461–470 of 797 posts
Interesting how a person's opinion can change: https://news.ycombinator.com/item?id=27594688
I assume it was a management decision to adopt Rust in APT similar to the decision to switch to the Rust version of coreutils.
> This extends at first to the Rust compiler and standard library, and the Sequoia ecosystem. By Sequoia, are they talking about replacing GnuPG with https://sequoia-pgp.org/ for signature verification? I really hope they don't replace the audited and battle-tested GnuPG parts with some new-fangled project like that just because it is written in "memory-safe" rust.
Sequoia-PGP is 8 years old at this point, their 1.0 happened half a decade ago. Meanwhile, GnuPG is well regarded for its code maturity. But it is a C codebase with nearly no tests, no CI pipeline(!!), an architecture that is basically a statemachine with side effects, and over 200 flags. In my experience, only people who haven't experienced the codebase speak positively of it.
It exits 0 when the verification failed, it exits 1 when it passed, and you have to ignore it all and parse the output of the status fd to find the truth.
It provides options to enforce various algorithmic constraints but they only work in some modes and are silently ignored in others.
Earlier quoted context omitted.
I think that's unkind absent any proof otherwise which you haven't posted. In reality a lot of the heavily used parts of Linux (and open source generally) are going to have some commercial involvement or are we suggesting that no one should be paid by any of the companies backing Linux who use Linux because to me that sounds worse . Whether you like it or not, Linux/Open Source hasn't been entirely the preserve of un…
https://en.wikipedia.org/wiki/The_Scorpion_and_the_Frog I don't understand how people keep giving the benefit of the doubt to corporations. They aren't people. They feel no guilt and have no shame.
It's about time. Critical infrastructure still written in C - particularly code that parses data from untrusted sources - is technical debt that is only going to get worse over time. It's not as if Rust is that much more difficult to write than C. Rust is explicitly designed to be what you'd get if you were to re-create C knowing what we know now about language design and code safety. If 32-bit x86 support can be dro…
Oh please, in a decade Rust will also be technical debt and people will be wanting to write it in Brust or whatever is the trendy new language.
They’d be better off just compiling the package manager with Fil-C No changes required. Bringing up the fil-C toolchain on weird ports is probably less work than bringing up the Rust toolchain
It also doesn't help you to attract new contributors. With the changes we made over in Ubuntu to switch to rust-coreutils and sudo-rs, we have seen an incredible uptake in community contributions amongst other things, and it's very interesting to me to try to push APT more into the community space.
At this time, most of the work on APT is spent by me staying awake late, or during weekends and my 2 week Christmas break, the second largest chunk is the work I do during working hours but that's less cool and exciting stuff :D
Adding Rust into APT is one aspect; the other, possibly even more pressing need is rewriting all the APT documentation.
Currently the APT manual pages are split into apt-get and apt-cache and so on, with a summary in apt(8) - we should split them across apt install(8), apt upgrade (8) and so on. At the same time, DocBook XML is not very attractive to contributors and switching to reStructuredText with Sphinx hopefully attracts more people to contribute to it.
It's about time. Critical infrastructure still written in C - particularly code that parses data from untrusted sources - is technical debt that is only going to get worse over time. It's not as if Rust is that much more difficult to write than C. Rust is explicitly designed to be what you'd get if you were to re-create C knowing what we know now about language design and code safety. If 32-bit x86 support can be dro…
> Rust is explicitly designed to be what you'd get if you were to re-create C knowing what we know now about language design and code safety. I don't know about that. Look at the code for the COSMIC desktop environment's clock widget (the cosmic-applet-time directory under https://github.com/pop-os/cosmic-applets >), for example. It's pretty much unreadable compared to a C code base of similar complexity (GNU coreuti…
I beg to differ.
It's about time. Critical infrastructure still written in C - particularly code that parses data from untrusted sources - is technical debt that is only going to get worse over time. It's not as if Rust is that much more difficult to write than C. Rust is explicitly designed to be what you'd get if you were to re-create C knowing what we know now about language design and code safety. If 32-bit x86 support can be dro…
> In particular, our code to parse .deb, .ar, .tar, and the HTTP signature verification code would strongly benefit from memory safe languages > Critical infrastructure still written in C - particularly code that parses data from untrusted sources - is technical debt that is only going to get worse over time. But hasn't all that foundational code been stable and wrung out already over the last 30+ years? The .tar and…
Not necessarily. The "HTTP signature verification code" sounds like it's invoking cryptography, and the sense I've had from watching the people who maintain cryptographic libraries is that the "foundational code" is the sort of stuff you should run away screaming from. In general, it seems to me to be the cryptography folks who have beat the drum hardest for moving to Rust.
As for other kind of parsing code, the various archive file formats aren't exactly evolving, so there's little reason to update them. On the other hand, this is exactly the kind of space where there's critical infrastructure that has probably had very little investment in adversarial testing either in the past or present, and so it's not clear that their age has actually led to security-critical bugs being shaken out. Much as how OpenSSL had a trivially-exploitable, high criticality exploit for two years before anybody noticed.
I really like to write programs in rust. But my stance has changed a bit over the years ever since other languages caught up a bit. On top of that I’m very skeptical if the rewrite of an ancient tool brings more less security. I don’t know the apt source code or how it actually works behind the cli interface so I leave this judgement to the pros. But there seems to be a very strong move to rewrite all core systems in…
I've heard two arguments for these rewrites that don't always come up in these discussions. There are fair counterpoints to both of these but I think they add valuable dimensions to the conversation, or perhaps may explain why a rewrite may not seem justified without them. * It's becoming increasingly difficult to find new contributors who want to work with very old code bases in languages like C or C++. Some open so…
On your second part. I wonder how aviation and space and car industry do it. They rely heavily on tested / proven concepts. What do they do when introducing a new type of material to replace another one. Or when a complete assembly workflow gets updated.
Earlier quoted context omitted.
> I know nobody that programms or even thinks about rust. That's you. At companies like Microsoft and Google, plenty of people think about and discuss Rust, with some products/features already using Rust.
Well, at Microsoft, plenty of people have thought about and discussed C#, with some products/features using it. Not that it's gone away, but it has not won the hearts and minds of the general (software-developing) public.
> This extends at first to the Rust compiler and standard library, and the Sequoia ecosystem. By Sequoia, are they talking about replacing GnuPG with https://sequoia-pgp.org/ for signature verification? I really hope they don't replace the audited and battle-tested GnuPG parts with some new-fangled project like that just because it is written in "memory-safe" rust.
Sequoia-PGP is 8 years old at this point, their 1.0 happened half a decade ago. Meanwhile, GnuPG is well regarded for its code maturity. But it is a C codebase with nearly no tests, no CI pipeline(!!), an architecture that is basically a statemachine with side effects, and over 200 flags. In my experience, only people who haven't experienced the codebase speak positively of it.
Does Sequoia-PGP have similar credentials and who funds it?