[rust-dev] Why are generic integers not usable as floats?

Matthew McPherrin mozilla at mcpherrin.ca
Thu Jun 19 16:16:43 PDT 2014

fn main() { println!("{}", 16777217f32) }

This program prints 16777216.  So I think allowing integer literals doesn't
really change anything, since you can already type unrepresentable float
literals.  That said, this ought to at least trigger a warning like
type_overflow does for integers.

On Thu, Jun 19, 2014 at 4:06 PM, Cameron Zwarich <zwarich at mozilla.com>

> Not all integer constants can be perfectly represented as floating-point
> values. What do you propose in that case, just a hard error?
> Cameron
> On Jun 19, 2014, at 4:02 PM, Matthew McPherrin <matt at mcpherrin.ca> wrote:
> This came up on IRC today, and it was something I've wondered in the past
> but nobody had an immediately good answer either way.
> I think it's fairly inconsistent that these two code samples aren't
> equivalent:
>     let a = 1f32;
>   let b: f32 = 1;
> It's fairly annoying in my opinion to have to occasionally add a .0 after
> floating point literals.
> Especially since we're getting rid of integer fallback in RFC 30, I think
> this issue ought to be thought about.
> _______________________________________________
> Rust-dev mailing list
> Rust-dev at mozilla.org
> https://mail.mozilla.org/listinfo/rust-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/rust-dev/attachments/20140619/ea3785ab/attachment.html>

More information about the Rust-dev mailing list