Rushmore Mushambi

> maybe create a new type under Number as NaN? We could give it explicit behavior for all of these ops I like this idea. > make sure NaN ==...

How about adding a restriction to floats in SurrealDB to say they can't be `NaN` [like what this crate does](https://docs.rs/ordered-float/latest/ordered_float/struct.NotNan.html)? This way we can sidestep issues with NaN.

Continuing the discussion from https://github.com/surrealdb/surrealdb/pull/1310#issuecomment-1267574930 > use `f32` and check when a NaN is possible I think this is error prone as we may forget to check for `NaN` in...

> One way to make them conform to the spec would be to eliminate any possibility of NaN by converting it to Value::Null like mysql seems to do. I don't...

> The issue I see with this is that it forces the NaN check to coincide with the low level mathematical operation in impl {Add, Sub, ...} for Number. Maybe...

> Are you suggesting the hypothetical change of dropping support for `Number::Float/`? No. Instead of ```rust pub enum Number { Int(i64), Float(f64), Decimal(BigDecimal), } ``` I'm suggesting something like ```rust...

@happysalada It's possible to save your environment variables in a file and make Systemd load them for you using `EnvironmentFile`.

Things I like about this approach:- - This should fix https://github.com/alexcrichton/futures-await/issues/11 just as [async_block](https://github.com/alexcrichton/futures-await/issues/11#issuecomment-326399783) works fine right now. I think this is HUGE. - I see you are using the...

> I've just pushed a branch implementing the first part of this for comparison purposes; including supporting returning both `impl Future` and `impl Stream` using a single macro: https://github.com/Nemo157/futures-await/tree/alternate-signatures Awesome!...

As you can probably tell from my RFC above, I really like the direction you are headed with this. One thing that's been bugging me even from your other experiment...