An independent software engineer applying compiler technology to the data space. I most often write about Rust, compilers, performance optimizations, and data querying technology.
The RSS's url is : https://predr.ag/atom.xml
1970-01-01 08:00:00
1970-01-01 08:00:00
Last month, I gave a talk titled "SemVer in Rust: Breakage, Tooling, and Edge Cases" at the FOSDEM 2024 conference.
The talk is a practical look at what semantic versioning (SemVer) buys us, why SemVer goes wrong in practice, and how the cargo-semver-checks
linter can help prevent the damage caused by SemVer breakage.
TL;DR: SemVer is impossibly hard for humans, but automated tools can cover our greatest weaknesses.
1970-01-01 08:00:00
My last post covered the key cargo-semver-checks
achievements from 2023. Here are the biggest challenges that lie ahead!
Many of the remaining challenges in cargo-semver-checks are obvious: we all want more lints, fewer false-positives, etc. etc. Let's set those aside.
Instead, let's talk about four non-obvious challenges we have yet to tackle:
1970-01-01 08:00:00
2023 was a big year for cargo-semver-checks
! We saw ecosystem-wide adoption in projects of all shapes and sizes: the tokio
and PyO3
ecosystems, company-backed OSS projects from companies like Amazon and Google, and even in cargo
itself. Here's a look back at the highlights of 2023!
1970-01-01 08:00:00
cargo-semver-checks
v0.25 squashes nearly all bugs related to doc(hidden)
items — its most common source of false-positives. What does doc(hidden)
mean in Rust, and why was handling it correctly so hard?
1970-01-01 08:00:00
This post is coauthored by Tomasz Nowak and Predrag Gruevski. It describes work the two of us did together with Bartosz Smolarczyk, Michał Staniewski, and Mieszko Grodzicki.
Anecdotally, cargo-semver-checks
is a helpful tool for preventing the semver violations that every so often cause ecosystem-wide pain.
This is why it earned a spot in the CI pipelines of key Rust crates like tokio
, and also why the cargo
team hopes to integrate it into cargo
itself.
While anedotal evidence is nice, we wanted to get concrete data across a large sample of real-world Rust code.
1970-01-01 08:00:00
A few days ago, I started polls on Mastodon and Twitter whether adding a new private type, or an import, can ever be a major breaking change. The consensus was that this should be impossible.
I agree with that. It should be impossible.
I've discovered a way to cause a previously-public type or function to disappear from a crate's public API by making innocuous-seeming changes like adding a private type or adding an import, etc. It is not a hypothetical problem, either — I've found at least one real-world Rust project that has been affected by it.
1970-01-01 08:00:00
For the longest time, I thought that "sealed trait" in Rust was a singular concept implementable in one specific way. To prevent downstream crates from implementing your traits, you make the traits sealed — done, end of story. I was wrong! It turns out there are multiple ways to seal traits, forming a pleasant spectrum of options:
1970-01-01 08:00:00
Happy April 1st! This post is part of April Cools Club: an effort to publish genuine posts on topics our usual audience would find unexpected. The tech content will be back soon!
Over the many years I spent heavily involved in intern and full-time recruiting at $PREVIOUS_JOB
, multiple people have commented something to the effect of: "How come Predrag always gets the best people?"
This post is a series of vignettes showing three of the less-obvious ideas that gave us an edge,
1970-01-01 08:00:00
We've already explored some of the dark corners of Rust semantic versioning on this blog:
1970-01-01 08:00:00
This post describes work in progress: how cargo-semver-checks
will benefit from the upcoming query optimization API in the Trustfall query engine. Read on to learn how a modern linter works under the hood, and how ideas from the world of databases can improve its performance.
1970-01-01 08:00:00
1970-01-01 08:00:00
1970-01-01 08:00:00
1970-01-01 08:00:00
1970-01-01 08:00:00
1970-01-01 08:00:00
I had a lot of fun spending nights-and-weekends time participating in the HYTRADBOI Jam, a global hack week aimed at building "exciting and weird" data-centric solutions to familiar problems. The name HYTRADBOI might sound familiar: the jam is associated with the same conference where I gave my "How to Query (Almost) Everything" talk talk in April this year.
I jammed on two projects: one solo and one with a friend. The projects ultimately were very successful and mostly-successful, respectively.
1970-01-01 08:00:00
The saying usually goes: "If at first you don't succeed, try, try again." But in the Safari web browser under the right conditions, trying again after succeeding once can get you in trouble. This is my recent debugging adventure.
1970-01-01 08:00:00
I recently built cargo-semver-checks
, a linter that ensures crates adhere to semantic versioning. This is why and how I built it.
Fearless development is a key theme throughout Rust. "If it compiles, it works", fearless concurrency, etc.
But there's one aspect of Rust (and nearly all other languages) that isn't entirely fearless yet: cargo update
, upgrading the versions of the project's dependencies.
1970-01-01 08:00:00
1970-01-01 08:00:00
1970-01-01 08:00:00
1970-01-01 08:00:00
1970-01-01 08:00:00