but in the same way that a 747 is terrible for delivering pizza.
You may be right that Java doesn't scale down well, but that's not normally the sort of work it's intended for. (Let's ignore the applets mistake.) It's a terrible solution for scripting, yes, but for heavyweight server roles I'd be surprised to see it significantly outperformed by .Net, and there's really no way Lua or Python would be able to compete.
OpenJDK has a very impressive (and heavyweight) JIT, for instance, and very impressive GCs of almost no use for short-lived applications.
With all that said, it'd be great if OpenJDK's new developments mature to the point that Java can become practical for 'lightweight work'.
well compared to dotnet core 3 the java memory management is really really bad.
in java memory is not more efficient than in node.js/php/ruby or python it's worse. what is better is the performance, especially when you have a lot of stuff on the heap the performance is predictable even with a memory size of > 16gb where the other languages lack of. GraalVM still misses a lot of stuff and is slower.
so I think the biggest buck for the bang actual is dotnet (maybe go is as good as well, but does not really play in the same league). I think once valhalla hits the tables might change, but even than ReadOnlySpan<T> is way too awesume than just value types.
Speaking of the mainstream scripting languages ...
Python often leaks memory, the GC being built on top of reference counting to solve cycles, isn't compacting and can also be left disabled for performance, which is a trap. This is made worse by native libraries that also leak. It has gotten better, Python apps being one of the most deployed. But I haven't seen a Python app that's memory efficient yet, except for short lived shell scripts.
PHP uses tens of MB per request, which due to its "shared nothing" policy means you need several GBs of RAM for serving a Wordpress website that can withstand traffic spikes.
Many companies actually deploy their Ruby apps on top of the JVM, via JRuby, due to better memory usage.
Node.js is the only one that's more efficient out of this bunch, the GCs have similar implementations to the JVM and optimized for devices that are memory constrained, the problem being that Node.js processes are single threaded, so to distribute the load, you end up with multiple processes, plus due to JavaScript you also get more garbage.
> the biggest buck for the bang actual is dotnet
Not in my experience, but granted I haven't tested .NET Core yet.
well he talked about the AOT stuff. which is not on par as dotnet core's aot approach in 3.0, however default graalvm is quite fast (but still memory hungry)
I recently learned that Scala.js doesn't yet support generating TypeScript *.d.ts type definitions. I intend to open a Gitcoin bounty to get this work done.
Have you had any success using Scala.js-generated JavaScript in TypeScript projects?
I'm quite interested in a full stack of create-react-app + TypeScript + graphQL + front end libraries written in Scala + back end services written in Scala.
In general I'd advise against using Scala.js for libraries. The Scala.js runtime is pretty heavy, and Scala's and TypeScript's type systems are different enough that it tends to be difficult to design an API that feels native in both.
Have you considered ScalaJS-React[0] or Outwatch[1] instead?
Are you not happy with JVM in general or not happy with how Scala uses the JVM? Things like Spark were written in Scala so I am curious what didn’t work for you.
Spark uses off heap memory via JNI in order to avoid JVM memory management drawbacks. It hides it behind a heavy abstraction, which makes it plausible to use such solution for data processing pipelines, for the smaller applications not so much.
But for most things that is not that much a problem - I'm not saying that JVM memory model makes Scala unusable, just that Spark is not a great example here.
Scala Native [1] has been really quiet of late (last commit July 29), not sure what's going on. There was a lot of excitement when the project kicked off, hopefully it continues to evolve.
The creator and lead contributor has been finishing up his school but I am not sure of his plans for the future. There is a book coming out on the subject, https://pragprog.com/book/rwscala/modern-systems-programming... and I am hoping it will continue to grow and evolve.
https://slinky.dev