Lightnews — Scholar-powered news

Michael Pratt @prattmic.com · 1d

Is “control handoff is easily avoided by not doing it” oversimplified? Don’t give control to anyone, but even professional maintainers add new maintainers eventually. Things like two party review requirements help reduce risk by ensuring another human reviews changes even by other maintainers.

1

Reposted by Michael Pratt

Filippo Valsorda @filippo.abyssdomain.expert · 1d

To implement robust mitigations across Geomys, I did a survey of open source project compromises in 2024/2025.

Three root causes dominate: phishing, control handoff, and unsafe GitHub Actions triggers. All three can be systematically avoided.

words.filippo.io/compromise-s...

A Retrospective Survey of 2024/2025 Open Source Supply Chain Compromises

Project compromises have common root causes we can mitigate: phishing, control handoff, and unsafe GitHub Actions triggers.

words.filippo.io

1 13 42

Michael Pratt @prattmic.com · 2d

Are you looking for slices.Collect or maps.Collect?

pkg.go.dev/slices#Collect

pkg.go.dev

1 2

Michael Pratt @prattmic.com · 3d

In all seriousness, check out Birria-Landia if you find yourself nearby.

Reposted by Michael Pratt

Go @golang.org · 4d

🥳 Go 1.25.2 and 1.24.8 are released!

📢 Announcement: https://groups.google.com/g/golang-announce/c/4Emdl2iQ_bI/m/qZN5nc-mBgAJ

📦 Download: https://go.dev/dl/#go1.25.2

#golang

$ go install golang.org/dl/go1.25.2@latest
$ go1.25.2 download
Downloaded 0.0% ( 0 / 58280426 bytes) ...
Downloaded 50.0% (29140213 / 58280426 bytes) ...
Downloaded 100.0% (58280426 / 58280426 bytes)
Unpacking go1.25.2.linux-riscv64.tar.gz ...
Success. You may now run 'go1.25.2'
$ go1.25.2 version
go version go1.25.2 linux/riscv64

20 50

Michael Pratt @prattmic.com · 10d

Time to go simplify my jj + Gerrit guide…

5

Michael Pratt @prattmic.com · 10d

jj v0.34 is out, including Gerrit upload support!

github.com/jj-vcs/jj/re...

Release v0.34.0 · jj-vcs/jj

About jj is a Git-compatible version control system that is both simple and powerful. See the installation instructions to get started. Release highlights Support for uploading changes to Gerrit ...

github.com

1 3 14

Michael Pratt @prattmic.com · 10d

There is also nice set of resellers that I’ve learned when I see they are the only seller listed it just means the hotel is sold out.

1

Michael Pratt @prattmic.com · 10d

I hate this so much. I’ve slowly grown heuristics for ignoring them, but it’s so annoying to have lovely map full of lies.

1 4

Reposted by Michael Pratt

Merriam-Webster @merriam-webster.com · 10d

We are thrilled to announce that our NEW Large Language Model will be released on 11.18.25.

660 8.3K 25K

Michael Pratt @prattmic.com · 10d

No, I wouldn't reach for execution tracing first for memory usage issues. I suggest looking at a heap profile from runtime/pprof, as well as the breakdown of memory types in runtime/metrics (pkg.go.dev/runtime/metr...) /memory/classes/... metrics to verify that the leak is in Go memory.

metrics package - runtime/metrics - Go Packages

pkg.go.dev

1

Reposted by Michael Pratt

Jake Bailey @jakebailey.dev · 11d

Fun little Go compiler CL merged today: go.dev/cl/706655

Uninlined generic functions have a "dict" arg, since Go generics are neither erased nor monomorphized, but instead instantiated for each "GC shape" (e.g. T=*int and T=*float64 get the same code, but T=int32 and T=int64 do not).

1 4 19

Reposted by Michael Pratt

Steve Klabnik @steveklabnik.com · 13d

#jj-vcs part 1: what is it

andre.arko.net/2025/09/28/j...

jj part 1: what is it

I’ve been working on a blog post about migrating to jj for two months now. Rather than finish my ultimate opus and smother all of you in eight thousand words, I finally realized I could ship increment...

andre.arko.net

1 9 47

Michael Pratt @prattmic.com · 15d

I'd love to hear from folks about your experiences. Do you use execution tracing often. If not, is it due to lack of need, lack of documentation, missing information, tooling issues, etc?

2 1 5

Michael Pratt @prattmic.com · 15d

If you haven't been keeping up, Go 1.21 reduced overhead from ~10% to ~1%. 1.22 overhauled the format to improve reliability and add information. 1.25 has a built in trace flight recorder. And we have an active proposal (go.dev/issue/62627) for a trace parsing package to enable arbitrary tooling.

proposal: x/debug/trace: add package for parsing execution traces · Issue #62627 · golang/go

As part of #60773 (tracking issue #57175) I've been working on a new parser for the execution tracer, and to save work down the line I've also been trying to come up with a nice API that would work...

go.dev

1 3 12

Michael Pratt @prattmic.com · 15d

Go's execution tracer is a woefully underutilized tool. It contains tons of information about what is happening at any given moment that you won't get with a profile.

In addition to showing flight recording capabilities, this post provides a nice example of how to use tracing to diagnose a problem.

Go @golang.org · 15d

“Flight Recorder in Go 1.25” by Carlos Amedee and Michael Knyszek — https://go.dev/blog/flight-recorder

#golang

1 9 30

Reposted by Michael Pratt

GopherCon Europe @gopherconeu.bsky.social · 16d

Spotlight on @prattmic.com from the @golang.org team and his talk on on Faster Go Maps With Swiss Tables.

youtu.be/aqtIM5AK9t4?...

#gopherconEU #golang

Faster Go Maps With Swiss Tables - Michael Pratt | GopherCon EU 2025

YouTube video by GopherCon Europe

youtu.be

1 4

Michael Pratt @prattmic.com · 18d

I got one of these a few days ago attempting to masquerade as a GitHub account security alert.

1 1

Michael Pratt @prattmic.com · 19d

Regarding ARM64, someone on go.dev/issue/71255 tried a prototype implementation and actually found it to be slower than the portable implementation, but it certainly warrants another look.

runtime: swiss map SIMD implementation for non-amd64 · Issue #71255 · golang/go

https://go.dev/cl/626277 (in 1.24) added SIMD implementations for swissmap (still 8-byte) control word operations for amd64. We could do similar for other architectures, if the SIMD implementations...

go.dev

1

Michael Pratt @prattmic.com · 19d

Great talk, I enjoyed all the performance comparisons!

While SIMD is helpful, the boost over the portable version is smaller than one might expect. go.dev/cl/626277 shows a few percent improvement over the portable version. The portable version is still parallel, which is the big win.

Gerrit Code Review

go.dev

3 1

Reposted by Michael Pratt

Bryan Boreham @bboreham.bsky.social · 19d

Video is up now - my take on how map is implemented in Go, and what changed from Go 1.23 to 1.24 and 1.25.

youtu.be/M05t7Q6LbFs

Swiss Maps in Go - Bryan Boreham

YouTube video by GopherCon UK

youtu.be

1 3 9

Reposted by Michael Pratt

roland.zone @roland.zone · 22d

I did a talk at the UK GopherCon last month about what my team does, and I only let my laptop fall asleep twice! www.youtube.com/watch?v=oLtq...

Go Security – Past, Present, and Future - Roland Shoemaker

YouTube video by GopherCon UK

www.youtube.com

1 8 20

Reposted by Michael Pratt

Haskell programming language @haskell.org · 23d

There is no thriving #Haskell community without its members of the LGBT community, and on that there cannot be any compromise.

9 49 240

Michael Pratt @prattmic.com · 22d

To be clear, arm64 is not being ignored. The intention is to only ship something with amd64 and arm64 support (to ensure the general idea works across different SIMD architectures, if nothing else). I simply meant the prototype implementation that exists today only supports amd64.

1 1

Michael Pratt @prattmic.com · 22d

I imagine you’ve seen it, but if not I believe the Go SIMD experimental proposal should support SIMD popcnt, allowing you to use the wider instruction directly from Go. (Ignoring that the proposal doesn’t implement arm64 yet)

go.dev/issue/73787

proposal: simd: architecture-specific SIMD intrinsics under a GOEXPERIMENT · Issue #73787 · golang/go

Update (08/20/2025): A preliminary implementation of AMD64 low-level SIMD package is being developed on the dev.simd branch. You're welcome to check it out and try it with your use cases. Feedback ...

go.dev

1 1 2