Open alamb opened 1 month ago
I've been playing around with this, and found that next_batch
is only used in benchmarks and tests, while as_any
truly does seem to be completely unused. For next_batch
we can add #[cfg(any(feature = "experimental", test))]
to silence the warning. For as_any
, we could either remove it (it appears unused, and array_reader
does not appear to be public), or simply allow dead code if we think as_any
might be of use down the road. We could also just throw #[allow(dead_code)]
on both. @alamb any preferences here?
Thanks @etseidl
I wonder if the issue is that the ArrayReader
trait itself is not pub
by default (maybe it needs an experimental flag?) - if it was a public trait then the error shouldn't happen
Maybe we can add a
#[cfg(any(feature = "experimental", test))]
To the overall trait?
Maybe we can add a
#[cfg(any(feature = "experimental", test))]
To the overall trait?
Tried it and that caused a huge ripple (which I find odd since the whole module is marked experimental, but I think I still don't fully understand the mechanics there 🤷). I'm also puzzled over why the warnings starting appearing in the first place. Was it a code change or just improvements in unused code detection? I haven't wanted to start bisecting to run that thought to ground 😅.
Describe the bug While publishing
52.2.0
I noticed this warning: #5998I have also seen it locally
To Reproduce
Expected behavior No warnings
Additional context