Hyrum’s Law in Golang

Solution to the specifically mentioned problem: Don't use string-based errors, use sentinel errors [1].

More generally: Don't produce code where consumers of your API are the least bit inclined to rely on non-technical strings. Instead use first-level language constructs like predefined error values, types or even constants that contain the non-technical string so that API consumers can compare the return value againnst the constant instead of hard-coding the contained string themselves.

Hyrum's Law is definitely a thing, but its effects can be mitigated.

[1]: https://thomas-guettler.de/go/wrapping-and-sentinel-errors

In your example, the onus is on the consumer not the provider. I could still be writing code that checks if `err.String() == "no more tea available."`. I agree, I shouldn't do that, but nothing is preventing me from doing that. Additionally, errors.Is is a relatively recent addition to Go, so by the time people would check for errors like this, it was just easier to check the literal string. But as an API provider in Go, you cannot prevent your consumers from checking the return values of .String().

Unfortunately true. The Go maintainers might not agree with me on this, but I think in this case consumers have to learn the hard way. Go tries to always be backwards compatible, but I don't think that trying to be backwards compatible with incorrect usage is ever the right choice.

So the people who decided to make a stringly type error with `errors.New("http: request body too large")` and make you suffer, now can remove a stringly typed error and make you suffer even more? What would the lesson be? What would consumers learn?

I don't understand your point. The lesson is "don't rely on magic strings, instead rely on exported and documented constants, otherwise your code might break".

My point is that a few years ago there was no exported and document constant. The lesson should be "provide sensible tools, otherwise your consumers will have to rely on implementation details for the most basic expected stuff".

>My point is that a few years ago there was no exported and document constant.

Then the feature didn't exist. Figuring out undocumented implementation details to "make it work" is asking for it to be broken in the future. So if you are unwilling or unable to support fixing it in the future then don't do that.

If it is "the most basic expected stuff" then quite literally make the determination that it isn't ready for use. A lot of Go was and maybe still be half baked and not ready for production. It is ok to recognize that and not use it.

I am glad that your circumstances are such that you can just stop working on a project when the tooling it uses turns out to be inadequate, wait five years, and then come back when it improves.

Unfortunately, many people can't really do that: when the ecosystem turns out to be somewhat inadequate in a project that's already been in use for couple of years, their options are either "just make it work one way or another, who cares if it's a hardcoded string, we have to ship the fix ASAP" or "rewrite it all in Rust/X, allegedly their ecosystem is production-ready".

> "just make it work one way or another, who cares if it's a hardcoded string, we have to ship the fix ASAP"

Sure, but now that there's a "correct" way to do this, you don't get to complain that the hacky thing you did needs to keep being supported. You fix the hacky thing you did, or you make peace that you're still doing the hacky thing, problems it causes and all.