ee7

I think that returning an error union here isn't particularly interesting. I think most of the exercises that currently contain (or previously contained) a single-item error set are better if...

Remove `QueenError` and the input validation tests. They weren't interesting - there's no good reason for `init` params to be `i8` when the valid range is between 0 and 7...

The tests currently use `Signal`: https://github.com/exercism/zig/blob/0c4805d47884d6ec790bc1cae43f13e9e9fadaeb/exercises/practice/secret-handshake/test_secret_handshake.zig#L7 But the docs only talk about "actions": https://github.com/exercism/zig/blob/0c4805d47884d6ec790bc1cae43f13e9e9fadaeb/exercises/practice/secret-handshake/.docs/instructions.md#L3-L8

Not just #343 or #344, but both together.

Not just https://github.com/exercism/zig/issues/342 or https://github.com/exercism/zig/issues/344, but both together.

But if we continue to state that Exercism supports 0.11 only, we would allow the tests to fail with zig master. This would give more visibility to code that is...

This can help catch some bugs (for example, an example solution with a function that returns a reference to function-local memory).

This would be a check that is allowed to fail, and should probably be required before we make the test runner use the x86_64 backend (see https://github.com/exercism/zig-test-runner/pull/72). The testing command...

From https://ziglang.org/download/

https://www.openmymind.net/learning_zig/