I think this + node:test makes Node.js a pretty compelling sensible default for most things now. Running things with `tsx` was such a QoL improvement when it happened, but it didn't solve everything.
Runtime type assertion at the edges is mostly solved through `zod` and tools like `ts-rest` and `trpc` makes it so much easier to do full-stack Typescript these days.
This. It's 2025 and the node ecosystem is finally usable by default!
ESM modules just work with both Node and Typescript, Node can run .ts files, and there's the a good enough test runner built in. --watch. The better built in packages - `node:fs/promises` - are nice with top-level await for easier async loops.
It took a while to convince everyone involved to just be pragmatic, but it's nice now.
What's the story with supporting CommonJS libraries? I've tried to update many projects to ESM multiple times over the years, and every time, I ended up backing out because it turned out that there was some important upstream library that was still CommonJS - or even if we fixed those issues, our downstream NPM consumers wouldn't be able to consume EJS. So then you have to go down this rabbit hole of dual compilation, which actually means using something other than tsc.
This is great to hear, but perhaps comes too late for people like myself. Node.js has been by go-to platform from around 2014 until last year. But around September last year, I found myself thrust into the .NET ecosystem (due to a client project). Within a few months, I realized that it too, had finally become usable by default (unlike the last time I tried it, when it was too tightly coupled to Windows). In fact, it felt like what Node.js would be, if it had strong typing built-in, and had a good standard library that eliminated a lot of the module management and churn. I'm now finding it hard to return to Node.js.
Yes, with the difference that Google would have to be compromised in order to poison the go distributable containing fmt tool. With js, it’s enough to poison any single one of the 1400 dependencies of the linter
I forgot that even though fmt will never suffer from middle man attacks downloading the Go toolchain, the standard library already covers 100% of the uses cases someone cares about using Go for, and no one is using CGO.
I'm very much in favor of TS support directly in node. vitest has made it easier these days, but I've lost too much time over the years getting the balance just right when configuring test environments for .ts files.
trpc and ts-rest are a different animal in my opinion. I'm happy to use either one but won't deal with them in production. For trpc that's mainly due to the lack of owning API URLs and being able to more clearly manage deprecating old URLs gracefully.
For ts-rest I just tend to prefer owning that setup myself, usually with zod and shared typings for API request/response pairs. It also does irk me every time I import what is clearly an RPC tool named "-rest"
It doesn’t. The comment you’re replying to is referring to tsx, the package that lets you execute ts files, not to running files with the tsx extension.
i switched to python a while ago. it has batteries included. i feel so much better now that i dont have to debug all the quirks of a half-baked system.
This is great up until you get to the fact that typescript will not be accepted under node_modules [0].
That leads me to ask, what about project dependencies? I wrote a lib for my data models in typescript and I want to import that into my app in node, in typescript? Does the rule only apply to npm packages? There’s opportunity here…
I wrote a runtime in golang that runs typescript (well, JavaScript in general). The grafana folks have sobek that all they need is to add type striping. I feel like if there’s one runtime where typescript could be adopted fully and it would change the world is Node.js. No transpiler, no typescript-go, no rust (well, maybe some rust ;) just a great parser that will keep track of the source map and types in debug mode (for tracing).
Either way, kudos to the node team, contributors, for pulling in the goal posts to make the kick to launch shorter. I’m still a fan of bun, and my own runtime, but node is the standard by which we all are kinda following. I also like that the embedding api is simple and clean to use now so if you want to make an executable, you can.
"Written only in TypeScript" might put it better. If your module ships TypeScript source and a JS build as it should, then this will never affect it. Otherwise, to support stripping arbitrary modules would immediately compromise the design goal of light weight, due to the torrent of ill-founded and -formed bug reports incorrectly raised on Node that would follow. ("Don't make the maintainers' lives too miserable to continue the work" being also of course an implicit goal.)
I switched to deno for new projects ~1 year ago and it’s only been joy. There’s a shockingly small amount of friction to switch over, and there are so so many benefits
I'm glad you've said this. I have a project at nearly a perfect point to try out that cutover. Not that it isn't nice to understand the circa 2018-2022 TS stack, but it sure would be nice not to have to. (Our ancestors had the same discussions about cfront(1). Everything old is new again.)
Last actually note-worthy improvement I heard of was properly supporting import/export (although do you still need to use the .mjs hack?), but I've been out of the loop here for sometime so would be nice to know what they've added since.
Sounds like the obvious correct solution, making .cjs and .mjs obsolete - unless of course someone uses import() statements exclusively, in which case I need to ask: why?
It is surprising for me to see these features finally being added to Node after such a long time. Especially so when I remember reading discussion after discussion about how something like this wasn't possible. I touched on this in a blog post some time ago [1]. Glad Node is catching up.
I don't see in your blogpost any sources cited regarding anyone saying that ES modules were infeasible.
Additionally, io.js actually forked off due to internal drama which started with Ben Noordhuis having changed some pronouns here and there and people wanting to cancel him for that, to which he picked up his toys and left the sandbox.
It so happened that aside from being competent himself, he had competent people on his side, which eventually forced those governing Node.js to concede.
using, memory64, undici, async local storage, ESM import improvements, type stripping, local storage / session storage, env file support, built in file watching. Those are just the ones I mainly remember. There is a lot more.
As a quick aside, “them” is an object pronoun, not a subject pronoun. The correct word you needed is “they”.
You couldn’t phrase your original question as a statement “Them have though.” That’s often a quick test for valid English grammar. With the correct pronoun, it makes more sense: “They have though.”
As another example, take this sentence: “Have you seen them though?”
“You” is the subject of that sentence, and “them” is the object.
It's short for "Have them [Node bozos improved it], though?"
Or, equally likely it, refers to deno and bun ("deno and bun has really made Node focus and improve", "Have them (deno and bun) really made Node focus and improve, though?")
The best thing is that they are shipping this as "type stripping" which means that there are no sourcemaps involved, making it zero-cost in production!
I’m not a heavy JS/TS dev so here’s an honest question: why not use Bun and forget about node? Sure I understand that not every project is evergreen but isn’t Bun a much runtime in general? It supports TS execution from day 1, has much faster dependency resolution, better ergonomics… and I could keep going.
I know I’m just a single data point but I’ve had a lot of success migrating old node projects to bun (in fact I haven’t used node itself since Bun was made public)
Again, I might be saying something terribly stupid because JS/TS isn’t really my turf so please let me know if I’m missing something.
I've mainly worked with Node for now 8 years, and recently switched to Deno. Even that switch was hard to do; not because things don't work, but you don't know when they won't.
Node has its share of flaws, but it's the de facto baseline against which things are tested and developed. I'm somewhat more comfortable working with The Main Thing.
The JavaScript ecosystem is nightmarish enough that many developers don't want to switch to the Next Cool Thing. I think many of us have had enough fatigue caused by new build tools, new bundlers, new runtimes, etc.
As of right now, Bun is not compelling enough for the potential headaches down the line.
(Maybe there won't be any, but I've spent weeks dealing with incompatibilities caused by a single TS minor update (which should've been breaking). Days chasing after dependency problems, after missing docs, etc.)
Bun is still a toddler: it's not ready for primetime.
Simple example: you know how at the command line you can type "npm run", and then type a character or two, hit tab, and the appropriate script from your `package.json` will autocomplete? And if you keep going (eg. "npm run knex") you can do the same thing to autocomplete arguments?
Bun still hasn't figured out how to do that (https://github.com/oven-sh/bun/issues/6037), even though they can all but copy NPM's (already written) completions. I really liked using bun when I played around with it (and it ran my codebase perfectly, without issue) ... but if they can't handle something as simple as Bash completions, they're clearly not ready for the big leagues.
I have tried fully switching to bun repeatedly since it came out and every time I got 90% of the way there only to hit a problem that couldn't be worked around. Last I tried I was still stuck on some libraries requiring napi functions that weren't implemented in bun yet, as well an issue I forget but it was vaguely something like `opendir` silently ignoring the `recursive` option causing a huge headache.
I'm waiting patiently for bun to catch up because I would love to switch but I don't think its ready for production use in larger projects yet. Even when things work, a lot of the bun-specific functionality sounds nice at first but feels like an afterthought in practice, and the documentation is far from the quality of node.js
> I’m not a heavy JS/TS dev so here’s an honest question: why not use Bun and forget about node?
Why would you switch from runtime A to runtime B? I mean, you presented no reason at all, let alone a compelling one, to pick either one. So what leads you to believe it is a reasonable idea to waste time switching runtimes?
For me - it doesn't support secure and reliable dependency vendoring.
The best way to do this atm. is using (and configuring) yarn for zero-installs.
This keeps dependencies inside the codebase so that:
* Issues can be easily traced to the code that actually ran - development and deployment are the same.
* Deployment doesn't depend on package repositories.
* Deployment is secure from many kinds of attacks.
* It is possible to transparently patch packages.
* Development is only internet dependent when adding a new package.
* and the best ease-of-use - no reinstall when changing branches.
I agree. I've tried the Node TS and test runner features, and they are still (not yet) as good as Bun's. So for now sticking with Bun for those.
Really, in the Node ecosystem you eventually learn not to put all your eggs in one basket. Different things excel in different aspects. Here is my preferred setup for now:
Bun.js:
As a Node runtime, and for TS execution and test running. I tried lots: TSX, TS-Node, Node itself
NPM
For executing tooling scripts
PNPM
For installing dependencies. It's simply better than the rest (npm, yarn, bun) for several reasons
Biome.js
For linting (superior to every other tool I tried)
The bun test runner is definitely a lot better for testing than Node's.
But really, any test runner is beter than Node's: that thing is awful. It's like they looked at all the test runners in existence, and instead of copying what they all did, decided "let's make things harder for no apparent reason."
I've been using this for helper scripts, where each script is its own entry point, to assist with various maintenance tasks in frontend projects. Much easier to clone the project and run `./scripts/frobnicate` than faff around with tsc. Previously they would have been written in pure JS or just bash.
It's faster and a good bit more convenient in greenfield, in my experience. Less safe in that you do still need a type checker, but nothing about Node's stripping is at all meant for more than experimental use anyway so that's fine.
It performs no type checking, and you don't need to load the compiler to compile. Tsc is a heavy package and having Node do this for you means much faster startup.
One of Typescript's design goals is that removing all type-related parts of the source text should yield a valid JavaScript file. A typescript compiler does not generate code (unlike, say, PureScript).
You can run a typechecker (such as tsc) that check various properties of your code statically, relying on the type information. It is then erased.
The same applies, say, to Python: type annotations are ignored at runtime. Somehow similarly, Java's type information is also partly erased in the bytecode; in particular, all the information about parametrized types. (This is to say nothing about actual machine code.)
This is true but these are also old features, and the TS team have stated that they will not add any more features like those, and to a certain extent regret adding them initially (particularly decorators, which iirc were added because the Angular framework wanted to use them). You can also see that these features aren't really being updated to match recent Typescript developments (parameter properties can't do true private properties, const enums don't work with isolated modules, etc).
I don't think those features are ever going to go away, because they've been around for so long and are so widely used. But I generally use erasableSyntaxOnly in new projects, because I find it's useful when my typescript source matches the generated Javascript code as much as possible.
How useful is it exactly that you accept to not use DX improving syntax like constructor properties, enums, etc? To me, someone who uses these features _a lot_, this would be a terrible trade. Seems more like people push this out of ideology and because TS is never going to be part of node itself (since its implementation is just way too slow)
> I find it's useful when my typescript source matches the generated Javascript code as much as possible.
Is this that worth? In the past I was able to read past async/await desugaring generated by TypeScript, and there are several useful non-JS syntaxes that are much easier to read than that (e.g. enums). Of course it would be great if ECMAScript eventually adopts them, but that doesn't seem like a strict requirement for me.
Agreed about TS, but Python type annotations are not ignored. They are executed as code (all type annotations are valid expressions) and the results are stored on the module/class/function object that contains the annotated variable
>so at best it saves you a transpilation pass and doesn't improve safety.
That's a bit misleading. Node being able to run TS code does not "improve safety", because that's not where the type checking happens. You can do type checking in your editor, or various other points in your toolchain.
Node being able to run TS code reduces the friction in writing TS code, which indirectly helps with type safety.
Except it doesn't. In anything serious, you have to wait for a full type check to happen before you run your TS code. Why would you run code that has not been checked yet and could throw very weird errors like undefined property access?
That just doesn't make sense. Yes, you can wait for your editor in your current open file, if you are lucky and the change in the open file doesn't break anything downstream in another file that is not yet open. In best case you have such simple code that nothing breaks, and in worst case, you have to still run it with type-checking - on top of running it in type-stripping-mode, because you got weird errors in runtime. This is a net negative.
This whole situation is there because we are trying to workaround the slow TSC. It's not a feature, it's something we actively work around. We try to whitewash now the obviously less useful "solution" of running code without its core features enabled: type checking. To me this is insane.
You can tsc on the code and then ship that git hash if it passes. You don't need to run it every single time the code executes, nothing of value is gained, because nothing has changed.
TypeScript never promised improving safety, maybe it’s a common misconception. But TypeScript has no runtime mode or information. You were always at the mercy of running and not ignoring the typechecker. Nothing stopped you from running ts-node or tsx on code with egregious type errors. TypeScript is more like a linter in that regard.
I think it's not fair to say that Typescript isn't about improving safety, just that the mechanism isn't the same as with other languages. Typescript had always allowed you to ignore the type checker (in fact, the default configuration will always attempt to emit compiled Javascript, even if the source Typescript has type errors). But if you run the type checker on every commit (via e.g. CI or a precommit hook), then you can be sure that the code you release is correctly typed, which will not guarantee it is safe, but makes it more likely.
I agree that it's better to think of Typescript as a linter that needs specialised annotations to work, rather than a type system like you might find in Java or Rust.
What, pray tell, would be the point of putting all that type information in there, and then have it checked (via tsc), if not for the sake of safety? What other use would this have in your opinion?
This is misleading. It is not transpiling TS in JS, it is transpiling a subset of TS into JS. If my normal TS code can not be "executed" by Node, then it is not executing TS per definition but something else. If you are good with Node supporting and "executing" only a subset of TS and lacking useful features, that's fine. But don't tell people it is executing TypeScript. That's like me saying my rudimentary C++ compiler supports C++ while in reality only supporting 50%. People would be pissed if they figure it out once they try to run it on their codebase.
There is always a compile step (JS -> Bytecode -> Machine code). The question is only if it is visible to you or not. They could have made it totally transparent to you by fully support TS including type checking under the hood including support full TS and not this subset of it, but decided not to do so. There is nothing inherently great to have less compile steps if you are not even aware of it. See v8 how many compile and optimizations steps they have - You don't care, because you don't see it. The only problem of TS is, you will always be able to see it because of it being slow.
I think running TS without type checks is almost entirely pointless.
I'm using tsx for a project to achieve the same effect. As you said, it saves you from having to set up a build/transpilation step, which is very useful for development. Tsx has a --watch feature built in as well, which allows me to run a server from the typescript source files and automatically restart on changes. Maybe with nodemon and this new node improvement this can now done without tsx.
To check types at runtime (if that can even be done in a useful way?) it would have to be built into v8, and I suppose that would be a whole rewrite.
Yeah agreed - saying "node can execute typescript files" is a bit misleading. More accurate would be "node can find-and-replace type information with spaces from .ts files and try and executing them as if they were plain JavaScript"
I suspect this would only handle the most rudimentary and basic typescript files. Once you start using the type system more extensively I suspect this will blow-up in your face.
It's kinda a shame. What a missed opportunity to do it properly (rather than relying on third party plugins etc)
Edit: if you are just using typescript for type checking at build time (i.e. the most basic rudimentary typescript files) the sure fine this may help you. But typescript also generates JavaScript code for missing features e.g. proper class access modifiers, generics, interfaces, type aliases, enums, decorators etc etc. typescript generates JS code to handle all that, so you're going to have a bad time if node just replaces it with the space character at run time.
Just to give context here, NodeJS doesnt support enums, namespaces and class parameter properties. All of these have been described as regrets by Anders Hejsberg, and none of them prevent advanced use of the type system at all.
Yes, they regret them because they hinder adoptions. Why? Because nobody chose to add TSC with all features in their runtime because TSC is extremely slow.
They know they can skyrocket adoption by limiting the language. That's the reason they regret it. This is just a strategy to increase adoption. Not because they are bad features. They are in fact very useful, and you should not stop using them just because your favorite runtime decided to go the easy way and only supporting a subset of TS by stripping types. You should rather switch the runtime instead of compromising your codebase.
TypeScript 5.8 added a configuration option[1], where the compiler emits errors when using language features that result in code generation (e.g. enums). It's expected that people wanting to run TypeScript through "erase-only" transpilers will use this option.
Of course, everyone else is free to use enums, decorators, class parameters, etc., but for quick prototyping or writing simple scripts in TypeScript, Bun has been good enough for me, and I assume Node will be "good enough" as well.
> node can find-and-replace type information with spaces from .ts files and try and executing them as if they were plain JavaScript
That’s what all the other tools like ts-node and tsx do already.
I’m not sure what more are you expecting to do?
Typescript is build time type checked, there is no runtime component to TypeScript. If you want type checking you run tsc.
I think this is a great step in the right direction by node. It will save transpiration on the server and improve stack traves and whatnot.
> Once you start using the type system more extensively I suspect this will blow-up in your face.
I don’t see why. There isn’t any more runtime information in “complex”
TypeSceipt types than in simple ones. It’s all build time - see above.
> What a missed opportunity to do it properly
Please explain in more detail what “doing it properly” means to you. Including the typechecker? If so that wouldn’t make sense - they would be competing with TypeScript itself, they shouldn’t, all the “third party plugins” rely on tsc for type checking.
> I think this is a great step in the right direction by node
I think it's the opposite. It will be a net negative, since people will now run TS by default without type checking. Wasting so much time chasing weird runtime errors - just to end up running the full blown TSC type checking again. They will also write very different TS now, trying to workaround the limitation and arguably very useful features like Enums, constructor properties, etc. This has real negative effects on your codebase if you rely on these, just because Node chose to support only a subset.
It's interesting to see the strategy now and to see people even gaslighting people into believing no type checks and less features is a good thing. All just because of one root cause - TSC being extremely slow.
I have been using this feature to remove the transpilation step during development for a rather large monorepo that makes extensive use of very complex types and haven’t noticed any errors whatsoever at runtime.
You’re downplaying this quite a bit. Node being able to execute TS files slashes a lot of tooling in half, especially during development when you want to modify files and retry in quick succession.
Besides, I’m not so sure this cannot be expanded in the future to adopt validation features before stripping the type information. For now it solves some major pain points for a lot of people.
It's not able to execute TypeScript, but a subset of it. The claim in the title is misleading if not totally wrong.
This will unfortunately drive people towards using TS only as a linter, and not use its powerful features that are inherently impossible to implement with just type stripping.
I guess decorators. Which are quite powerful but I’ve never seen them used outside of libraries/frameworks that use them (angular, certain ORMs, nestjs?)
As a personal taste I don’t really like decorators that much, but it’s true that nestjs projects (which is probably a majority of new backend TS projects) will not gain anything from this release. Then again, you always set nestjs up with a template anyway that has all of the tooling and building baked in. So whatevs.
It’s still a huge huge win, and I finally have hope for typescript-ifying some horrible legacy node apps at work!!
That is the real reason we get a less feature-rich TypeScript in the future and Node not supporting full TypeScript. Because they want to be supported by browser.
Next is where it's at these days in my opinion. You get a full-featured client-side React framework (the only one that supports modern React SSG), and then on top of it you get a better-organized approach to doing everything you can do with Express.
And do mean everything: I run an entire Postrgraphile server through Next (and you can easily do the sme with Supabase or a similar tool)!
Express is still popular, but a lot projects these days use a full-stack framework like Next.js, SvelteKit, etc.
Fastify, NestJS (bleh), Koa, Hono are the modern replacements for express, none of them have caught on as a standard though. My personal favorite for small projects is Polka (https://github.com/lukeed/polka), when I'm not using Go instead.
IMHO the only reason you are using JS in the backend is because of some meta framework, otherwise is not worth. So at least for Nuxt is nitro, not sure for SvelteKit or the other React meta frameworks.
Is there any movement in the node world for the runtime to be able to ascertain types? Obviously nodejs doesn't do this, but does Bun or Deno or are there plans for any of this in the future?
This has nothing to do with the "node world". Such an enormous feature would have to go into ECMAScript. Which is very, very unlikely to ever happen, they may as well design a new language. All those runtimes implement that spec. Expecting them to write an extremely complex new feature that is easily more complicated than everything already implemented (especially with backwards compatibility, and given that the language was not designed for this) would be a bit much.
TypeScript is for compile time checking of a language that was not designed to have them. Runtime types have very different requirements! It has to be in the language from the early design phase, otherwise it will just be a hack with many conditions, restrictions and holes.
TS Types are only partially a description of the underlying types in the code, a very big part instead is that it provides guard rails that prevent you from using a lot of perfectly fine and valid JS code that would however be incompatible with type guarantees. You pay the price of using only a part of the large space of JS code possibilities for guarantees. If you were to put that into the runtime you would end up with two different versions of the language. If you still want to support the full JS you would end with two runtimes in one (or one that has so many branches and conditions that maintaining that runtime is a real beast).
Now of course, this would only add type syntax to the language, not true processing: there's nothing in the spec about actually handling them. Still, it's a step in that direction, so I wouldn't say "very unlikely to ever happen" ... "still a long ways off (if ever)" would be more accurate.
Can't be too hard imo. Primitives can contain an extra metadata field for the types defined in the code. This doesn't really interfere with the runtime and will be backwards compatible. The runtime of course still doesn't actually have to do any type checking, it's just forwarding the type information on a new metadata field so it's not in any way interrupting the core flow of the runtime logic. As a result you can still have the wrong type tagged onto a primitive just like TS has it currently.
Only downside I see is that It can slow down the code as the runtime now has to evaluate type level functions in order to know what to place in the metadata.
But you're right in the sense that it has to go into the core ECMA specification rather then being a node project.
i don’t think you’ll see this at the typescript level, and you won’t see anything like this that compiles to javascript. specifically, compiling typescript to javascript with runtime checks would not actually be very useful.
typescript is pretty ambiguous about a lot of the things that would need explicit definition for runtime safety, and anyways we already have tools for that - it’s called zod.
and comprehensive checks would incur a significant runtime penalty, unless they were restricted to external interfaces, which is what you’re really concerned about. we already have tools for that - protobuf, swagger, etc.
anything else is sharing a runtime with you. so either it’s in your ide, and you just don’t write shitty code; or you’re trapped in some kind of demonic javascript prisoner’s dilemma, and you are mutable.
so typescript is basically ‘good enough’ for developers.
thinking forward anyway, and assuming you’re really willing to share a runtime with a stranger…
node doesn’t really operate in that kind of context, but maybe browser code does. i could imagine a framework based on web components, workers, and maybe iframes, taking advantage of message boundaries to enhance analysis and conceal code generation. it’s not that much better than typescript.
but if you want efficient runtime checks, and you want to leverage static analysis and strong module boundaries to scope the type-checking codegen, and you probably need additional syntax, you might as well target wasm.
Yeah, but Zod and all of the runtime schema libraries all kind of add verbosity to the type system compared to say something like Typia[0] which AOT compiles the type checks (and ends up being way more elegant).
Caveat is that there are some restrictions with the compiler and some possible footguns (duplicated declarations bloating code).
I believe it would be too hard to keep up with pace of TypeScript development. We should probably at some point formally define the system and allow for alternative implementations outside of the control of Microsoft.
Contrary to the popular trend, I usually stick with Node and NPM for most of my projects. That said, I’ve run into plenty of headaches with CommonJS vs ESM quirks. Sharing code between frontend and backend (via shared libraries) still feels messy because of those differences. In one project I actually had to switch to Bun as the runtime, since Node kept failing at runtime, and strangely enough, Bun just worked without issues.
For one I do not want to run my startup at the mercy of VC funded tech. Node.js is open source and maintained by a foundation, it will not "run out of funds" or be abandoned if there is no profitability in the near future.
Yeah, but it's not a big bet. Deno can do it, Bun can do it, if they die a tragic VC-fueled death then somebody else (maybe Node) can do it. Using Bun to me is just like using a microwave oven in 1980 — there weren't a lot of microwavable convenience foods yet, but you could sure heat up some leftovers more conveniently and quickly.
I think this + node:test makes Node.js a pretty compelling sensible default for most things now. Running things with `tsx` was such a QoL improvement when it happened, but it didn't solve everything.
Runtime type assertion at the edges is mostly solved through `zod` and tools like `ts-rest` and `trpc` makes it so much easier to do full-stack Typescript these days.
This. It's 2025 and the node ecosystem is finally usable by default!
ESM modules just work with both Node and Typescript, Node can run .ts files, and there's the a good enough test runner built in. --watch. The better built in packages - `node:fs/promises` - are nice with top-level await for easier async loops.
It took a while to convince everyone involved to just be pragmatic, but it's nice now.
What's the story with supporting CommonJS libraries? I've tried to update many projects to ESM multiple times over the years, and every time, I ended up backing out because it turned out that there was some important upstream library that was still CommonJS - or even if we fixed those issues, our downstream NPM consumers wouldn't be able to consume EJS. So then you have to go down this rabbit hole of dual compilation, which actually means using something other than tsc.
It's possible, but it can be weird and difficult: https://nodejs.org/docs/latest-v17.x/api/esm.html#esm_common...
Thankfully, actively-maintained CommonJS-only packages are quite rare by this point (in my experience).
> our downstream NPM consumers wouldn't be able to consume EJS
Node.js 20.17 and later supports loading ESM using `require()`: https://nodejs.org/api/modules.html#loading-ecmascript-modul...
The next version of Babel (currently in beta) is even going ESM-only.
It's almost like Python 2 to Python 3 upgrade. Took a decade but everyone is finally happy.
This is great to hear, but perhaps comes too late for people like myself. Node.js has been by go-to platform from around 2014 until last year. But around September last year, I found myself thrust into the .NET ecosystem (due to a client project). Within a few months, I realized that it too, had finally become usable by default (unlike the last time I tried it, when it was too tightly coupled to Windows). In fact, it felt like what Node.js would be, if it had strong typing built-in, and had a good standard library that eliminated a lot of the module management and churn. I'm now finding it hard to return to Node.js.
Interesting. I haven't looked hard at .Net despite some advocacy from past colleagues. Perhaps I should.
does it have a go fmt / lint command yet?
does that require a config?
depends, will you keep finding pendatic faults after any answer?
Having a typo in "pendantic" is a masterstroke
> "pendantic"
Genius!
One more time and it's pentadic. But imagine someone finding a way to bikeshed anti-bikeshedding tools. "Take with food."
No, it has good defaults. See also: https://prettier.io/docs/option-philosophy
jslint/tslint are an install away.
werent one of the js linters part of a supply chain attack recently?
Maybe, are you sure Go dependencies are immune to similar attacks?
Yes, with the difference that Google would have to be compromised in order to poison the go distributable containing fmt tool. With js, it’s enough to poison any single one of the 1400 dependencies of the linter
I forgot that even though fmt will never suffer from middle man attacks downloading the Go toolchain, the standard library already covers 100% of the uses cases someone cares about using Go for, and no one is using CGO.
Use biome, it doesn't have any external dependencies. eslint should have been put to rest a long time ago.
Good advice. That was my conclusion as well after years of fighting with eslint.
Nope
I'm very much in favor of TS support directly in node. vitest has made it easier these days, but I've lost too much time over the years getting the balance just right when configuring test environments for .ts files.
trpc and ts-rest are a different animal in my opinion. I'm happy to use either one but won't deal with them in production. For trpc that's mainly due to the lack of owning API URLs and being able to more clearly manage deprecating old URLs gracefully.
For ts-rest I just tend to prefer owning that setup myself, usually with zod and shared typings for API request/response pairs. It also does irk me every time I import what is clearly an RPC tool named "-rest"
vitest is incredible; it makes one wonder how/why jest, with its larger user base and community, couldn't get its TS support sorted.
Let's see if Sveltr converts their codebase back to TypeScript
Does this run tsx? Even with the types stripped you still need the JSX transformed to JavaScript
It doesn’t. The comment you’re replying to is referring to tsx, the package that lets you execute ts files, not to running files with the tsx extension.
https://github.com/privatenumber/tsx
i switched to python a while ago. it has batteries included. i feel so much better now that i dont have to debug all the quirks of a half-baked system.
This is great up until you get to the fact that typescript will not be accepted under node_modules [0].
That leads me to ask, what about project dependencies? I wrote a lib for my data models in typescript and I want to import that into my app in node, in typescript? Does the rule only apply to npm packages? There’s opportunity here…
I wrote a runtime in golang that runs typescript (well, JavaScript in general). The grafana folks have sobek that all they need is to add type striping. I feel like if there’s one runtime where typescript could be adopted fully and it would change the world is Node.js. No transpiler, no typescript-go, no rust (well, maybe some rust ;) just a great parser that will keep track of the source map and types in debug mode (for tracing).
Either way, kudos to the node team, contributors, for pulling in the goal posts to make the kick to launch shorter. I’m still a fan of bun, and my own runtime, but node is the standard by which we all are kinda following. I also like that the embedding api is simple and clean to use now so if you want to make an executable, you can.
[0] https://nodejs.org/api/typescript.html#type-stripping-in-dep...
I made the same comment here https://news.ycombinator.com/item?id=44931575
"To discourage package authors from publishing packages written in TypeScript"
I tried to use it with private packages but that doesn't work either, apparently node doesn't even read the "private" field.
"Written only in TypeScript" might put it better. If your module ships TypeScript source and a JS build as it should, then this will never affect it. Otherwise, to support stripping arbitrary modules would immediately compromise the design goal of light weight, due to the torrent of ill-founded and -formed bug reports incorrectly raised on Node that would follow. ("Don't make the maintainers' lives too miserable to continue the work" being also of course an implicit goal.)
Why would I want to ship a JS build for my private package? That's just extra machinery I don't need. Switching to a superior runtime would be easier.
It’s a missed opportunity for sure
https://github.com/nodejs/node/issues/57215
Not supporting type stripping in node_modules is unfortunate
But... that's like half the reason why I wanted this feature...
Writing a library in TypeScript (with typechecks in CI/CD as devDependencies) and just importing it directly from Node.js...
It was the first thing I tried and of course it didn't work.
It might finally be time to switch to Deno or Bun =(
I switched to deno for new projects ~1 year ago and it’s only been joy. There’s a shockingly small amount of friction to switch over, and there are so so many benefits
I'm glad you've said this. I have a project at nearly a perfect point to try out that cutover. Not that it isn't nice to understand the circa 2018-2022 TS stack, but it sure would be nice not to have to. (Our ancestors had the same discussions about cfront(1). Everything old is new again.)
Bun has been awesome for me and my team fwiw
Impressed with what Node is doing the last years, deno and bun has really made Node focus and improve. It was stuck for a while
What are the recent improvements in node itself?
Last actually note-worthy improvement I heard of was properly supporting import/export (although do you still need to use the .mjs hack?), but I've been out of the loop here for sometime so would be nice to know what they've added since.
Here’s a nice overview:
https://kashw1n.com/blog/nodejs-2025/
It doesn’t cover everything, but as an old-school Node user I found several interesting features I didn’t know about.
That is a nice article. It does a great job summing thing up with real examples too.
Small but lovely addition for me is the ability to load .env files natively. There’s more like this; small, focused, real-world-improving features.
> (although do you still need to use the .mjs hack?)
Syntax detection is enabled by default in v22.7.0, v20.19.0:
https://nodejs.org/api/packages.html#syntax-detection
Sounds like the obvious correct solution, making .cjs and .mjs obsolete - unless of course someone uses import() statements exclusively, in which case I need to ask: why?
It is surprising for me to see these features finally being added to Node after such a long time. Especially so when I remember reading discussion after discussion about how something like this wasn't possible. I touched on this in a blog post some time ago [1]. Glad Node is catching up.
[1] https://kilo.bytesize.xyz/an-incorrect-specification
I don't see in your blogpost any sources cited regarding anyone saying that ES modules were infeasible.
Additionally, io.js actually forked off due to internal drama which started with Ben Noordhuis having changed some pronouns here and there and people wanting to cancel him for that, to which he picked up his toys and left the sandbox.
It so happened that aside from being competent himself, he had competent people on his side, which eventually forced those governing Node.js to concede.
Bun is just a cash grab in comparison.
And for a few earlier versions `type: module` in package.json was all that was needed for .js files to be treated as ESM.
using, memory64, undici, async local storage, ESM import improvements, type stripping, local storage / session storage, env file support, built in file watching. Those are just the ones I mainly remember. There is a lot more.
Adding to the list: permissions, CLI styling/colouring, require(esm), globs, test runner.
Do you mean.. node-worthy?
Have them though?
On the projects I am involved, they could even not exist, only node LTS releases matter, and the most recent projects are still node 20.
As a quick aside, “them” is an object pronoun, not a subject pronoun. The correct word you needed is “they”.
You couldn’t phrase your original question as a statement “Them have though.” That’s often a quick test for valid English grammar. With the correct pronoun, it makes more sense: “They have though.”
As another example, take this sentence: “Have you seen them though?”
“You” is the subject of that sentence, and “them” is the object.
Them is fine.
It's short for "Have them [Node bozos improved it], though?"
Or, equally likely it, refers to deno and bun ("deno and bun has really made Node focus and improve", "Have them (deno and bun) really made Node focus and improve, though?")
That is nonstandard English, at best. It's found in some uncommon dialects.
Without the expansion I don't know of any native English speaker who would say it.
Your expanded version is also incorrect.
22 is LTS. The future is now.
All even versions are LTS btw, maybe what you mean is that version 22 entered in maintenance mode (hence stable) and new features will not be added.
Sure, who is going to budget project upgrade effort of ensuring all dependencies work equally as well?
There is a reason why so many Java, Python, .NET/C#, C, C++,.. projects are stuck several versions behind.
No one said developing software wasn't going to be work.
Work isn't free beer in most places.
The best thing is that they are shipping this as "type stripping" which means that there are no sourcemaps involved, making it zero-cost in production!
Very well done Node team!
I’m not a heavy JS/TS dev so here’s an honest question: why not use Bun and forget about node? Sure I understand that not every project is evergreen but isn’t Bun a much runtime in general? It supports TS execution from day 1, has much faster dependency resolution, better ergonomics… and I could keep going.
I know I’m just a single data point but I’ve had a lot of success migrating old node projects to bun (in fact I haven’t used node itself since Bun was made public)
Again, I might be saying something terribly stupid because JS/TS isn’t really my turf so please let me know if I’m missing something.
I've mainly worked with Node for now 8 years, and recently switched to Deno. Even that switch was hard to do; not because things don't work, but you don't know when they won't.
Node has its share of flaws, but it's the de facto baseline against which things are tested and developed. I'm somewhat more comfortable working with The Main Thing.
The JavaScript ecosystem is nightmarish enough that many developers don't want to switch to the Next Cool Thing. I think many of us have had enough fatigue caused by new build tools, new bundlers, new runtimes, etc.
As of right now, Bun is not compelling enough for the potential headaches down the line.
(Maybe there won't be any, but I've spent weeks dealing with incompatibilities caused by a single TS minor update (which should've been breaking). Days chasing after dependency problems, after missing docs, etc.)
Bun is still a toddler: it's not ready for primetime.
Simple example: you know how at the command line you can type "npm run", and then type a character or two, hit tab, and the appropriate script from your `package.json` will autocomplete? And if you keep going (eg. "npm run knex") you can do the same thing to autocomplete arguments?
Bun still hasn't figured out how to do that (https://github.com/oven-sh/bun/issues/6037), even though they can all but copy NPM's (already written) completions. I really liked using bun when I played around with it (and it ran my codebase perfectly, without issue) ... but if they can't handle something as simple as Bash completions, they're clearly not ready for the big leagues.
I have tried fully switching to bun repeatedly since it came out and every time I got 90% of the way there only to hit a problem that couldn't be worked around. Last I tried I was still stuck on some libraries requiring napi functions that weren't implemented in bun yet, as well an issue I forget but it was vaguely something like `opendir` silently ignoring the `recursive` option causing a huge headache.
I'm waiting patiently for bun to catch up because I would love to switch but I don't think its ready for production use in larger projects yet. Even when things work, a lot of the bun-specific functionality sounds nice at first but feels like an afterthought in practice, and the documentation is far from the quality of node.js
> I’m not a heavy JS/TS dev so here’s an honest question: why not use Bun and forget about node?
Why would you switch from runtime A to runtime B? I mean, you presented no reason at all, let alone a compelling one, to pick either one. So what leads you to believe it is a reasonable idea to waste time switching runtimes?
For me - it doesn't support secure and reliable dependency vendoring.
The best way to do this atm. is using (and configuring) yarn for zero-installs.
This keeps dependencies inside the codebase so that: * Issues can be easily traced to the code that actually ran - development and deployment are the same. * Deployment doesn't depend on package repositories. * Deployment is secure from many kinds of attacks. * It is possible to transparently patch packages. * Development is only internet dependent when adding a new package. * and the best ease-of-use - no reinstall when changing branches.
I tried to replace Node with Bun, but had the following compatibility problems.
localAddress on TCP connections ignored, last time I tried it its no-op
Incompatibility with Node module APIs (https://github.com/spamscanner/spamscanner wouldn't work)
EventEmitter race problems (partially worked around with https://www.npmjs.com/package/eventemitter2)
Svelte vites dev server sometimes forever freeze until I wiped node_modules and reinstalled it.
I agree. I've tried the Node TS and test runner features, and they are still (not yet) as good as Bun's. So for now sticking with Bun for those.
Really, in the Node ecosystem you eventually learn not to put all your eggs in one basket. Different things excel in different aspects. Here is my preferred setup for now:
Bun.js: As a Node runtime, and for TS execution and test running. I tried lots: TSX, TS-Node, Node itself
NPM For executing tooling scripts
PNPM For installing dependencies. It's simply better than the rest (npm, yarn, bun) for several reasons
Biome.js For linting (superior to every other tool I tried)
The bun test runner is definitely a lot better for testing than Node's.
But really, any test runner is beter than Node's: that thing is awful. It's like they looked at all the test runners in existence, and instead of copying what they all did, decided "let's make things harder for no apparent reason."
I’m curious what the benefit of stuff like this is vs tsc --watch and running the JS?
I’ve always just run tsc to a .gitignored’d directory and execute my JS from there.
Edit: Thanks for the responses. There’s some great examples in there!
I've been using this for helper scripts, where each script is its own entry point, to assist with various maintenance tasks in frontend projects. Much easier to clone the project and run `./scripts/frobnicate` than faff around with tsc. Previously they would have been written in pure JS or just bash.
It's faster and a good bit more convenient in greenfield, in my experience. Less safe in that you do still need a type checker, but nothing about Node's stripping is at all meant for more than experimental use anyway so that's fine.
It performs no type checking, and you don't need to load the compiler to compile. Tsc is a heavy package and having Node do this for you means much faster startup.
As a note, Node has a built in --watch now, too
It looks like this works by stripping away the type information, so at best it saves you a transpilation pass and doesn't improve safety.
One of Typescript's design goals is that removing all type-related parts of the source text should yield a valid JavaScript file. A typescript compiler does not generate code (unlike, say, PureScript).
You can run a typechecker (such as tsc) that check various properties of your code statically, relying on the type information. It is then erased.
The same applies, say, to Python: type annotations are ignored at runtime. Somehow similarly, Java's type information is also partly erased in the bytecode; in particular, all the information about parametrized types. (This is to say nothing about actual machine code.)
> A typescript compiler does not generate code
Except for where it does: Enums, namespaces, parameter properties, etc.
This is true but these are also old features, and the TS team have stated that they will not add any more features like those, and to a certain extent regret adding them initially (particularly decorators, which iirc were added because the Angular framework wanted to use them). You can also see that these features aren't really being updated to match recent Typescript developments (parameter properties can't do true private properties, const enums don't work with isolated modules, etc).
I don't think those features are ever going to go away, because they've been around for so long and are so widely used. But I generally use erasableSyntaxOnly in new projects, because I find it's useful when my typescript source matches the generated Javascript code as much as possible.
> because I find it's useful
How useful is it exactly that you accept to not use DX improving syntax like constructor properties, enums, etc? To me, someone who uses these features _a lot_, this would be a terrible trade. Seems more like people push this out of ideology and because TS is never going to be part of node itself (since its implementation is just way too slow)
> I find it's useful when my typescript source matches the generated Javascript code as much as possible.
Is this that worth? In the past I was able to read past async/await desugaring generated by TypeScript, and there are several useful non-JS syntaxes that are much easier to read than that (e.g. enums). Of course it would be great if ECMAScript eventually adopts them, but that doesn't seem like a strict requirement for me.
Just use erasableSyntaxOnly = true and you're fine.
Hadn't heard of that. Thanks
Agreed about TS, but Python type annotations are not ignored. They are executed as code (all type annotations are valid expressions) and the results are stored on the module/class/function object that contains the annotated variable
Except in Python you can easily access the type hints at runtime, which allowed people to build ergonomic libraries such as Pydantic
> Somehow similarly, Java's type information is also partly erased in the bytecode;
Just for generics, I believe
>so at best it saves you a transpilation pass and doesn't improve safety.
That's a bit misleading. Node being able to run TS code does not "improve safety", because that's not where the type checking happens. You can do type checking in your editor, or various other points in your toolchain.
Node being able to run TS code reduces the friction in writing TS code, which indirectly helps with type safety.
Except it doesn't. In anything serious, you have to wait for a full type check to happen before you run your TS code. Why would you run code that has not been checked yet and could throw very weird errors like undefined property access?
That just doesn't make sense. Yes, you can wait for your editor in your current open file, if you are lucky and the change in the open file doesn't break anything downstream in another file that is not yet open. In best case you have such simple code that nothing breaks, and in worst case, you have to still run it with type-checking - on top of running it in type-stripping-mode, because you got weird errors in runtime. This is a net negative.
This whole situation is there because we are trying to workaround the slow TSC. It's not a feature, it's something we actively work around. We try to whitewash now the obviously less useful "solution" of running code without its core features enabled: type checking. To me this is insane.
You can tsc on the code and then ship that git hash if it passes. You don't need to run it every single time the code executes, nothing of value is gained, because nothing has changed.
TypeScript never promised improving safety, maybe it’s a common misconception. But TypeScript has no runtime mode or information. You were always at the mercy of running and not ignoring the typechecker. Nothing stopped you from running ts-node or tsx on code with egregious type errors. TypeScript is more like a linter in that regard.
I think it's not fair to say that Typescript isn't about improving safety, just that the mechanism isn't the same as with other languages. Typescript had always allowed you to ignore the type checker (in fact, the default configuration will always attempt to emit compiled Javascript, even if the source Typescript has type errors). But if you run the type checker on every commit (via e.g. CI or a precommit hook), then you can be sure that the code you release is correctly typed, which will not guarantee it is safe, but makes it more likely.
I agree that it's better to think of Typescript as a linter that needs specialised annotations to work, rather than a type system like you might find in Java or Rust.
> TypeScript never promised improving safety
What, pray tell, would be the point of putting all that type information in there, and then have it checked (via tsc), if not for the sake of safety? What other use would this have in your opinion?
> TypeScript is more like a linter
that's exactly the point--GP is pointing out that node can't do that part
Which is why the title is "Node can now execute Typescript files" and not lint, check, or even run TypeScript files.
I'm not sure what the distinction between "execute" and "run" is; is there a difference?
Node is just inline transpiling the TS into JS, then running the JS.
This is misleading. It is not transpiling TS in JS, it is transpiling a subset of TS into JS. If my normal TS code can not be "executed" by Node, then it is not executing TS per definition but something else. If you are good with Node supporting and "executing" only a subset of TS and lacking useful features, that's fine. But don't tell people it is executing TypeScript. That's like me saying my rudimentary C++ compiler supports C++ while in reality only supporting 50%. People would be pissed if they figure it out once they try to run it on their codebase.
Not needing a separate compile step just to run some script sounds great to me. I will run tsc if I want a type check.
There is always a compile step (JS -> Bytecode -> Machine code). The question is only if it is visible to you or not. They could have made it totally transparent to you by fully support TS including type checking under the hood including support full TS and not this subset of it, but decided not to do so. There is nothing inherently great to have less compile steps if you are not even aware of it. See v8 how many compile and optimizations steps they have - You don't care, because you don't see it. The only problem of TS is, you will always be able to see it because of it being slow.
I think running TS without type checks is almost entirely pointless.
I'm using tsx for a project to achieve the same effect. As you said, it saves you from having to set up a build/transpilation step, which is very useful for development. Tsx has a --watch feature built in as well, which allows me to run a server from the typescript source files and automatically restart on changes. Maybe with nodemon and this new node improvement this can now done without tsx.
To check types at runtime (if that can even be done in a useful way?) it would have to be built into v8, and I suppose that would be a whole rewrite.
Node has had a built-in --watch flag for a while too:
https://nodejs.org/docs/latest/api/cli.html#--watch
Surprisingly this gone unheralded.
I agree, that said if the main reason people use TypeScript is security they should use a decent programming language instead.
Yes, it can't parse enums, for example.
It is available behind an experimental flag: --experimental-transform-types
Exactly! Type checking can be long and costly.
Ocaml does it fast
Typescript mission is to strip itself of typing, what’s your point? Typescript will never be Java, you’ll never have runtime reflection.
Yeah agreed - saying "node can execute typescript files" is a bit misleading. More accurate would be "node can find-and-replace type information with spaces from .ts files and try and executing them as if they were plain JavaScript"
I suspect this would only handle the most rudimentary and basic typescript files. Once you start using the type system more extensively I suspect this will blow-up in your face.
It's kinda a shame. What a missed opportunity to do it properly (rather than relying on third party plugins etc)
Edit: if you are just using typescript for type checking at build time (i.e. the most basic rudimentary typescript files) the sure fine this may help you. But typescript also generates JavaScript code for missing features e.g. proper class access modifiers, generics, interfaces, type aliases, enums, decorators etc etc. typescript generates JS code to handle all that, so you're going to have a bad time if node just replaces it with the space character at run time.
Just to give context here, NodeJS doesnt support enums, namespaces and class parameter properties. All of these have been described as regrets by Anders Hejsberg, and none of them prevent advanced use of the type system at all.
Source: 49m 43s https://m.youtube.com/watch?v=NrEW7F2WCNA
Yes, they regret them because they hinder adoptions. Why? Because nobody chose to add TSC with all features in their runtime because TSC is extremely slow.
They know they can skyrocket adoption by limiting the language. That's the reason they regret it. This is just a strategy to increase adoption. Not because they are bad features. They are in fact very useful, and you should not stop using them just because your favorite runtime decided to go the easy way and only supporting a subset of TS by stripping types. You should rather switch the runtime instead of compromising your codebase.
TypeScript 5.8 added a configuration option[1], where the compiler emits errors when using language features that result in code generation (e.g. enums). It's expected that people wanting to run TypeScript through "erase-only" transpilers will use this option.
Of course, everyone else is free to use enums, decorators, class parameters, etc., but for quick prototyping or writing simple scripts in TypeScript, Bun has been good enough for me, and I assume Node will be "good enough" as well.
[1] https://www.typescriptlang.org/docs/handbook/release-notes/t...
> node can find-and-replace type information with spaces from .ts files and try and executing them as if they were plain JavaScript
That’s what all the other tools like ts-node and tsx do already.
I’m not sure what more are you expecting to do?
Typescript is build time type checked, there is no runtime component to TypeScript. If you want type checking you run tsc.
I think this is a great step in the right direction by node. It will save transpiration on the server and improve stack traves and whatnot.
> Once you start using the type system more extensively I suspect this will blow-up in your face.
I don’t see why. There isn’t any more runtime information in “complex” TypeSceipt types than in simple ones. It’s all build time - see above.
> What a missed opportunity to do it properly
Please explain in more detail what “doing it properly” means to you. Including the typechecker? If so that wouldn’t make sense - they would be competing with TypeScript itself, they shouldn’t, all the “third party plugins” rely on tsc for type checking.
> I think this is a great step in the right direction by node
I think it's the opposite. It will be a net negative, since people will now run TS by default without type checking. Wasting so much time chasing weird runtime errors - just to end up running the full blown TSC type checking again. They will also write very different TS now, trying to workaround the limitation and arguably very useful features like Enums, constructor properties, etc. This has real negative effects on your codebase if you rely on these, just because Node chose to support only a subset.
It's interesting to see the strategy now and to see people even gaslighting people into believing no type checks and less features is a good thing. All just because of one root cause - TSC being extremely slow.
> Typescript is build time type checked, there is no runtime component to TypeScript.
Not exactly. Typescript enums and decorators are examples.
I have been using this feature to remove the transpilation step during development for a rather large monorepo that makes extensive use of very complex types and haven’t noticed any errors whatsoever at runtime.
You’re downplaying this quite a bit. Node being able to execute TS files slashes a lot of tooling in half, especially during development when you want to modify files and retry in quick succession.
Besides, I’m not so sure this cannot be expanded in the future to adopt validation features before stripping the type information. For now it solves some major pain points for a lot of people.
But still, bun is the winner here. I recently started using bun and now not moving to node.js again. It just works.
Node still has a better repl, but I do agree. Bun is really good. It’s my goto, too.
Even though this should be the way to go, it still leaves a bitter taste.
Remember ESM fiasco? Now we have a few years of that all over again, this time with different versions of TypeScript, settings, and tsgo. Good news!
I've tried it, still doesn't work as well as Bun's Typescript and test execution capabilities. So basically just went back to Bun.
I still want enums
It's not able to execute TypeScript, but a subset of it. The claim in the title is misleading if not totally wrong.
This will unfortunately drive people towards using TS only as a linter, and not use its powerful features that are inherently impossible to implement with just type stripping.
I thought TS abandoned stuff that can't just be stripped. Besides enum, what do you use that isn't strippable?
I guess decorators. Which are quite powerful but I’ve never seen them used outside of libraries/frameworks that use them (angular, certain ORMs, nestjs?)
As a personal taste I don’t really like decorators that much, but it’s true that nestjs projects (which is probably a majority of new backend TS projects) will not gain anything from this release. Then again, you always set nestjs up with a template anyway that has all of the tooling and building baked in. So whatevs.
It’s still a huge huge win, and I finally have hope for typescript-ifying some horrible legacy node apps at work!!
I wish browsers also supported directly running typescript files.
Someday! https://github.com/tc39/proposal-type-annotations
That is the real reason we get a less feature-rich TypeScript in the future and Node not supporting full TypeScript. Because they want to be supported by browser.
Already a proposal exists, recent news
That proposal has existed for 4 years, it hasn't gone anywhere yet.
What backend framework is the go to these days? Still Express?
Next is where it's at these days in my opinion. You get a full-featured client-side React framework (the only one that supports modern React SSG), and then on top of it you get a better-organized approach to doing everything you can do with Express.
And do mean everything: I run an entire Postrgraphile server through Next (and you can easily do the sme with Supabase or a similar tool)!
Express is still popular, but a lot projects these days use a full-stack framework like Next.js, SvelteKit, etc.
Fastify, NestJS (bleh), Koa, Hono are the modern replacements for express, none of them have caught on as a standard though. My personal favorite for small projects is Polka (https://github.com/lukeed/polka), when I'm not using Go instead.
IMHO the only reason you are using JS in the backend is because of some meta framework, otherwise is not worth. So at least for Nuxt is nitro, not sure for SvelteKit or the other React meta frameworks.
Wooooo! Finally! <3
does it support watch mode? does it support path aliases?
Is there any movement in the node world for the runtime to be able to ascertain types? Obviously nodejs doesn't do this, but does Bun or Deno or are there plans for any of this in the future?
This has nothing to do with the "node world". Such an enormous feature would have to go into ECMAScript. Which is very, very unlikely to ever happen, they may as well design a new language. All those runtimes implement that spec. Expecting them to write an extremely complex new feature that is easily more complicated than everything already implemented (especially with backwards compatibility, and given that the language was not designed for this) would be a bit much.
TypeScript is for compile time checking of a language that was not designed to have them. Runtime types have very different requirements! It has to be in the language from the early design phase, otherwise it will just be a hack with many conditions, restrictions and holes.
TS Types are only partially a description of the underlying types in the code, a very big part instead is that it provides guard rails that prevent you from using a lot of perfectly fine and valid JS code that would however be incompatible with type guarantees. You pay the price of using only a part of the large space of JS code possibilities for guarantees. If you were to put that into the runtime you would end up with two different versions of the language. If you still want to support the full JS you would end with two runtimes in one (or one that has so many branches and conditions that maintaining that runtime is a real beast).
ECMAScript community is actually dipping their toes in the water of adding typing to the spec, and it's at Stage 1: https://github.com/tc39/proposal-type-annotations.
Now of course, this would only add type syntax to the language, not true processing: there's nothing in the spec about actually handling them. Still, it's a step in that direction, so I wouldn't say "very unlikely to ever happen" ... "still a long ways off (if ever)" would be more accurate.
Can't be too hard imo. Primitives can contain an extra metadata field for the types defined in the code. This doesn't really interfere with the runtime and will be backwards compatible. The runtime of course still doesn't actually have to do any type checking, it's just forwarding the type information on a new metadata field so it's not in any way interrupting the core flow of the runtime logic. As a result you can still have the wrong type tagged onto a primitive just like TS has it currently.
Only downside I see is that It can slow down the code as the runtime now has to evaluate type level functions in order to know what to place in the metadata.
But you're right in the sense that it has to go into the core ECMA specification rather then being a node project.
i don’t think you’ll see this at the typescript level, and you won’t see anything like this that compiles to javascript. specifically, compiling typescript to javascript with runtime checks would not actually be very useful.
typescript is pretty ambiguous about a lot of the things that would need explicit definition for runtime safety, and anyways we already have tools for that - it’s called zod.
and comprehensive checks would incur a significant runtime penalty, unless they were restricted to external interfaces, which is what you’re really concerned about. we already have tools for that - protobuf, swagger, etc.
anything else is sharing a runtime with you. so either it’s in your ide, and you just don’t write shitty code; or you’re trapped in some kind of demonic javascript prisoner’s dilemma, and you are mutable.
so typescript is basically ‘good enough’ for developers.
thinking forward anyway, and assuming you’re really willing to share a runtime with a stranger…
node doesn’t really operate in that kind of context, but maybe browser code does. i could imagine a framework based on web components, workers, and maybe iframes, taking advantage of message boundaries to enhance analysis and conceal code generation. it’s not that much better than typescript.
but if you want efficient runtime checks, and you want to leverage static analysis and strong module boundaries to scope the type-checking codegen, and you probably need additional syntax, you might as well target wasm.
> you won’t see anything like this that compiles to javascript
https://github.com/microsoft/TypeScript/issues/47658
Yeah, but Zod and all of the runtime schema libraries all kind of add verbosity to the type system compared to say something like Typia[0] which AOT compiles the type checks (and ends up being way more elegant).
Caveat is that there are some restrictions with the compiler and some possible footguns (duplicated declarations bloating code).
[0] https://typia.io/
I believe it would be too hard to keep up with pace of TypeScript development. We should probably at some point formally define the system and allow for alternative implementations outside of the control of Microsoft.
Contrary to the popular trend, I usually stick with Node and NPM for most of my projects. That said, I’ve run into plenty of headaches with CommonJS vs ESM quirks. Sharing code between frontend and backend (via shared libraries) still feels messy because of those differences. In one project I actually had to switch to Bun as the runtime, since Node kept failing at runtime, and strangely enough, Bun just worked without issues.
Deno has sandboxing and much more - use it instead
Great, it does type stripping by replacing TypeScript with whitespace. Can it also now load ES5 modules? That way we can use them for everything.
> Can it also now load ES5 modules?
ES5 standard (released in 2009) didn't have modules.
That answers it. I guess it still can’t! :-/
Quite for some time, assuming they are actually ES6 modules.
Some previous discussion: https://news.ycombinator.com/item?id=44597966
the js ecosystem is so sad.
all threads saying the truth, js on the server could implement actual ts and not yet another transpiler gets downvoted.
js "experts" think they are smarter because they know ts is just annotations for a linter. they don't even question why that is so and why that sucks.
I’ve been happy with JSDoc
Bun can do it for years now. I think it's time to move on.
Bun has lots of issues and lots of incompatibilities with a lot of packages.
What are the issues
For one I do not want to run my startup at the mercy of VC funded tech. Node.js is open source and maintained by a foundation, it will not "run out of funds" or be abandoned if there is no profitability in the near future.
Yeah, but it's not a big bet. Deno can do it, Bun can do it, if they die a tragic VC-fueled death then somebody else (maybe Node) can do it. Using Bun to me is just like using a microwave oven in 1980 — there weren't a lot of microwavable convenience foods yet, but you could sure heat up some leftovers more conveniently and quickly.