Dominic

> > custom test runner probably does not want it's stdout to be consumed by the build runner > > Why not? This allows conditional logic in the build runner...

> > * Is there a way to make a test that will invoke the build system from a subdirectory, (i.e. simulate `cd /path/to/zig/test/standalone/test_runner_path/subdir && zig build test`)? > >...

Yes, I agree that custom log and panic for wasm is not straightforward - you could either not allow it, or force the user to get it right at their...

I can try to explain a bit more about my thoughts on logging (which is the main feature of the above I am interested in at the moment). ## Logging...

For anyone wondering why xxhash uses `anytype` currently, the idea is to allow passing both arrays and slices (you can't use arbitrary types). Passing an array should allow the compiler...

I agree that we should have more consistency, but I don't think I like having an options struct that pretty much always has 0 or 1 fields. I don't think...

> * uses a consistent programming model for both single and multi-value hashes which does not require multiple code usage scenarios for compile time vs. runtime contexts for a simplified...

I guess the issue is that `std.Build` uses an array list, so it doesn't notice when two steps have the same name, so the root of the problem isn't really...

> Side note: one thing that would be nice from a fuzzing perspective would be the option to do everything in-memory instead of hitting the filesystem, but not sure how...

Until a nicer solution is merged, you can generate documentation for individual modules compiling static libraries rather than tests, via `zig build-lib -fno-emit-bin -femit-docs` on the CLI or using `std.Build.addStaticLibrary`...