Tsrs: A Port of TypeScript to Rust (blog.cloudflare.com)
292 points by hnisjafx40 8 hours ago
DanRosenwasser 6 hours ago
Hey there, TypeScript team member here. The tsrs project has been extremely interesting to see come together. We've been working with the maintainers to better understand how tsrs is fast and what sorts of optimizations it uses are safe. We even did our own Rust vibe port and what we currently suspect is that a lot of the speed does not just come down to language choice, and some of what is going to make TS 7.1 faster and leaner (see the charts in the blog post) actually comes from optimizations discovered in tsrs.
We think the wasm and oxlint integration makes a lot of sense for Cloudflare's needs, so this investment makes sense. At the same time, we'll be working hard and collaborating to make the official TypeScript toolset faster and even more efficient.
Congrats to the tsrs team on making this official. We're excited and we look forward to continuing on collaboration!
samuelstros 6 hours ago
At this point, why invest further in TSGO?
No matter how many optimizations can be ported from tsrs to ts-go, Rust will be more efficient and enable the WASM use case.
The premise of porting the compiler from TS to GO was ease of migration due to Go's garbage collector. That doesn't hold anymore? AI doesn't care how hard or easy a migration is if a test suite exists.
magicalist 5 hours ago
> That doesn't hold anymore? AI doesn't care how hard or easy a migration is if a test suite exists.
Sure, but that only matters if migration is going to be the last thing anyone does on a project. Unless you think typescript 7.1 ought to be enough for anybody, it also matters what the typescript team wants to write/read/reason in.
collabs 2 hours ago
platinumrad 4 hours ago
I just vibecofed tsasm. tsrs is obsolete now and anyone who still uses it is hopelessly behind the times. The tsrs team should delete the repository immediately and join my project.
written-beyond 2 hours ago
cute_boi 2 hours ago
wavemode 5 hours ago
Your answer is within the comment you wrote. AI is maintaining tsrs whereas humans are maintaining tsgo.
Your question boils down to "why would humans maintain something?" which could have a variety of answers depending on context.
greenavocado 5 hours ago
melodyogonna 4 hours ago
I never understand why people say things like this. It is not uncommon for languages to have multiple implementations, everything must not be abandoned for Rust man.
fallingbananna 4 hours ago
Because TSGO is stable and compatible?
I assume that most of the speedup in ts-rs does not simply stem from using a language without garbage collection, but actually from doing smart optimizations that can be done in any langauge.
Similar to how uv in the Python world explained that their implementation being in Rust has very little to do with the most impactful speedups they did.
skybrian 2 hours ago
Another advantage of AI, though, is that it becomes easier to maintain more than one TypeScript compiler. Perhaps there's some benefits to having some diversity?
pier25 5 hours ago
TSGO is already plenty fast and overkill for most projects.
Sure the Rust version is faster but the difference is probably only noticeable in the top 1% largest projects.
theappsecguy 5 hours ago
alain_gilbert 5 hours ago
lifeisloving an hour ago
Please you port it if its so easy, the tests are available on gh. I wont be using an llm ported ts compiler anyways, and neither would many others
Its a great way to divide and kill the popularity of your project. I made sure bun is no longer used in my devices or projects that depend on it for example, I'm not alone in this.
egonschiele 6 hours ago
Thanks for the response and it is encouraging to hear that there's more to this project than vibe coding! Sounds like some interesting things were found in the process, and I'm all for vibe coding as a way to learn new things.
lifeisloving an hour ago
Im not because nobody actually ever learns anything. For every one person that learns something the other 99% of the population gets lazier and dumber, ponder how that might play itself out.
This is probably the most harmful tech ever created simply because how many people trick themselves into thinking they're smarter while actually becoming much dumber. We're actually cooked.
zyf 5 hours ago
Been with TS for many years now and the work your team's done over the years is amazing. Thank you
lemagedurage 6 hours ago
Thanks for the input, glad to hear both parties benefit.
Could you elaborate on what type of optimizations were added to TS 7.1? As in, were they more algorithmic or related to memory management?
DanRosenwasser 5 hours ago
If I recall correctly, a few of both - there were instances where some work could be done more lazily/minimally (https://github.com/microsoft/TypeScript/pull/64476), and there were instances where we changed our allocation patterns and played around with some struct layouts. We originally strived not to drastically reorganize our data structures to make things easier to port, and we have more flexibility now. We're continuing to experiment more here.
satvikpendem 5 hours ago
Ah I was gonna ask what happened with tsc-rs by Theo but looks like you guys are merging it in, at least it won't be a waste of his $500k in token spend, lol.
afdbcreid 6 hours ago
It appears they're using a unique memory model (https://github.com/oxc-project/tsrs/blob/main/docs/PORTING.m...) to overcome the challenge that prevented the TypeScript team from choosing Rust to begin with: Rust has no GC and porting from a GC language to it often requires restructure of data.
Essentially, they fake a GC via never-freed arenas for long-living objects. It's interesting that Rust can do that (hopefully soundly)!
zamalek 6 hours ago
From the file:
>Create with P::new(value). p.get() returns &'static T. [...] unless an explicit recycling site proves them dead
It's unsound, the LLM/human is doing the proving. They've hidden unsafe in an abstraction without removing the unsafeness. Using the arena allocator's lifetime would have been safe and easy (bumpalo recommends this AFAIK).
I'm guessing a sycophantic LLM either found or praised this load-bearing seam.
Never deallocating would have been sound (missile memory management: https://devblogs.microsoft.com/oldnewthing/20180228-00/?p=98... ).
tengbretson 4 hours ago
> Never deallocating would have been sound (missile memory management: https://devblogs.microsoft.com/oldnewthing/20180228-00/?p=98... ).
Calling this a breakthrough in memory management is an understatement. Call up the Turing Award committee. We just discovered a solution to the halting problem! Turns out that for certain applications the answer is unequivocally "yes"
afdbcreid 5 hours ago
With sanitizers (https://github.com/oxc-project/tsrs/blob/main/notes/mem-recy...). We're back to C I guess. I hate this project now.
lioeters 5 hours ago
> load-bearing seam
Zing!
jcmontx 5 hours ago
Wouldn't that eat the whole ram?
panzi 4 hours ago
Compilers often just build up more and more data structures in memory until the end when they then release everything more or less at once. With an arena you can then just throw away the whole arena at the end, assuming there are no resource handles that need to be closed in drop().
A long time ago I read an article about someone speeding up a C++ compiler that is itself written in C++ (was it g++? maybe clang?). They simply patched out delete so it did nothing. The compiler used marginally more memory but ran much faster, since releasing all the small objects one by one is just useless work at the end of the compilation.
A GCed language has a similar behavior and could therefore be fast in that case. It just builds up all the objects and then when the program exists it doesn't have to GC them all, it just throws away the whole process memory. So I can imagine one could "optimize" a compiler written in a GCed language by deactivating the GC.
sheepscreek 6 hours ago
Ah took me a while to understand the reason behind this.
It wasn’t “let’s build a faster Rust based Typescript compiler” at all. That was a nice side-effect.
They needed a Typescript compiler that can run on their WASM only sandbox. So the only existing option was to export tsc Go to WASM, but that comes with the full Go runtime.
And to avoid that cost, they experimented with a Rust compiler and turns out it’s even faster than the reference tsc/Go implementation.
nitsky 6 hours ago
Another reason: type aware listing in oxlint.
anamexis 6 hours ago
Oxlint already has type-aware linting, integrated with TS7/typescript-go.
nitsky 2 hours ago
egonschiele 7 hours ago
It's exciting to see the speed gains. I use TypeScript every day on a large codebase, and it is too slow to run locally. I don't love this change though:
> However, frontier AI models can now do large-scale ports almost independently in a matter of days
> I wrote zero code. Not "barely any". Zero
So we are going to move from a carefully handwritten port to an entirely AI-generated port. I know this is the way the industry is headed, I use coding agents a ton. But I still review their output because they often don't make sound architectural decisions. Maybe this is a sign that my way of working is going away. I'm not sure how to justify that kind of time spent when someone can do the same thing in a matter of days with barely any effort.
I am working on my own language now, I've spent most of this year on it. The idea that someone could just replace it in a matter of days is wild. Wonder how the TS team at Microsoft feels about this.
TheSamFischer 6 hours ago
> The idea that someone could just replace it in a matter of days is wild.
They didn’t design a language that millions of human engineers enjoy (?, lol) using. They took an existing codebase and ported it. Yes, the porting part is easier now but only because another project and spec exist.
FuckButtons 6 hours ago
While this is true - and I see the same sentiment as i am about to make ad nauseam - it is kind of amazing how far the goal posts have moved in such a short space of time.
TheSamFischer 6 hours ago
egonschiele 6 hours ago
That's fair, I was thinking the same – porting must be simpler, especially if there's already a substantial test suite
onlyrealcuzzo 6 hours ago
If TypeScript has the integration testing framework in place to verify that the compiler actually works (and it should) - then, presumably, you should be able to swap in a Chinese Room black-box and - at least - know that it works.
You do have to be worried that it's possibly installing trojan horses on your computer while compiling code, or any other number of unwanted side effects.
But you can and should be able to have integration testing in place to know that a compiler - at a minimum - actually works.
egonschiele 6 hours ago
Totally fair. The thing is, though, agents often write code in ways I don't expect, and I don't think any human would write, which makes me think that code may also break in a way that no human expects or would have written a test for.
spiderice 4 hours ago
Your way of working is definitely going away. And quickly. The idea of reviewing code today is already a questionable use of time. In a few more months to a year it will clearly be a waste of time.
_sinelaw_ an hour ago
Or we can just drop TypeScript completely and type-check JavaScript directly.
With HM-style type inference, row-polymorphism, sum types, type classes for overloading + operations, recursive types for object methods, and type annotations for edge cases.
I've been working on this since before TypeScript existed. Latest incarnation: https://github.com/sinelaw/inty
Demo: https://sinelaw.github.io/inty/ (you may need to reload once)
brazukadev an hour ago
isn't it better to use JSDoc ?
_sinelaw_ 38 minutes ago
I assume you mean jsdoc + tsc in js type checking mode (tsc --allowJs --checkJs --strict)?
inty infers more - including generics (parametric polymorphism), while tsc is more of a simpler local inference.
inty is more strict - it accepts less programs than tsc would. This is mostly on purpose (to our benefit) as some of those programs have unsound code like array covariant (downcasting array members of the wrong type). tsc makes it easier to slip through these bugs.
And if tsc + jsdoc is good then the same question remains (do we really need the TypeScript language?)
AbuAssar 7 hours ago
According to voidzero, this port is up to 8x faster than TS 7.0 & ~70% less memory.
Microsoft should seriously consider porting typescript to rust at this point
aleksandrm 6 hours ago
I'm pretty sure they already considered it before settling on Go. Allegedly, they chose Go because it was easier to translate to.
satvikpendem 5 hours ago
That was pre-AI. Ironically if they waited just another year they could've done this port directly to Rust themselves.
nish__ an hour ago
ndr 7 hours ago
I wonder how more frustrating is it to work on the compiler in rust vs go.
IIRC at the time of the port to go they wanted fast test and compile time, of course it's a different era now.
lemagedurage 6 hours ago
That's interesting. A go version to iterate on the code and a rust version to run the code might be an optimal split here.
tgv 7 hours ago
Idk. Apart from the fact that benchmarks are generally of questionable value, what's the point? Is the TS compiler unbearably slow?
ruther4d 7 hours ago
They say the point in the article - WASM
msdz 7 hours ago
TheMrZZ 7 hours ago
The success of "modern" CLI tools written in Rust (uv, oxlint, turbopack) show there's a clear demand for faster tools in general.
Also, faster feedback loop is crucial for AIs, so I guess this makes sense.
pjmlp 7 hours ago
pier25 5 hours ago
> Is the TS compiler unbearably slow?
It's not. Even the JS compiler is still fast enough for many projects given the single core improvements in hardware (eg Apple Silicon).
eterm 7 hours ago
Faster feedback loops almost always deliver improvements elsewhere.
Rapzid 6 hours ago
Up to 8x faster seems kinda of a crazy claim. 8x faster at what?! TS 7.0 is reasonably well optimized and compiles VSCode in well under 10 seconds..
This can type check and compile VSCode in less than a second?
lemagedurage 6 hours ago
For VSCode: 4x speedup vs TS 7, 3x vs TS 7.1
rs_rs_rs_rs_rs 6 hours ago
Yes, it's in the article.
magicalist 6 hours ago
Leynos 7 hours ago
So that's three now. This, tsgo-rs, and `bun check`
alexanderson 7 hours ago
They mention in the post they are working with Theo to absorb tsgo-rs into tsrs.
Rapzid 6 hours ago
Why would they bother merging his vibe slop if theirs is better?
preommr 2 hours ago
shimman 6 hours ago
Leynos 6 hours ago
Nice. I missed that. This is good news.
gzapp 6 hours ago
Shameless plug of my Typescript C# port. It's a case study with some interesting (perhaps) notes from me in the README about approach, performance, and cost(~$75).
jrop 3 hours ago
What strange times. A few years ago seeing a headline like this, my jaw would have literally dropped. Instead I'm wondering what kind of software projects will have meaning in the future if this headline feels common.
yabirgb an hour ago
I wonder if something similar has been attempted with Python. CPython is a project where I don't know how they are taking advantage of AI.
LudwigNagasena 35 minutes ago
The problem with Python is that the whole ecosystem is based on CPython’s implementation details, so every change inevitably breaks something and requires effort of package maintainers to fix compatibility. There is no test suite an agent can follow.
mohsen1 2 hours ago
I ran out of patience before models got good enough to get this done.
Congrats to the team on getting this out
sashank_1509 3 hours ago
I guess it will be soon that typescript won’t matter anymore since agents will be writing code is Rust compiled to WASM or similar for everything. typescript added human readability or JavaScript, if no human is coding, I don’t see the point of wasting tokens on this.
hollowturtle 2 hours ago
> typescript added human readability
Ehm no, added type safety, and dealing with a type checker it's exactly what agents need.
> if no human is coding
Nope outside the bubble there are many still hand coding(me too)
> I guess it will be soon that typescript won’t matter anymore since agents will be writing code is Rust
Oh hold on a second isn't Rust the same as Typescript "adding human readability"?! Why not trashing Rust and just have LLMs output WASM directly? XD Wasm which is basically useless outside the browser or some microvms on the edge, and in the browser you don't even have an interface for directly rendering anything and still have to deal with the dom
You either are highly non technical, very junior or I just don't know
hollowturtle 2 hours ago
All of this just for enabling the WASM use case? Why not keep supporting Deno then that is written in Rust? To me it just sounds like a technical ai journey and a blow to the go port
brazukadev 2 hours ago
I think it is a response to tsc-ts and scriptc from Vercel. CloudFlare and vercel CEOs have a personal competition going on.
norman784 7 hours ago
I'm tired of all the new tools that comes out every day, but I would say, this is basically the Vite team and if they implement it with Vite+, then I'm all for it, less decisions about tooling to be made, one config file, etc.
ComputerGuru 2 hours ago
Apparently only supported under darwin and linux for arm64 and x64? At least via npm.
wannabe44 7 hours ago
Y'all building the same thing.
esafak 31 minutes ago
It's a nit, but do you guys need to embed the TS version in semver? https://www.npmjs.com/package/tsrs?activeTab=versions
jscheel 6 hours ago
I benchmarked tsrs a few weeks ago. I was testing out tsgo, ts-rust, tsrs, and bun for typechecking on our monorepo. tsrs is definitely awesome, it is very memory efficient when bumping up the number of threads. Not as efficient as bun, but it has incremental checks, which bun does not have.
egeozcan 2 hours ago
I think we'll see TypeScript ported to minecraft ported to rust ported to typescript before GTA6.
panzi 4 hours ago
I wonder how fast "Doom in TypeScript" runs in tsrs now. XD
frangonf 7 hours ago
I use oxc tooling and it's great to see the rust rewrite singularity in action and get to ride the fastest horses, but I would like that they made a branch for 7.2 as experiment to see how it compares. Maybe this ends in a win for Microsoft, they could cut operating expenses by keeping just Anders Hejlsberg in a hammock reciting invariants to claudpilot.
nsm 7 hours ago
Obligatory reminder that their needs from a TS compiler are extremely different from a regular web developers iteration cycle. They are trying to run millions of these while saving on costs.
You almost definitely don't need to immediately plan to switch.
ameliaquining 44 minutes ago
I expect I'll switch once Vite starts shipping it by default.
cryptonym 6 hours ago
> This enables new possibilities when combined with agents: You are now able to run a development toolchain end-to-end in a fully sandboxed environment on Cloudflare.
True. Looks like main use-case is AI Agent Sandboxing
sinameraji 3 hours ago
this is insane! saw a quote tweet of this by an infra engineer at chat gpt who said “ I'm looking at migrating over for the ChatGPT codebase and saw ~70% improvements on both wall time perf + peak cpu/memory usage.
this is great for getting work done faster but also much better if you want to run concurrent agents/worktrees at scale”. https://x.com/feedthejim/status/2112584843368570921?s=46
then i went and tried this on my own open source harness. type checking went from 10 seconds to under half a second and uses a quarter of the memory. madness
tonyhart7 7 hours ago
I thought its TS port of theo
AntonyGarand 7 hours ago
Theo's port is tsc-rs[0] as reference, and is significantly slower and uses more memory than this new version.
francislavoie 7 hours ago
A new one but they're merging it with Theo's.
jdw64 7 hours ago
These days the trend seems to be rewriting everything in Rust.
zeroq 6 hours ago
I'm staying stainless
satvikpendem 5 hours ago
You should say you're staying fungal infection free as it's named after the Rust fungus not the chemical reaction even if that's what people think these days.
jdw64 6 hours ago
Did you know? Even stainless steel RUSTs in sea wind.
g0ld3nrati0 5 hours ago
so rust is now go-to slop language? can't you guys use C or something. I want rust back.
preg_match 9 minutes ago
Rust is basically the perfect language for LLMs because it has a strong, comprehensive, and strict compiler. For coding you really need a tight loop for models so they can iterate. C doesn’t accomplish this because it’s far too “anything goes” so bugs slip through. Dynamically typed languages don’t work either because their source code doesn’t contain enough context.
cpursley 3 hours ago
What? You realize how many people now are excited about Rust that were not before? I am, for one.
bakugo an hour ago
So excited that you avoid writing a single line of it at all costs.
nalekberov 6 hours ago
Oh god, I wish this Rust fetishism would end soon.
nalekberov an hour ago
For those who downvotes my comment, I don't give a shit. Use your brain.
retropragma 7 hours ago
please support TSRX (https://github.com/tsrx-org/tsrx)