jwr1@kbin.earth to Rust@programming.dev · 4 days agoAnnouncing Rust 1.91.0 | Rust Blogblog.rust-lang.orgexternal-linkmessage-square24fedilinkarrow-up161arrow-down10
arrow-up161arrow-down1external-linkAnnouncing Rust 1.91.0 | Rust Blogblog.rust-lang.orgjwr1@kbin.earth to Rust@programming.dev · 4 days agomessage-square24fedilink
minus-squareFizzyOrange@programming.devlinkfedilinkarrow-up2·4 days agoThe non-strict versions also panic by default, but only in debug mode. So if you were willing to use abs() you should be willing to use strict_abs(). Arguably a bit of a mistake to have the “obvious” function names be surprisingly unsafe, but I guess it’s too late to fix that.
minus-squareflying_sheep@lemmy.mllinkfedilinkarrow-up1·3 days agoI think this is where safety and performance are actually in conflict. I think guaranteeing a panic leads to much worse code gen in this case.
minus-squareFizzyOrange@programming.devlinkfedilinkarrow-up1·3 days agoYes that’s true, but the point is you should have to opt in to surprising or risky behaviour.
The non-
strictversions also panic by default, but only in debug mode. So if you were willing to useabs()you should be willing to usestrict_abs().Arguably a bit of a mistake to have the “obvious” function names be surprisingly unsafe, but I guess it’s too late to fix that.
I think this is where safety and performance are actually in conflict.
I think guaranteeing a panic leads to much worse code gen in this case.
Yes that’s true, but the point is you should have to opt in to surprising or risky behaviour.