Cocalus

An other API choice, would be such a closure and a Duration. The password(connection string) would get refreshed when a password is needed and the cached password is older than...

I believe I can do time limited authentication caching with a closure. It'll probably need interior mutability since I expect the closure to be just a Fn instead of FnMut....

I missed those. Doesn't look like either claims to be complete for mp3 yet. Sonata is testing mp3 and puremp3 is version 0.1.0. Sonata looks like a self contained replacement...

I'm not that familiar with the MP3 standard but I'd start with a more complete test vectors and fuzzing. https://github.com/lieff/minimp3 seems to have a pretty good set of vectors and...

I fixed some issues like this in #249 which might help but I suspect this is a different issue with the same root cause of generated samples not respecting the...

It might be enough for now to mention in the docs that between the interfaces for #[cfg(target_feature = ..)], is_target_feature_detected! and #[target_feature(enable = "")], only the later accepts (requires?) multiple...

> What does that buy us? That multiplies the memory requirements of the SIMD run-time with the number of features (~64x increase on x86), and would require the compiler to...

I think if the relaxed load and check could be inlined it would speed up things with lots of dynamic checks. the other ideas are probably detrimental. > > The...

I tend to use futures::stream::StreamExt::for_each_concurrent a lot when I need to batch any async work. Would this Batch type work with that API. tokio-postgres's implicit approach should.

The original is about 177MB and I doubt I could share it, I could try to trim it but I suspect any tools I have that accept it would likely...