Rendered at 08:43:18 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
pjmlp 32 minutes ago [-]
As posted to reddit thread,
He kind of has a point, I think the whole hype of Rust rewrites makes a disservice to the language, a bit like the C into C++ and Objective-C rewrites back on my day, and yet we could not win as expected, on UNIX like platforms, naturally C as the platform language wins out.
While interesting, and a way to grow the ecosystem, eventually it creates a sentiment among those not interested into adopting the language, that creates memes like the Rust Evangelism Strike Force.
This ends up shunning potential users that shut off at the first signs of reference to the language, many times folks that otherwise could have been willing to learn the language.
Note this applies not only to Rust, rather all kinds of rewrites, as mentioned on my youth it was the C++ Evangelism Strike Force, and the flame wars ground took place on USENET.
viraptor 21 hours ago [-]
These issues take away from the article:
"This latter one claims to have spent 8,600 USD to make it happen." - no, people use subscriptions for this, not API pricing. It's extremely unlikely someone paid more than $200.
"A Rust implementation of curl would thus still have had 67 of those vulnerabilities." - some of these could still be prevented due to more expressive language. For example https://github.com/curl/curl/commit/d37d79df1e01066417aff0a9... exists partially because someone thought full parsing and comparison in C is just hard and the alternative likely won't have issues. What are the chances the same decision would be made in Rust with much better struct/enum handling, standard library, and an Eq trait?
goodmythical 1 days ago [-]
Wow, I'm kind of ashamed how eye-opening this article is.
curl is definitely one of those 99% invisible "so perfect I never notice it" things.
I mean, I use it. I see it pop up in dnf update, but I had no idea of the update cadence. 1,306 contributors, 99 versions, ~17k closed PRs, ~5,700 closed issues
Anyway, interesting to see "of the 81 vulnerabilities we have confirmed in curl so far during 2026, 14 were the result of mistakes that most likely would have been avoided had we used a memory-safe language such as rust. That’s 17.3%. A Rust implementation of curl would thus still have had 67 of those vulnerabilities."
fschuett 1 days ago [-]
It's not just about vulns, it's that Rust has a "compile anywhere" (even WASM) that C just doesn't have. I can pull a few Rust libraries that do different things together and make something new from it, with C, good luck with all the Makefiles.
And the small improvements that accumulated can be ported too, when doing Rust rewrites, I would not say that this is an argument for sticking with C per se. And 20% less vulns is still 20% less.
cozzyd 1 days ago [-]
Using libcurl is trivial and you probably already have it installed in your distro.
pjmlp 32 minutes ago [-]
CMake + conan
I find funny the "compile anywhere" remark, given the reason GCC frontend project exists in first place is that the LLVM based compiler isn't able to compile anywhere there is a C compiler available.
ahartmetz 22 hours ago [-]
>agents converted 185,000 lines of C code into 340,000 lines of Rust code (I counted)
Porting to a higher level language with 1.9x the LOC? Ew.
He kind of has a point, I think the whole hype of Rust rewrites makes a disservice to the language, a bit like the C into C++ and Objective-C rewrites back on my day, and yet we could not win as expected, on UNIX like platforms, naturally C as the platform language wins out.
While interesting, and a way to grow the ecosystem, eventually it creates a sentiment among those not interested into adopting the language, that creates memes like the Rust Evangelism Strike Force.
This ends up shunning potential users that shut off at the first signs of reference to the language, many times folks that otherwise could have been willing to learn the language.
Note this applies not only to Rust, rather all kinds of rewrites, as mentioned on my youth it was the C++ Evangelism Strike Force, and the flame wars ground took place on USENET.
"This latter one claims to have spent 8,600 USD to make it happen." - no, people use subscriptions for this, not API pricing. It's extremely unlikely someone paid more than $200.
"A Rust implementation of curl would thus still have had 67 of those vulnerabilities." - some of these could still be prevented due to more expressive language. For example https://github.com/curl/curl/commit/d37d79df1e01066417aff0a9... exists partially because someone thought full parsing and comparison in C is just hard and the alternative likely won't have issues. What are the chances the same decision would be made in Rust with much better struct/enum handling, standard library, and an Eq trait?
curl is definitely one of those 99% invisible "so perfect I never notice it" things.
I mean, I use it. I see it pop up in dnf update, but I had no idea of the update cadence. 1,306 contributors, 99 versions, ~17k closed PRs, ~5,700 closed issues
Anyway, interesting to see "of the 81 vulnerabilities we have confirmed in curl so far during 2026, 14 were the result of mistakes that most likely would have been avoided had we used a memory-safe language such as rust. That’s 17.3%. A Rust implementation of curl would thus still have had 67 of those vulnerabilities."
And the small improvements that accumulated can be ported too, when doing Rust rewrites, I would not say that this is an argument for sticking with C per se. And 20% less vulns is still 20% less.
I find funny the "compile anywhere" remark, given the reason GCC frontend project exists in first place is that the LLVM based compiler isn't able to compile anywhere there is a C compiler available.
Porting to a higher level language with 1.9x the LOC? Ew.