If you want to read more about the bisect idea itself, this link has a bunch of interesting use cases:

https://research.swtch.com/bisect

TLDR: you can bisect on more than just "time".

We don't allow? Who doesn't? And what are they going to do about it?

The majority of nations? majority of people on earth? We are going to a multilateral world and to win a war you need secure the appeal of majority. If the majority think your war is illegal they can cut you off from the world economy.

It is a distributed consensus-based algorithm, and the young people who are writing those algorithms will shape the future of governance.

The majority of the world thinks the Russia-Ukraine war is illegal.

The majority of the world thinks the Israeli/US-Iran war is illegal.

They meant the Gulf. You cross the straight into the Gulf, then what?

Iran hit an E-3's antenna in an airport in Riyadh with a precision strike. Was it not worth defending?

How many tankers inside the Gulf do they need to hit before the rest of the world decides it's a bad idea to send new tankers to the Gulf?

And if new tankers don't go into the Gulf, then it's simply not open for business. That's their leverage.

I've added it, to one of my repos, and yes, it's slower than using emulation.

Particularly for my use case, Go cross compilation, QEMU and binfmt work really well together.

Still, for some things, it's nice to test on actual hardware.

Here's a workflow so you can see both approaches working: https://github.com/ncruces/wasm2go/blob/main/.github/workflo...

GitHub CI lets me test my package, for free, in all these: https://github.com/ncruces/go-sqlite3/wiki/Support-matrix

Which is actually useful.

GitHub also generously gives me a bunch of free CI, in exchange for whatever they benefit from me being there.

It's worth $50 just this month, according to them, but I don't see anyone else offering the mac runners that account for most of it.

For all the complaints, I test my packages that actually need it across dozens of architecture and OS combinations with a mix of runners, nested virtualization and qemu binfmt, all on their free platform.

AI and Copilot increase the load on git workflows.

> Go compatibility. So code is valid Go code. No exceptions.

But then they have block scoped defers.

I built my latest side project (a Wasm to Go "transpiler") precisely as a way to push the limits of what I could do with an LLM/agent.

It sped me up (and genuinely helped with some ideas) but not 10x.

The bits I didn't design myself I definitely needed to inspect and improve before the ever eager busy beaver drove them to the ground.

That said, I'm definitely impressed by how a frontier model can "reason" about Go code that's building an AST to generate other Go code, and clearly separate what's available at generation time vs. at runtime. There's some sophistication there, and I found myself telling them often "this is the kind of code I want to generate, build the AST."

I also appreciated how faster models are good enough at slightly fuzzy find and replace. Like I need to do this refactor, I did two samples of it here, can you do these other 400? I have these test cases in language X, converted 2, can you do the other 100? Even these simple things saved me a lot of time.

In return I got something that can translate SQLite compiled to Wasm into 500k lines of Go in about a month of my spare time.

https://github.com/ncruces/wasm2go

Not so far. All my contributions where welcomed.