❌

阅读视图

Announcing Rust 1.100.0

Nov. 12, 2026 · The Rust Release Team

The Rust team is happy to announce a new version of Rust, 1.100.0. Rust is a programming language empowering everyone to build reliable and efficient software.

If you have a previous version of Rust installed via rustup, you can get 1.100.0 with:

$ rustup update stable

If you don't have it already, you can get rustup from the appropriate page on our website, and check out the detailed release notes for 1.100.0.

If you'd like to help us out by testing future releases, you might consider updating locally to use the beta channel (rustup default beta) or the nightly channel (rustup default nightly). Please report any bugs you might come across!

What's in 1.100.0 stable

Never type

After 10 years and 5 failed stabilization attempts, the never type is finally stable! This also completes the project goal for this work.

The never type, spelled ! in code, is the canonical uninhabited type. It represents the type of diverging computations – computations which never resolve to any value – where any code that follows is unreachable. Expressions that terminate control flow, such as return, break, and std::process::exit, also have the type !.

Since having a value of the never type is impossible, Rust permits a never-to-any coercion, allowing to turn ! into any other type (this is what distinguishes it from any other uninhabited types). This can be used, for example, to not return a value from one of the branches of a match or an if:

fn mrrrow(option: Option<u32>) {
    let value = match option {
        // `x` has type `u32`
        Some(x) => x,
        // `return` has type `!`, which is then coerced to `u32`,
        // allowing the `match` to pass type checking.
        None => return,
    };
    // ...
}

While the never type could already be used and observed this way, and even named in the return type of functions (-> !), this release allows naming the never type anywhere:

fn infallible_unwrap<T>(res: Result<T, !>) -> T {
    // ^^ writing `Result<_, !>` was not allowed before
    let Ok(v) = res;
    v
}

Additionally this release changes Infallible to be an alias to the never type. That makes all old APIs that used Infallible be compatible with using !. This was the plan ever since the introduction of Infallible and is one of the reasons why the never type stabilization was so hard (waffle gave a talk about some of the troubles of the stabilization).

Unfortunately, to allow changing Infallible, we had to introduce a breaking change. This release changes the "never type fallback" from () to ! in editions before 2024 (i.e. 2015, 2018, and 2021), similarly to what has already been done in edition 2024. This should largely be unnoticeable, but can affect code in some rare cases (for example when ignoring a result of a function returning a generic parameter). Note that this change might also make the compiler detect more code as unreachable.

We tried to minimize the impact of this breaking change as much as possible, by providing future compatibility warnings far in advance and releasing fixes to public crates that have been broken by this. If a dependency is failing due to issues with the never type, consider updating it, as a fix might have already been released for it.

Allocator API

This release brings about the stabilization of the core parts of allocator_api, enabling library authors to define custom allocators that can be used together with certain standard library types - currently Vec and Box. These are enabled through the newly-stabilized Allocator trait.

This is the foundation of the Allocators 1.0 project goal. A lot remains to be done still, including expanding the API surface usable with custom allocators and stabilizing their use on more collections and pointer types; with the core trait and design out of the way though, expect to see more being stabilized soon!

DropGuard: run a closure on every exit path

The Drop trait in Rust makes it possible to run code when a variable goes out of scope, regardless of whether we return, ?, or panic. Drop being a part of types means that in Rust you never have to close a socket.

Implementing Drop can however feel a little inconvenient when all you want is to run some code unconditionally when a function exits. This is why we’re introducing core::mem::DropGuard: a new helper type that runs a closure when dropped.

DropGuard takes two arguments: a variable and a closure. It implements Deref and DerefMut so you can keep working with the variable as normal. But when dropped it runs the closure and consumes the variable.

DropGuard also implements a dismiss associated function[1] which returns the variable from the guard without running the destructor. This makes it possible to quickly implement all-or-nothing transaction-like APIs with a rollback. Here is an example implementing a parser which uses a guard to roll back mutations in case of failure:

use std::mem::DropGuard;
use std::num::ParseIntError;

/// Attempt to parse all `strings` and push them into `nums`.
/// If parsing any of them fails, `nums` remains unchanged.
fn parse_all(nums: &mut Vec<u32>, strings: &[&str]) -> Result<(), ParseIntError> {
    let len = nums.len();
    let mut nums = DropGuard::new(nums, move |nums| {
        nums.truncate(len);    // ← 1. Reset the vec on failure
    });

    for s in strings {
        nums.push(s.parse()?); // ← 2. If `?` returns an error, roll back
    }

    DropGuard::dismiss(nums);  // ← 3. Parsing successful, dismiss the guard
    Ok(())
}

Conjuring zero-sized types

As Rust's unsafe code guidelines are still being decided, it became clear that it would be very useful to decide what exactly it means to create zero-sized types in an unsafe way, and what we should recommend for people to do this.

This question is answered in Rust 1.100 with a single function: std::mem::conjure_zst.

Instead of transmuting (), MaybeUninit::uninit().assume_init(), or various other incantations, this is the dedicated way to create a value of a zero-sized type when you know that doing so is safe to do. If an inhabited type has no size, we know exactly what value it must be: nothing.

Of course, APIs often rely on the inability to construct such types for safety, and that's why this function is unsafe. For example, PyO3's Python<'py> is a zero-sized marker token that represents holding the Python GIL (global interpreter lock). Many of Python's APIs require holding this lock to be safe, but conjuring a token from thin air would allow safe PyO3 calls to violate that precondition.

Reporting unused dependencies in Cargo

Building dependencies that are no longer used or are only needed for specific platforms can slow down builds. Rustc has the unused_crate_dependencies lint for this but it has false positives when used with Cargo because dependencies may be shared between multiple build targets, e.g. dependencies.tracing-subscriber being available for every bin in a package when only some may use it.

Cargo now reports unused dependencies to users, e.g.

warning: unused dependency `tracing-subscriber`
  --> Cargo.toml:23:1
   |
23 | tracing-subscriber.workspace = true
   | ^^^^^^^^^^^^^^^^^^
   |
   = note: `cargo::unused_dependencies` is set to `warn` by default
help: consider removing the dependency on `tracing-subscriber`

A limitation of this warning is that it will only appear when building every build target that may use an entry in a dependencies table. For example, running cargo check --bin foo won't show warnings for unused [dependencies] because another bin may be present that uses them. No warning will be shown for unused dev-dependencies because there isn't a way to ensure every development build target is built.

This warning is being introduced as part of a new linting system for Cargo, controlled by the [lints.cargo] table in Cargo.toml, completing this project goal. See also the other lints that have been introduced.

Reducing Cargo cache collisions

Cargo builds all libraries and binaries in the build-dir and then uplifts final artifacts to the target-dir. At this time, these share a common parent directory. Different build configurations, like --features, can lead to different copies of libraries and binaries in the build-dir. These copies are kept unique by a build-configuration hash but this hash was not always applied due to platform constraints. This can lead to extra rebuilds, missed rebuilds, confusion over what binaries or libraries are used, and so on. These problems get exacerbated when sharing a build-dir across git worktrees.

The build-dir now consistently applies the build-configuration hash to all content by organizing the directories around their build step, rather than their file type, and moving the hash from a file's name to the parent directory.

This work began with the build directory layout project goal in 2025, and its stabilization is a milestone toward a new project goal, unblocking future improvements such as cache garbage collection, cross-workspace cache de-duplication, cross-workspace caching, and fine grain locking.

We previously published a call for testing for the changes that are now taking effect. While this is an internal change, this may break build processes and tests that had workarounds for missing or incomplete features. In some cases, the workaround is no longer needed. See cargo#17287 for possible solutions and to add any new use cases.

Cargo min-publish-age a.k.a. dependency cooldowns

Malicious versions of packages are often caught within hours or days after being published, by either automated scanners or by the community. Rust 1.100 stabilizes Cargo's min-publish-age aware dependency resolution, which lets you wait for that window before it picks up newly published versions. This practice is also known as dependency cooldowns.

You can opt in via .cargo/config.toml:

[registry]
global-min-publish-age = "3 days"

When selecting versions, Cargo now skips any version published more recently than that, and tells you what it held back:

$ cargo update
    Updating crates.io index
     Locking 2 packages to highest compatible version as of min-publish-age
      Adding cc v1.11.0 (available: v1.18.0, published 3 hours ago)
      Adding clap 4.6.7 (available: 4.6.9, published 2 days ago)

When a new version is needed right away, you can override the check for a single update:

$ CARGO_RESOLVER_INCOMPATIBLE_PUBLISH_AGE=allow cargo update clap --precise 4.6.9
    Updating crates.io index
    Updating clap 4.6.7 -> clap 4.6.9 (published 2 days ago, minimum age 3 days)

The newer version will be recorded in Cargo.lock and kept from then on until a later cargo update re-resolves it.

A few things to keep in mind:

  • This depends on the registry recording when each version was published. Versions without a pubtime field in the index are never held back.

  • cargo install ignores the minimum publish age, for both the installed crate and its dependencies. Pin a version with cargo install foo@1.2.3 if needed.

  • Cooldowns reduce the risk from some supply chain attacks but are not a complete defense on their own. Please keep reviewing dependency updates as you otherwise would.

Further work to protect the ecosystem is planned under the crates.io supply chain hardening project goal, including quarantine support and automated detection of suspicious releases.

Stabilized APIs

Other changes

Check out everything that changed in Rust, Cargo, and Clippy.

Contributors to 1.100.0

Many people came together to create Rust 1.100.0. We couldn't have done it without all of you. Thanks!

  1. it has deliberately been chosen not to be a method to avoid name collisions with the DerefMut-enabled method resolution on the inner type. ↩

  •  

Google Summer of Code 2026 results

The Rust Project participated in Google Summer of Code (GSoC) for the third time this year. This time we had 13 projects, and our dear contributors have been working very hard on them for several months. As usual, the projects had various durations, so while some of them have ended in August, the last one has only finished in November, so we have been waiting until we have all the results before publishing this blog post.

We are very happy to announce that all 13 projects were successful and passed their final evaluation! We consider it to be a testament to the diligence of our contributors, and the helpfulness of our mentors, that we are able to keep a very high rate[1] of successful GSoC projects several years in a row.

At the start of summer, we had a video call with our contributors, so that we could all get to know each other a bit, and learn about their projects. Over the following few months, we continuously communicated with the contributors, provided guidance and answered their questions and reviewed their pull requests. And most importantly, we learned many new things together! It was an overall great experience. We are grateful for all Rust contributions done by our GSoC participants and for the time we spent working together, and we hope that we will continue seeing at least some of them around in the future.

We would like to thank all our Rust GSoC 2026 contributors. You did a great job!

Even though this year we had fewer projects than in 2025, mostly due to limited mentoring capacity, Google Summer of Code 2026 was again a success for the Rust Project. We think that GSoC is a great way of introducing new contributors to our community, and we are looking forward to participating in GSoC (and other mentorship programs, such as Outreachy) again in the near future. If you are interested in becoming a (GSoC) contributor, check out our GSoC project idea list and our guide for new contributors.

Below you can find a brief summary of our GSoC 2026 projects. You can find more information about the original goals of the projects here. For easier navigation, here is an index of the project descriptions in alphabetical order:

And now strap in, as there is a ton of great content to read about here!

A Frontend for Safe GPU Offloading in Rust

Marcelo worked on improving the usability of cross-vendor GPU programs in Rust using the std::offload module. At the start of the project, the offload prototype was only a proof of concept. It assumed that the offloaded code was bug-free, and it required users to write a lot of boilerplate code. Marcelo resolved both issues over the summer.

First, he implemented proper type checking. Marcelo extended the compiler to verify that the arguments passed to any offload call are of the correct type, as expected by the GPU kernel. Kernel launches are thus now safe.

After that, he added support for generic kernels. The compiler now collects all generic instantiations launched from the CPU side, and loads that info later when compiling GPU kernels to make sure we instantiate them as needed. Marcelo also added support to select which GPU to use, which is helpful for laptops with both an integrated and a dedicated GPU, as well as for servers with multiple GPUs. He also added two macros that removed the need for any boilerplate when writing and launching GPU kernels.

The hardest part of this project by far was dealing with references. Due to Rust semantics, we cannot just map types containing mutable references to GPUs and then process them using multiple threads. Previous Rust GPU projects used raw pointers as a workaround, but we obviously wanted to avoid those. Marcelo managed to implement a sound abstraction so that users can now safely declare how their kernels will access mutable arguments.

As a cherry on top, Marcelo also ported the RajaPerf performance suite from C++ to Rust and used it to benchmark his pull requests. His safe Rust version matches the performance of the original C++ code, and is now described in a GPU offload paper that he is working on together with his mentor, Manuel. This was Marcelo's second Rust GSoC project, and same as last year, we are very happy with the results. Great job!

Adding WebAssembly Linking Support to Wild

The goal of this project was simple on paper: teach the Wild linker to link WebAssembly (Wasm) programs. However, adding support for a new linker target is of course far from simple. Luckily, Kei already had a lot of experience with Wild, as this was his second GSoC project working on the Wild linker.

And as expected, Kei did a great job! Prior to GSoC 2026, Wild had no Wasm output support. Now, the Wasm support is sufficiently good that Wild can link a range of complex programs to Wasm. For example, it is able to link itself, SQLite, ripgrep, CPython and even the SpiderMonkey JavaScript engine.

And the linking performance is excellent! When compared to wasm-ld, the de-facto standard Wasm linker, which is based on the LLD linker, Wild is 2-15x faster, and uses 2-5x less memory. You can find more benchmarks in his final report, where he also explains that achieving this performance took quite some work, because Wild's ELF linking performance did not immediately translate to being fast when linking Wasm.

Even after this GSoC project ended, Kei has continued improving Wild's Wasm support, as well as contributing to other parts of Wild. Thank you for bringing fast linking to a new set of use-cases, Kei! You are a valued member of the Wild team.

Bringing autodiff and offload into Rust CI

The goal of this project was to build the experimental std::autodiff module, which implements automatic differentiation of Rust code, and the std::offload module, which allows offloading Rust code to GPUs, in the CI of the Rust compiler.

Shota has already been contributing to this area in the months leading up to this year's GSoC, so he had no issues getting started. After valiantly fighting with our CI, he made both modules available in the official nightly compiler channel. Rust users can thus now simply install them with a single rustup command, rather than having to compile rustc from source! This is a huge benefit for gathering feedback on these features in the wild, as it makes it much easier to experiment with them.

Beyond making these projects available via rustup, he also worked on making it simpler to build projects that use std::offload. The original compilation pipeline was derived from Clang's support for GPU offload, so it involved calling clang helper binaries. That is clearly not optimal, so Shota worked on replacing it by directly calling the underlying LLVM APIs from rustc.

The last piece of his project was enabling testing in CI for both offload and autodiff, which already helped prevent multiple regressions. Right now, the GPU offload pipeline only tests the CPU setup side, but his work allowed us to talk to GPU hardware vendors, two of which already agreed to provide the Rust Project with GPU hardware so that in the future we can also run those tests on dedicated GPUs.

Overall, the project was a success. Shota had to show a lot of patience and endurance when working with various CMake quirks and CI challenges. We would like to extend a special thank you for that!

Debugger for Miri

miri is a tool that can detect undefined behavior (UB) in unsafe Rust code. It used to have a prototype of an interactive debugger called priroda, but it became unmaintained some time ago. The goal of this project was to re-establish priroda, to make it easier to use miri interactively, debug how and why is UB introduced in Rust programs, and also help rustc developers analyse MIR, one of the intermediate representations of Rust programs.

Mohamed successfully reached both main goals of this project. He created a miri interface for the gdb debugger, which allows debugging programs executed by miri. You can set breakpoints, step through individual instructions, analyse MIR local variables and inspect the debugged program's state, with miri acting as a "virtual CPU". So when you encounter your next miri UB report, you can actually debug the program state that caused UB by running priroda! Mohamed also created a command-line interface for priroda, which is used to test it in miri's CI. This will ensure that it keeps working properly and will not bitrot again.

Mohamed also added basic support for the Debugger Adapter Protocol, which allows using priroda together with IDEs such as Visual Studio Code. Currently, connecting priroda to an IDE is a bit involved, so the next step would be to build a plugin for an IDE that would allow showing all the extra state that miri has, which gdb doesn't know how to interpret.

Debugging programs executed by miri is very exciting. Thank you for all your work!

Implementing impl and mut restrictions

The goal of this project was to implement two new language features that were accepted in RFC 3323:

  • Impl restrictions (impl(crate) trait Foo {}), which allow restricting in which modules you are able to implement a given trait.
  • Mut restrictions (struct Foo { pub mut(crate) bar: u8 }), which allow restricting in which modules you are able to mutate a given field.

Ryosuke had to familiarize himself with the compiler's codebase and do a lot of implementation work to get both features working. This included implementing parsing of the new syntax, recording its metadata in the compiler's Abstract Syntax Tree, implementing resolution of the restricted crate paths, and finally adding checks to enforce both restriction rules. This all amounted to around five thousand lines of Rust code!

Once the new features were implemented and available on nightly, Ryosuke wrote a Call for Testing blog post. It had to be very exciting to announce to the world that he implemented two new features from scratch, and people can now play with them on the nightly channel! After the blog post was published, users reported a few issues, which Ryosuke promptly fixed. He also contributed to rustdoc to improve how the impl and mut restrictions are rendered in the generated documentation pages.

Now that the features are available on nightly, feedback is being gathered on them, with the plan being to stabilize them in 2027. If you would like to provide feedback on the impl and mut restrictions, you can do so in this issue.

The goal of this project was clear, and Ryosuke achieved it completely. Awesome work!

Improving Ergonomics and Safety of serialport-rs

The goal of this project, which was done in collaboration with the Rust Embedded Working Group, was to improve the API ergonomics and type safety of the serialport-rs crate.

The serialport project is a bit different from some other projects which participated in this year's Google Summer of Code. It is more on the side of "facility management" than rocket science. Its challenges are taming finicky operating system APIs and making good abstractions, rather than full-blown software engineering.

Tanmay led many fruitful discussions about potential design decisions, and did a lot of field research on how other crates that depend on serialport actually use its API, to figure out how to make the upcoming breaking API changes less disruptive and painful for them.

Tanmay also worked on sanding off some rough edges of the serialport API, such as replacing Boxed trait objects with enums, and making errors explicit in places where the legacy API used to panic. Some of the work still waits for integration. It is targeting the upcoming serialport 5.0 release.

Christian (Tanmay's mentor) found working alongside Tanmay very motivational, and he was able to catch up with a lot of other unrelated work in order to integrate Tanmay's changes. He noted that working with Tanmay was fun and productive, and that it was pure joy to see him grow from his first contribution in the bonding phase to seasoned contributions. Thank you, Tanmay!

This was my first Google Summer of Code as well. It was challenging, motivating, fun and definitely worth it. I would show up again.

Christian Meusel

libc: transition differing bit-width time and offset variants and deprecate bug-prone constants

libc is a foundational crate that is used by thousands of other crates across the Rust ecosystem. For the past several years, its maintainers have been attempting to publish a stable 1.0 release of the crate. Because this crate has to support interoperating with C on many compilation targets with weird quirks, this requires an enormous amount of work. We already had a GSoC project for moving libc towards 1.0 last year, and the goal of Adam's project was to make more progress in that direction.

And we can happily say that progress was indeed made! The original goal of Adam's project was to deprecate a number of constants that were difficult to maintain and cleanup inconsistencies in handling 32-bit/64-bit variants of file and time handling functions[2]. Adam was able to resolve that in a few weeks, and then he moved on to resolve a number of various issues blocking the 1.0 milestone.

Saying that Adam was productive would be… an understatement. We could follow his work in real-time, because he was posting daily GSoC reports for 116 days (!) in a row. The diligence and consistency with which he tackled so many complicated issues was an inspiration to us. In fact, sometimes the libc reviewers were not able to keep up!

You can find all Adam's libc pull requests here. You might notice that many of them have lots of comments, and even more discussions happened in GitHub issues and on Zulip. In addition to implementing the necessary code, making changes that affect interesting C platforms and architectures also required him to do a ton of research, and a lot of considerations of edge cases and possible backwards compatibility breakages. His work also took him on a whirlwind tour of diverse platforms; Adam's contributions spanned FreeBSD, NetBSD, Android, Hurd, NuttX, AIX, VxWorks, Fuchsia, L4Re and more.

Even though more work will still be required to get libc to 1.0, Adam's contributions made a significant dent in this effort. Thank you very much for all your work!

Linker scripts are arcane commands that can be used to instruct a linker to construct the final binary in a custom way. The Wild linker used to implement only basic linker script functionality, which Vishruth greatly expanded during his GSoC project.

Thanks to that, Wild can now successfully link Linux kernel modules, and it's even close to being able to link the Linux kernel itself! Vishruth's work involved greatly expanding expression evaluation support, location counters, program headers, etc. In addition to adding support for many missing bits of linker script support, Vishruth implemented support for partial linking (-r) as this is needed in order to link Linux kernel modules. You can find all the linker script features that he implemented here.

Adding all those peculiar features to Wild required performing many cross-cutting changes to its codebase, which Vishruth managed to do really well. And it was not at the cost of performance! Wild now manages to link Linux kernel modules such as btrfs or amdgpu 2-4x faster than the BFD (GNU ld) or LLD linkers[3]. You can find more benchmarks in his final report.

Vishruth has continued to contribute to Wild even after his GSoC project ended and has become a valued contributor to the project. Thank you very much for making Wild better, Vishruth!

Migrating rust-analyzer assists to SyntaxEditor

rust-analyzer uses rowan for its syntax trees, a library for Concrete Syntax Trees. rowan has served us well over the years, but it has some downsides that cause major performance problems and also smaller convenience issues. In particular, memory fragmentation causes rust-analyzer to consume much more memory than it actually needs (in the order of redundant gigabytes).

We wanted to rewrite rowan to change that. The problem was that many parts in rust-analyzer touch syntax trees, and in particular assists (code actions) of which we have a lot. They also use mutable tree editing, a feature of rowan that can be convenient but was a major obstacle in rewriting it to be more efficient.

During his project, bit-aloo first migrated all rust-analyzer code from mutable tree editing to SyntaxEditor, a new abstraction that was created specifically to avoid mutation. After that, he removed mutable editing from rowan altogether.

The migration process is incremental and rowan is still inefficient in various aspects. However, we believe that the parts which required changes to large amounts of code are behind us, and we can complete the migration with more focused changes. bit-aloo is continuing to work on that even after his GSoC project ended. And he was also accepted as a member of the rust-analyzer team, so congratulations! We would like to thank him for his work, and hope to continue working with him to improve rust-analyzer for everyone.

Port std::arch test suite to rust-lang/rust

The std::arch module of the Rust standard library contains various intrinsic functions that provide architecture and vendor-specific functionality. Their implementation lives in the stdarch repository, which is included as a Josh subtree in the main compiler repo (rust-lang/rust).

When someone changes the intrinsics in the main repo, breakages can silently creep in, because the main repo does not execute all the stdarch tests in its CI. That can cause friction when we synchronize the changes between the two repositories. Ideally, we would run all of stdarch's tests in the main repo to detect any issues as soon as possible.

And that is where Sumit came in! Over the course of his GSoC project, he examined the (very!) complex ways the intrinsics are tested in stdarch, and started porting them over to the main repo. But this was not a matter of copying over a few lines of YAML. The compiler repo uses its own build system (called bootstrap) and a CI workflow structure that is quite different from the stdarch repository. Sumit thus had to first understand what is each test suite doing, reimplement it as a part of bootstrap and finally integrate it into the compiler's CI workflows.

We run ~90 CI jobs across several hardware architectures, operating systems and OS versions. Getting all that to work while testing vendor-specific intrinsics required a lot of diligent work… and a lot of patience. It took five months for one of the test suites to finally succeed on the main repo's CI, and it required Sumit to completely refactor how the suite worked, for it to be more compatible with bootstrap. As another example, a different test suite that Sumit also got to work required running the tests in an Intel CPU emulator on CI.

Sumit ported most of the stdarch test suites to bootstrap, including intrinsics verification against reference C implementations, checks for automatically generated source code being up-to-date, and various unit tests for individual intrinsics and CPU feature detection logic. There is one remaining test suite to port, but the ported tests are already helping us find issues, which is great.

You can check Sumit's rust-lang PRs here; notice that he was also contributing to other parts of Rust in the meantime! Sumit's work will help us make LLVM upgrades and land intrinsics changes more easily. Thank you, Sumit!

Reorganizing tests/ui/issues

The goal of this project was to get rid of a very scary place in the Rust compiler's test suite: a directory called tests/ui/issues. It contained thousands of test files based on historical Rust issues, dating back all the way to 2012. The problem with those tests was that they were named only with an issue number (e.g. issue-37291). When such a test would break, it was quite difficult to figure out what was it even supposed to test, because there was little context attached to it.

Even though we already had a GSoC project to tackle this problem last year, due to the sheer amount of uncategorized tests, there were still over four hundred tests left in this directory by the time zedddie started working on this project.

For each test in this directory, zedddie had to first understand what was the test doing and dig through its original GitHub issue and any other related issues and pull requests. Then they had to figure out what would be the best location to move the test under, search for any possible duplicates, come up with a descriptive name for the test (that tends to be very difficult!), and finally add a short description to the test that explains the context of why the test was added in the first place. Rinse and repeat, for hundreds of tests.

zedddie rolled up their sleeves, and categorized over 300 tests in 30 pull requests. Together with other contributors helping out over this summer, zedddie was finally able to remove the last batch of uncategorized tests at the start of August. As zedddie said in their final report: "What is left to do: nothing, tests/ui/issues is gone".

While this work might sound a bit bland, its positive effect on the maintainability of our test suite, and improving the lives of Rust compiler contributors, is considerable. It is also apparently a great way to learn about cool niche edge cases of Rust that most Rust developers might not know about! After migrating the remaining uncategorized UI tests, zedddie continued contributing to the next trait solver and const generics. Thank you very much for helping us finally get rid of tests/ui/issues, zedddie!

Utilize debugger APIs to improve debug info test accuracy and error reporting

The Rust debugging experience leaves a lot to be desired. One reason why we struggled to improve it is that many of our debuginfo test suites did not even run in CI for the past few years, as they were difficult to maintain. When something changed in the type visualizers, the implementation details of a standard library type or in debugger internals, someone had to go through tens of tests and manually update them. This was not sustainable, so the test suite rotted.

Luckily for us, Anthony took that affront personally. Even before his GSoC project, he spent considerable effort on improving the quality of debug visualizers of various Rust standard library types, and gained an impressive amount of knowledge about debuggers.

During his GSoC project, he completely overhauled how our debuginfo test infrastructure works, both for LLDB and GDB. Thanks to his improvements, we can now easily "rebless" the reference output when something changes, which makes it much easier to keep our test suite up-to-date. The new tests also record much more information about the internal state of the debugger recorded during the test, which helps with diagnosing issues that cause type visualizers to stop working.

In addition to redesigning the test suite, Anthony also actually managed to re-enable the LLDB tests on both Windows and Linux in our CI again (after several years of being skipped there!). This will allow us to avoid debuginfo regressions and make it easier to improve the Rust debugging experience. Oh, and it goes without saying that while doing all that, he also made numerous improvements to Rust type visualizers. You can find his PRs here.

Anthony's amazing contributions and his deep expertise in debuggers led to him being invited to the Rust Compiler team, which was much deserved. What else can we add, other than: thank you for your awesome work!

XDG path support for rustup

The support for platform-specific directories (platformdirs) has been a long-standing feature request for rustup, to have more fine-grained filesystem directory choices. This project investigated how that issue could be resolved.

The real complexity of this feature request lies within a decade worth of different assumptions that previous iterations of rustup and cargo have made: it turns out that quite a lot of them are no longer true if we were to implement platformdirs support truly faithfully.

Guicheng first worked on cleaning up the interactions between rustup self uninstall and $CARGO_HOME/bin as a warmup. This issue itself is a very good demonstration of all that compatibility baggage: somehow $CARGO_HOME/bin has become the default binary directory for all things Rust, so rustup ended up installing itself and all its proxies into it. As a result, it has been super annoying to many that everything in $CARGO_HOME is gone with rustup self uninstall.

After that, he started the actual implementation of platformdirs support in rustup. This was no easy task! Not because the code itself was hard to write, but because it takes a lot of effort to figure out what should be the desired behavior in the first place. Also, in order to determine the appropriate platformdirs semantics, some prior research is often required to see how things became what they are today, so the decoupling of different concerns could proceed, and migration paths could be designed accordingly. It was in this sense that this is different from a typical GSoC project, and luckily thanks to the effort he has put into this project, we have not only come up with the final patch, but also discovered and fixed some tiny preexisting issues with this long-untouched aspect of the rustup codebase along the way.

Thank you for digging into this complex topic, Guicheng!

Conclusion

We would like to thank all contributors who have participated in Google Summer of Code 2026 with us! Your contributions helped make Rust better, and introduced you as new members of our community, which is a reason to celebrate. We are looking forward to participating in GSoC again next year!

We would also like to thank Google for organizing the Google Summer of Code program and for allowing us to host a large number of projects every year. And last, but not least, we would like to thank all the Rust mentors who were tirelessly helping our contributors to complete their projects. Without you, Rust GSoC would not be possible.

  1. 97.5% over a three-year period, if you are wondering (40 out of 41 projects were successful so far). ↩

  2. If you were wondering, then yes, this is actually related to the famous 2038 problem :) ↩

  3. In case you were wondering, the mold linker does not support linker scripts required to link Linux kernel modules at the moment, so Vishruth did not compare the performance of Wild against mold here. ↩

  •  

Announcing rustup 1.29.3

Nov. 3, 2026 · The Rustup Team

The rustup team has published a new point release of rustup, 1.29.3.

Rustup is the recommended tool to install Rust, a programming language that empowers everyone to build reliable and efficient software.

What's new in rustup 1.29.3

This new patch release has fixed some minor regressions over the previous one:

  • Fixed "Illegal instruction" error on certain x86_64 CPUs without SSE4.1. pr#5162

  • Fixed "lock() not supported" error on Solaris. pr#5167

Further details are available in the changelog!

How to update

If you have a previous version of rustup installed, getting the new one is as easy as stopping any programs which may be using rustup (e.g. closing your IDE) and running:

$ rustup self update

Rustup will also automatically update itself at the end of a normal toolchain update:

$ rustup update

For this particular release, if you have encountered any issues during self-update, you may try to update rustup manually by following the steps below:

  1. Download the corresponding rustup-init binary for your platform.
  2. Make double sure that no other rustup processes are running.
  3. Replace your existing rustup binary with the downloaded rustup-init renamed to rustup.
    • As a reminder, this binary is often located at $HOME/.cargo/bin/rustup, or $CARGO_HOME/bin/rustup if you have overridden the $CARGO_HOME environment variable.

If you don't have it already, you can get rustup from the appropriate page on our website.

Rustup's documentation is also available in the rustup book.

Caveats

Rustup releases can come with problems not caused by rustup itself but just due to having a new release.

In particular, anti-malware scanners might block rustup or stop it from creating or copying files, especially when installing rust-docs which contains many small files.

Issues like this should be automatically resolved in a few weeks when the anti-malware scanners are updated to be aware of the new rustup release.

Thanks

Thanks again to all the contributors who made this rustup release possible!

  •  

How to fund Rust contributors

TLDR:

There is always so much to do in the Rust Project! New features to implement, bugs to fix, performance optimizations to make, issues to triage… It takes a lot of work to keep our high bar of quality while also shipping new features and improvements. And the workload keeps increasing with the continuous growth of Rust adoption.

As is unfortunately common in open source projects, a large fraction of that work is currently done by volunteers. That causes various issues, such as people burning out or complex features taking many years to ship.

Unsurprisingly, one of the best ways to help improve that is to get more people who work on Rust funded! This year, several funding initiatives started ramping up, and we already experience many positive effects. They helped us improve the maintenance status of critical Rust projects and make significant progress on several long-awaited features, such as allocators, the never type, the next trait solver or the Polonius borrow checker. However, only a fraction of people working on Rust are currently funded for their work.

We know that there are Rust users, both individuals and companies, who would like to help improve Rust by funding it, but it might not always be obvious how exactly to do that. This post thus provides suggestions on how you can support the development of Rust in three ways.

If you're unsure which option fits, or would like to discuss funding on behalf of your company, please contact the Funding team.

Fund general maintenance

Maintenance is a continuous process required to keep Rust (or any software project for that matter) working. It consists of two main activities. The first is general housekeeping, such as ensuring that Continuous Integration works, dependencies are kept up-to-date and issues, bugs and security incidents are promptly triaged and resolved. The second is performing refactorings and code reviews, which ensures that the codebase will be able to evolve and continuously ship new features, without accruing extensive technical debt.

Having quality maintenance is crucial for the long-term health of Rust, yet it is typically very difficult to find funding for it; funders often want to sponsor specific new features, rather than just "keeping the lights on". However, without sufficient maintenance, it becomes much more difficult, and much slower, to ship new features. Improving the maintenance of Rust thus also helps us ship features and improvements faster, though the effects are sometimes only observable over longer periods of time.

Funding maintenance is a great choice if you want to let us decide how best to improve Rust. The Rust Funding team continuously looks for Rust teams and projects that require help with their maintenance efforts, and allocates available funds to Rust maintainers who could help improve the situation.

To fund Rust maintainers, consider donating to the Rust Foundation Maintainers Fund or sponsoring the RustNL Maintainers Team.

Fund specific features and improvements

There are always many Rust improvements and new features that are in progress. We track some of those initiatives through Rust Project Goals. Each Goal describes a specific improvement to Rust: what it is, who is going to develop and drive it forward and which Rust teams are willing to help reviewing and maintaining it.

Some of those Goals are looking for funding. Either to help their owner dedicate time and focus to push the given feature forward, or to fund maintainers who will perform refactorings necessary to unblock that feature and review its implementation. Often funding is required for both. When a Goal is properly funded, the probability that it will achieve its milestones is much higher, which then translates in the given improvement being shipped sooner to Rust users.

Funding Project Goals is a great choice if you want to help push forward specific Rust features or Rust toolchain improvements that will benefit you or your company.

To fund contributors working on Project Goals, consider taking a look at Goals that need funding and reaching out to their funding contact. If in doubt how to proceed, reach out to the Rust Funding team.

Fund individual contributors

There are hundreds of people in the Rust Project who continuously work on improving Rust. Some of them have personal fundraising profiles (e.g. on GitHub Sponsors), which allows you to support them directly.

There might be various reasons for supporting a specific person. Maybe they are maintaining a crate that you love to use, maybe you like reading their blog posts, or maybe you are simply a fan of the work that they do. Whatever the reason, we are sure that they would appreciate your support!

Funding individual contributors is a great choice if you want to directly support Rust contributors whose work you appreciate.

To fund individual Rust contributors, check out their GitHub profile or personal website to see what options there are for supporting them. You can also examine our list of Rust Project members who can be supported through GitHub Sponsors[1].

Conclusion

We hope that this post is useful to anyone who would potentially like to invest funds into making Rust better. You can also find a shortened version of the "how to fund Rust guide" at the Funding page of the Rust website, or check out a RustConf 2026 talk by Niko Matsakis about how Project Goals and their funding works.

If you are a company looking to fund the development of Rust, consider joining the Rust Commercial Network. Joining the network is free, and Rust Foundation membership is not required. It can help you find opportunities for funding improvements in specific areas related to Rust usage (e.g. user interfaces, language interop, supply chain security, etc.) and collaborate on funding Rust development with other companies.

There are of course other ways of helping Rust other than donating funds. For example, you can check out Project Goals that are looking for contributors or more generally check our contributor onboarding guide. Companies can also help by giving employees paid time to contribute upstream, or by hiring Rust contributors to do that work. One possible model is the Engineer in Residence role, that was recently established in the Rust Foundation.

  1. Note that this list is not exhaustive; there are some Project members who accept donations through GitHub Sponsors (or another fundraising platform), but did not opt into being listed on this page. ↩

  •  

Announcing rustup 1.29.2

Oct. 20, 2026 · The Rustup Team

The rustup team is happy to announce the release of rustup version 1.29.2.

Rustup is the recommended tool to install Rust, a programming language that empowers everyone to build reliable and efficient software.

What's new in rustup 1.29.2

This is a small release which only updates the rustls-webpki and webpki-root-certs dependencies to address the problematic eMudhra intermediate certificates issue described in CVE-2026-111943.

See also RUSTSEC-2026-0382.

Further details are available in the changelog!

How to update

If you have a previous version of rustup installed, getting the new one is as easy as stopping any programs which may be using rustup (e.g. closing your IDE) and running:

$ rustup self update

Rustup will also automatically update itself at the end of a normal toolchain update:

$ rustup update

If you don't have it already, you can get rustup from the appropriate page on our website.

Rustup's documentation is also available in the rustup book.

Caveats

Rustup releases can come with problems not caused by rustup itself but just due to having a new release.

In particular, anti-malware scanners might block rustup or stop it from creating or copying files, especially when installing rust-docs which contains many small files.

Issues like this should be automatically resolved in a few weeks when the anti-malware scanners are updated to be aware of the new rustup release.

Thanks

Thanks again to all the contributors who made this rustup release possible!

  •  

Announcing Rust 1.99.1

Oct. 15, 2026 · The Rust Release Team

The Rust team has published a new point release of Rust, 1.99.1. Rust is a programming language that is empowering everyone to build reliable and efficient software.

If you have a previous version of Rust installed via rustup, getting Rust 1.99.1 is as easy as:

rustup update stable

If you don't have it already, you can get rustup from the appropriate page on our website.

What's in 1.99.1

Rust 1.99.1 fixes the following regressions from prior releases:

  • Missing symbol wcslen: Some Rust targets were getting link errors for missing wcslen, and that function is now present in Rust's compiler-builtins library as a fallback.

    The LLVM loop-idiom-wcslen pass may introduce calls to wcslen as an optimization, and some Rust targets did not have that function readily available in libc or elsewhere. The UEFI targets in particular were broken for any project using std, as some of these calls were introduced in the precompiled standard library, but the issue was also reproducible for similar user code patterns.

  • Miscompilation of FFI bool return values on x86_64: Rust functions returning bool through FFI could leave bits 1-7 of the return register unspecified, causing callers to observe raw values other than 0 or 1. This return type is now marked for zero-extension again, as required by the ABI.

  • rust-lld failed to load libLLVM on macOS: This linker could fail to run if the compiler was invoked without the rustup proxy, as there was no direct setting for the appropriate library search path. Now rust-lld is built with an rpath entry for the directory containing libLLVM.dylib.

If you'd like to help us out by testing future releases, you might consider using the beta (rustup default beta) and nightly (rustup default nightly) channels locally and in your CI. Please report any bugs you might come across!

Contributors to 1.99.1

Many people came together to create Rust 1.99.1. We couldn't have done it without all of you. Thanks!

  •  

Demoting i686 Windows targets to std-only

Oct. 2, 2026 · Mateusz Mikuła, Ralf Jung on behalf of the Compiler team

With Rust 1.100.0, the following changes to 32-bit Windows targets will happen:

  • i686-pc-windows-msvc Tier 1 with host tools target will be demoted to Tier 1 without host tools.
  • i686-pc-windows-gnu Tier 2 with host tools target will be demoted to Tier 2 without host tools.

Builds of the standard library will continue to be distributed, but host tools such as the compiler will be no longer available. i686-pc-windows-msvc as a Tier 1 target still undergoes CI testing.

To build 32-bit Windows binaries, cross-compiling from a still-supported host toolchain (such as a 64-bit Windows ones) will be required from now on.

Background

Desktop and Server 32-bit only x86 CPUs are no longer sold for over 15 years, and general 32-bit Windows support has ended in October 2025. This means that the development platforms these targets are meant for hardly exist these days, and even if they do exist they typically aren't capable enough for development.

Even on the modern x86_64 hardware, building i686 Windows toolchains has proven to be problematic. We have encountered compiler binaries crashing when built with the i686 MSVC target, and the GNU C++ toolchain failing with OOMs during LLVM build.

Considering all these things, cross-compiling these targets from a better supported one is what we have found to be the best solution forward. As part of that, we stopped producing host tools for these targets. For the time being, the prebuilt standard library is still available, and in case of i686-pc-windows-msvc still tested on CI.

What Changes?

After Rust 1.100, it will no longer be possible to install toolchains on 32-bit Windows hosts. We recommend cross-compiling from a still-supported host (such as a 64-bit Windows toolchain) instead. Other 32-bit platforms are not impacted by this change.

For more details about these demotions, see RFC 3999 for i686-pc-windows-msvc demotion, and MCP 1020 for i686-pc-windows-gnu demotion.

  •  

Announcing Rust 1.99.0

Oct. 1, 2026 · The Rust Release Team

The Rust team is happy to announce a new version of Rust, 1.99.0. Rust is a programming language empowering everyone to build reliable and efficient software.

If you have a previous version of Rust installed via rustup, you can get 1.99.0 with:

$ rustup update stable

If you don't have it already, you can get rustup from the appropriate page on our website, and check out the detailed release notes for 1.99.0.

If you'd like to help us out by testing future releases, you might consider updating locally to use the beta channel (rustup default beta) or the nightly channel (rustup default nightly). Please report any bugs you might come across!

What's in 1.99.0 stable

extern "C" variadics

Rust 1.99.0 stabilizes defining C-ABI variadic functions with "C" and "C-unwind" ABIs. Variadic functions defined this way use a variable argument list (...) and accept an arbitrary number of arguments. Rust could already call externally-defined variadic functions (e.g., libc::printf). With Rust 1.99, these functions can now be written in Rust itself:

/// SAFETY: must be called with (at least) 2 i32 arguments.
unsafe extern "C" fn sum(mut args: ...) -> i32 {
    // SAFETY: guaranteed by the caller.
    let a = unsafe { args.next_arg::<i32>() };
    let b = unsafe { args.next_arg::<i32>() };
    a + b
}

fn foo() -> i32 {
    unsafe { sum(0i32, 2i32) }
}

The type of ... is VaList, which is ABI-compatible with the C va_list type across targets. What types can be read from a VaList is guarded by the VaArgSafe trait.

For more details on c-variadic functions, see the Reference. This release also stabilizes support for defining naked variadic functions with non-"C" ABIs, which must be written via inline assembly.

Layout information from raw pointers

This release settles the safety requirements for retrieving the size and alignment on raw pointers to both Sized (trivially safe, already possible on stable) and non-Sized types.

This is done by stabilizing three functions:

Recommend against round-trip unleaking after Box::leak

While there are no changes to the language semantics in Rust 1.99, we have updated the documentation on Box::leak to recommend against patterns that later deallocate that memory. This was done because such code was found to have problematic interactions with current and future potential compiler optimizations, and is especially problematic with the upcoming stabilization of custom allocators. Instead, Box::into_non_null or Box::into_raw should be preferred.

This guidance also applies to other leak functions in the standard library.

Stabilized APIs

Other changes

Check out everything that changed in Rust, Cargo, and Clippy.

Contributors to 1.99.0

Many people came together to create Rust 1.99.0. We couldn't have done it without all of you. Thanks!

  •  

Announcing a Maintainer in Residence: Scott Schafer for the Cargo team

At the end of August, we announced our first Maintainers in Residence, Rust Project contributors who are funded for their upstream contributions and maintenance work from the Rust Foundation Maintainers Fund (RFMF). Since then, the Rust Leadership Council has dedicated more funds from its Project Priorities budget to RFMF, and together with AWS also providing additional funds, this allowed us to open a new full-time Maintainer in Residence (MiR) position to support the Cargo team. We would like to thank the Rust Leadership Council, AWS, and also the Rust Foundation for providing us with this opportunity! If you would like to help us hire more maintainers to improve Rust, consider donating to RFMF.

This post explains why we chose to support the Cargo team specifically, and introduces Scott Schafer, the new Cargo Maintainer in Residence.

Why Cargo?

The new MiR full-time position is dedicated to helping with the maintenance of Cargo, our build system and package manager. The Cargo project is deeply involved in many new Rust features, improvements, and Project Goals. Combined with its cross-cutting nature, where it has to support many different use-cases and integrate with several other tools, it takes a lot of work just to keep up with its maintenance needs, let alone support so many feature requests and proposed changes.

Because of that, the Cargo team has sometimes struggled with meeting its maintenance demands. You might remember that for several years, it actually held a feature freeze, to reduce Cargo's internal tech debt, perform necessary refactorings, go through the issue and pull request backlog, and come up with scalable internal development and design processes, so that they could eventually go back to even thinking about adding new features.

Recently, some changes occurred within the team, which made it more difficult for them to meet their maintenance baseline. Some members of the team left, while others lost their dedicated funding for working on Cargo maintenance and had to scale down their involvement. The Funding team thus considered it very important to support this team, given that we had an opportunity to do so. And thus we decided to hire a full-time maintainer to work on Cargo for (at least) the next 12 months.

Even though we know that a single full-time maintainer will not completely solve the maintenance struggles of the Cargo team, we hope that it will improve the situation, and provide a bit of a relief for the team.

Introducing Scott Schafer

We are very happy to welcome Scott Schafer (@muscraft) into the Maintainer in Residence role! Scott has joined the Cargo team three years ago, and apart from working on Cargo, he is also the lead of the Rust Docker team, which prepares official Docker images for every Rust version.

Apart from working on general maintenance of Cargo, Scott has implemented Cargo's Workspace inheritance feature, and has also spearheaded a complex multi-year effort to switch the rendering of diagnostics in the Rust compiler to use the annotate-snippets crate. This effort has been completed in the Rust 1.93.0 release. Thanks to it, the same diagnostics interface can now be shared between the compiler and Cargo (and also other tools), which amongst other things unblocked further development of the Cargo linting system, which has now been stabilized and will ship in the Rust 1.100.0 release.

Everyone we talked about was very excited about Scott becoming a Cargo Maintainer in Residence, and we share that feeling. We wish Scott all the best in his new role, and we are very happy that we can support his maintenance work.

Here is what Scott thinks about it:

I am incredibly excited to work on Cargo full-time! There have been so many things that I wish I could've worked on over the years, that I will now be able to get to. I hope that my efforts will bring Cargo into a more maintainable state.

Conclusion

We are incredibly happy that we keep getting more funds for the Rust Foundation Maintainers Fund, which allows us to support Rust Project contributors. The funding team will be working with the supported maintainers, and also the funders, to ensure that they are all happy with the arrangement, so that we can secure stable funding for Rust maintenance for years to come.

If you would like to help us support more Rust maintainers, consider donating to RFMF!

  •  

GitHub Actions leaking secrets when Miri output is cached

The Rust Security Response Team was notified that Miri stores all environment variables to target/, allowing secrets to persist in caches.

While not necessary a vulnerability in and of itself, when paired with GitHub Actions caching behavior, it is possible for this to expose secrets to PRs.

Overview

GitHub Actions makes it possible to cache directories between runs. Typical setups allow CI runs on main (and other branches) to write to cache, and PRs can only read from cache (preventing cache poisoning). Rust projects tend to speed up CI by caching binaries built by cargo install and sometimes the contents of target/.

PR CI can be triggered by anyone who can open PRs on your repository. GitHub requires maintainer approval for the first PR, but future PRs will rerun CI on every push. Anyone who has previously landed a change can trigger a CI run extracting information from cached target/ and then cover their tracks by pushing a second commit to the PR.

GitHub sometimes hides overwritten commits in its UI, making this kind of attack harder to detect. CI run logs and overwritten commits are also deleted after a few months.

When cargo miri is invoked, Miri needs to retain build-relevant environment variables between runs[1]. The current code to do so achieves this by storing all environment variables to target/. This, of course, persists when target/ is cached.

If your environment contained secrets, these can now be accessed by PRs via the cache.

Our fix

Our short term fix for this is to make Miri only preserve CARGO_* environment variables (excepting CARGO_*_TOKEN) and OUT_DIR. In the longer term, Miri and cargo may figure out better ways to inform Miri of the relevant list of environment variables. Note that this patch may not be available on nightly yet.

We also performed an ecosystem scan of GitHub repositories and identified 1 repository with this issue and 7 repositories that do not appear to be vulnerable but should be cautious anyway. We have reached out to those maintainers.

Am I affected?

It is likely that our scan was imperfect, so we recommend you check your own GitHub Actions setups if you run Miri.

You are vulnerable if:

  • You run cargo miri in CI
  • The step that runs cargo miri has access to secrets as an environment variable:
    • By being passed in to the step itself as an environment variable
    • By being set in env for the workflow
    • By being passed in to a previous step that persists it in the environment somehow
  • The workflow being used caches the target directory, usually done via actions/cache or swatinem/rust-cache
  • The cache is accessible to PRs (common and often the intended use case)

Possible quick fixes include:

  • Disabling cache for that job.
  • Scoping secrets to steps in that job that do not call Miri.
  • Temporarily disabling Miri.

Once done, please clear the cache. Consider rotating any secrets that might have leaked.

The Miri release in the upcoming nightly (2026-09-22) will no longer have this problem.

Even if you do not run Miri, ensure jobs that can write to public caches do not have access to secrets. Many tools do not have special handling for secrets, and assume the entire environment can be written to the filesystem.

Threat model

We consider it bad practice to have a cache that can easily be tainted by secrets.

If caching target/, it is worth making sure that the inputs to processes that create target/ (anything invoking cargo) do not have secrets available. It is generally rare for standard cargo build/test subcommands to need any secrets or tokens[2], so this is mostly a matter of being careful about having secrets exposed as environment variables to the entire job.

Cargo/Miri/Rust does not guarantee that environment variables will be safe from being copied into target/. While we are treating this as a security issue and patching it out of an abundance of caution, this is not something you should rely on in general. Beyond official Rust tooling, it is possible for build scripts to be doing things that lead to the environment being stored in compilation artifacts.

Acknowledgements

Thanks to Predrag Gruevski of OpenAI for reporting this issue to us. Furthermore, the ecosystem scan was performed using Codex access and credits donated by OpenAI, which we also thank them for.

Issue triage and remediation was performed by Manish Goregaokar, Ralf Jung, Ben Kimock, Weihang Lo, Jacob Finkelman, Walter Pearce, Josh Stone, and Mark Rousskov.

  1. Miri is invoked multiple times by cargo miri for complicated reasons ↩

  2. In theory it could come up with build scripts reading from the network ↩

  •  

Be alert: targeted attacks on prominent Rustaceans

We believe that there is an ongoing campaign targeting rust-lang members and owners of popular crates that is attempting to compromise devices and accounts in order to use them to publish malware.

What we've seen

A video call is set up for something positive — maybe for a job, maybe for a project, maybe for a contract opportunity — and then that's used as a vector to either get the target to install something on their computer (such as a purportedly missing audio codec) or execute another command (for example, via putting a command on the clipboard).

These attackers are setting up new but legitimate seeming company profiles, including plausible LinkedIn presences, in order to pass cursory inspection.

A previous attack of this form targeted many prominent Rust developers in June, and, last month, the arrayref crate was briefly compromised through similar attacks. At this moment we do not know if these are all a part of the same campaign.

This attack style is known to be used by the DPRK, and has been seen outside of the Rust community as well.

What you can do

Please take extra care in the near term. Be appropriately suspicious of cold outreaches, and ensure that any calls you have with new people are on platforms you trust — ideally, try to be the one who sets up the call on a platform you already use.

Please also re-check that your accounts look normal: MFA enabled, no unexpected logins on platforms that can track that, and so on.

If you have any concerns about your accounts, please reach out to help@crates.io (for crates.io account concerns) and/or security@rust-lang.org (for any other concerns). We're very happy to help.

  •  

Rust debugging survey 2026 results

One of the biggest challenges Rust developers report in our annual surveys is a subpar debugging experience. So, back in February, we ran our first Rust Debugging Survey, in the hopes of identifying how Rust developers are using debuggers and what problems they are facing when doing so. We received over 2,300 responses, and we'd like to thank everyone who took the time to participate in the survey!

In this report, we'll go over some of the results of the survey. If you'd like, you can also check out the complete results of the survey.

If you'd like to skip ahead to any particular section, you can do so with this index:

Who Uses Debuggers?

The first step to making sense of the survey results is understanding who took the survey. We asked respondents to rate their Rust expertise, from "Never used it" to "Advanced". Over 80% reported themselves as "Advanced" or "Intermediate", split roughly evenly between the two:

We also asked respondents if they currently use or have used debuggers in Rust. Over 46% said they currently do, with the remaining responses split between "have in the past" and "never have". That means that over half of respondents do not currently use a debugger for Rust!

Categorized by expertise, the responses reveal that roughly half of "beginners" have never used debuggers in Rust! On the other hand, nearly half of "advanced users" currently do use debuggers in Rust:

For respondents who indicated they had previously used Rust but no longer did, we asked if challenges with debugging support were why they stopped. For nearly 3%, the answer was "yes", with an additional 24% reporting debugging issues as being partially responsible (though mind the small response count; most respondents were active users of Rust):

How Are Debuggers Used?

Knowing what debuggers developers are using and how is another important part of understanding the challenges they face. To this end, we asked respondents how they were debugging their programs. Unsurprisingly, most developers make use of print debugging and the dbg! macro. Excluding those, using lldb inside an IDE was the most popular choice, followed by gdb on the command line:

We can get a more detailed breakdown of these results if we include the operating system on which the respondents use a given debugging approach. We examine this from two different angles. The first angle being, "On operating system X, what percent of responses are using debugger Y?". Print debugging and the dbg! macro are consistently the top two yet again, but looking beyond that, things get more interesting. On Linux, using gdb on the command line was the most popular choice by a thin margin, beating lldb in an IDE by only 0.4%. On Windows, Windows Subsystem for Linux (WSL), and macOS, lldb in an IDE was the top pick by at least 6%, making it a very popular choice in general. On Windows, the three least popular choices were the command line debuggers (gdb CLI, lldb CLI, and BugStalker), and on both Windows and macOS the third most popular pick was, "I don't know". Those who were debugging on operating systems not listed (Other) most frequently used some kind of special embedded debugger or gdb:

The other angle we can look at these responses from is, "For users of debugger X, what percent of responses are using it on operating system Y?". For most debuggers, Linux makes up the largest portion of uses, ranging from about 45% to about 77%, followed by Windows, then macOS. The most notable exceptions are WinDbg and the Visual Studio debugger, which are primarily used on Windows, and lldb, which is used more on macOS than Windows in an IDE and on the command line:

To the 6 respondents who use WinDbg on Linux: we wish you luck!

As for how people actually use their debugger of choice, the aggregate results are not particularly surprising. Roughly 87% of users are using debuggers for stepping line-by-line through programs and a little over half of users are using debuggers to obtain stack traces from hung/crashed processes. Only a quarter of the respondents use a debugger to debug async code. That might be partially caused by the async Rust debugging experience being clumsy and incomplete, or it could just be that users aren't writing much async code:

If we break these results down by expertise, we can learn a bit more about usage patterns. As users become more experienced with Rust, their use of debuggers for learning purposes decreases, and they get more stack traces from crashed processes:

The final bit of insight into how Rustaceans use debuggers is if they are debugging programs that use Rust alongside other programming languages. For 44% of respondents, the answer is "yes", which is a pretty high number!

As for which languages those are, C dominates the scene at a little over 70%, followed by C++ at about 43% and Python at about 20%:

Challenges

Instead of diving right into asking, "what problems do you face when using debuggers?", or something to that effect, we first asked respondents why they decide against using debuggers whenever they do, including for reasons that aren't necessarily "problems with debuggers".

The most commonly reported reason was that it was easier or faster to use logs or print debugging to solve problems, reported by a little over 81% of respondents. This could partially be explained by the open responses, which featured complaints that debuggers were too difficult to set up and/or use (especially on Windows, when dealing with Web Assembly, or in embedded contexts) and sentiment suggesting that small and/or simple problems just don't really need a debugger. It does leave one wondering if the user experience could be made convenient enough to dethrone print debugging, but it seems hard to beat something so intuitive. This is followed by roughly 37% of respondents who write code that Just Works. Fair enough. After that, about 26% of respondents indicated that they've decided not to use debuggers in situations where the language features they were working with had poor support. This is slightly more than issues with standard library types, at about 22%, which is slightly more than issues with external library types, at about 20%:

As stepping through code was anticipated to be one of the most common uses for debuggers, we directly asked respondents if they faced any issues when doing so. A little over 51% of respondents said they did! Of those who reported that they experienced issues stepping through code, we asked when they were experiencing issues. Async code was the most common case reported at slightly over 28%, followed by code involving macros at about 23%. The least common case reported was code involving function pointers, at almost 6%:

We also directly asked respondents which types in the standard library were hard to work with, if any. This was an open-response question, and reading through the responses, some particularly common complaints were with enums and collections, particularly std::collections::HashMap and std::vec::Vec. This is also visible in the word cloud in the full report.

We asked respondents to indicate which pain points, if any, they have encountered when using debuggers with Rust. At slightly over 74%, poor representation of values was the most common pain point by a decent margin, followed by being unable to print variables at just over 55%:

Debugger Visualizers

We asked respondents to indicate if they were library authors, and if so, if they were aware of and using the debugger_visualizer attribute. Nearly 62% of respondents indicated that they were library authors who were not aware of this attribute:

For those who indicated that they were library authors who knew about the attribute but did not use it, we also asked why. This represented a much smaller fraction of respondents, so keep that in mind! That said, half of these library authors indicated that they didn't have the time to maintain visualizer attributes, and just under half indicated they didn't know how to write visualizer scripts:

For those of you who have been reading this section asking yourself what the debugger_visualizer attribute is, you can read up on it in The Rust Reference: Debugger Attributes. The quick explanation is that the debugger_visualizer attribute can be applied to modules or the crate root to embed files in the debug information which improve the display of values with certain debuggers. The two currently supported file types are Natvis files, used by Microsoft debuggers such as WinDbg, and GDB "pretty printers", which are structured Python scripts used by GDB.

Thanks to your participation in this survey, we've gained some great insights about how Rustaceans are using debuggers and what issues they are facing. For example, knowing that such a high number of users are dealing with poor representation of values pairs well with knowing which standard library types are causing issues, knowing that many library authors haven't heard of the debugger_visualizer attribute, and knowing that many of those who have but don't use it either don't know how or don't have time to maintain visualizer scripts.

Looking to the future, the survey results have suggested that there are a few notable ways we could most significantly improve the debugging experience in Rust, such as:

  • Fixing the way enums are represented by debuggers so they show actual variants
  • Fixing the way collections (such as HashMap) are represented by debuggers so they show their contents, rather than their implementation details
  • Fixing the way string types (such as String and CString) are represented by debuggers so they render as text, rather than their implementation details
  • Improving the async debugging experience, particularly with stack traces
  • Improving stepping through certain state machines (such as iterators and Futures)
  • Providing documentation on basic set up and use of some common debuggers

A common suggestion that could resolve those first three points is to use the Debug implementation of types to display them in debuggers. There are challenges to that approach, such as the fact that the Debug implementation is not present in the final binary unless it is actually used somewhere in the program, but it isn't impossible. Notably, this is already supported by the BugStalker debugger (given the same condition that the Debug implementation must actually be used), which some of you first heard about from the survey! It also appears to have some support for async, with plans to expand.

One notable way the debugger experience is currently being improved is through the ongoing Google Summer of Code project improving how we test debug info and visualizer scripts, making it easier to maintain and improve our own visualizer scripts and general compatibility with visualizer scripts without silent breakage or regressions.

Once again, we'd like to thank everyone who took the time to participate in the survey!

  •  

Announcing Rust 1.98.1

Sept. 3, 2026 · The Rust Release Team

The Rust team has published a new point release of Rust, 1.98.1. Rust is a programming language that is empowering everyone to build reliable and efficient software.

If you have a previous version of Rust installed via rustup, getting Rust 1.98.1 is as easy as:

rustup update stable

If you don't have it already, you can get rustup from the appropriate page on our website.

What's in 1.98.1

Rust 1.98.1 fixes a miscompilation in vtable generation.

In Rust 1.98.0, in some circumstances, rustc would incorrectly generate a trait object vtable with a null pointer where a function pointer should be. This leads to undefined behavior in the emitted code. In some cases this may 'just' cause segfaults due to the null pointer being loaded, but it is possible for it to be justification for arbitrary effects (as is typical for UB).

If you'd like to help us out by testing future releases, you might consider using the beta (rustup default beta) and nightly (rustup default nightly) channels locally and in your CI. Please report any bugs you might come across!

Contributors to 1.98.1

Many people came together to create Rust 1.98.1. We couldn't have done it without all of you. Thanks!

  •  
❌