David Hoppenbrouwers

I submitted a [patch](https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=257301) for 0.4.0 a while ago and it seems to work fine so far. I don't know what the holdup is though admittedly I haven't been following...

I'd prefix macros with `unsafe_` if they use unsafe code without being able to verify whether it's actually safe. As for verbosity: something that annoys me is the need to...

IMO `select!` should be renamed to `select_random!` to emphasize that the ordering is random. I too spent a few days figuring out why my tests weren't reproducible until I read...

> Can you tell me more about how random selection caused a problem for you? Was starvation of the later futures desirable? In my case it made a specific test...

I have a SM-G960F. It took me a while before I could flash TWRP but I found that unplugging and replugging the device may help if it times out.

I've found that using `must_use` in combination with `forbid(unused_must_use)` is effective as a crappy form of linear types, e.g.: ```rust #![forbid(unused_must_use)] #[derive(Debug)] #[must_use = "Must be manually destroyed with `T::destroy()`"]...

You can't accurately represent 1.3 and 1.23 as a FP numbers and thus `1.3 * 10 != 13`. Use `is_equal_approx` to compare FP numbers (as you suggested yourself). To demonstrate...

It should be pretty easy to register (you don't need an email), so I don't think this is necessary.

I think the best way is to make an account with a randomized username when a non-registered user comments. The storage used up by a single account should be negligible....

I didn't make it possible to make threads anonymously :facepalm: