Twitter | Pretraživanje | |
Andrew Gallant
Brainstorm with me. How plausible is it for quickcheck to break ties with rand?
I am no longer happy about depending on the rand crates. There is too much churn, too many crates, and IMO, worst of all, there is no desire to add a minimal version check to their CI. Which means ...
GitHub GitHub @github
Reply Retweet Označi sa "sviđa mi se" More
\u221e 25. kol
Odgovor korisniku/ci @burntsushi5
What if there were an alternative, simpler rng implementation with less breakage than the rand crate?
Reply Retweet Označi sa "sviđa mi se"
Andrew Gallant 25. kol
Odgovor korisniku/ci @hdevalence
Right, yeah, I think that's what I meant by there being some design space in the ecosystem for something exactly like that. :-)
Reply Retweet Označi sa "sviđa mi se"
Sean Griffin 🏳️‍⚧️ 25. kol
Odgovor korisniku/ci @burntsushi5
Commenting here since it's tangential, and I don't want to derail your issue, but I think we really need to decide as a community what we consider to be the norm for compiler version support. I think we can be more aggressive than other langs, but different crates have diff needs
Reply Retweet Označi sa "sviđa mi se"
Andrew Gallant 25. kol
Odgovor korisniku/ci @sgrif
I agree. To be clear though, if it were only about MSRV, I could overlook that, at least for rand. I still like the LTS idea that boats/aturon proposed. It provides a rallying point. Adding MSRV to Cargo.toml in order to improve failure modes is also a good start.
Reply Retweet Označi sa "sviđa mi se"
Brian Smith 25. kol
Odgovor korisniku/ci @burntsushi5
Using ‘getrandom’ directly would probably help you and may be enough.
Reply Retweet Označi sa "sviđa mi se"
Andrew Gallant 25. kol
Odgovor korisniku/ci @BRIAN_____
It would need the `Rng` trait (or similar) layer on top of it, ideally. I guess that's probably the (longer) pole in the tent.
Reply Retweet Označi sa "sviđa mi se"
Florian Gilcher ∠(・.-)―〉 →◎ 25. kol
Odgovor korisniku/ci @burntsushi5
As `rand` is a project controlled crate, we can also lobby for stabilisation. You're not the only one frustrated.
Reply Retweet Označi sa "sviđa mi se"
Dirkjan Ochtman 25. kol
Odgovor korisniku/ci @Argorak @burntsushi5
+1 from me, seems like this is much more annoying than any MSRV issues that you might also have. The rand folks should agree on a minimal API already and start taking backwards compatibility seriously as befitting a project with its amount of downstream users.
Reply Retweet Označi sa "sviđa mi se"
Brook Heisler 25. kol
Odgovor korisniku/ci @burntsushi5
I'd support this as a Quickcheck user. Rand has been more of a pain in my backside for than it's worth, and I hardly even use it except transitively through Quickcheck.
Reply Retweet Označi sa "sviđa mi se"
Tony “Abolish ICE” Arcieri 🦀 25. kol
Odgovor korisniku/ci @redattack34 @burntsushi5
As a criterion user, I'd love to see it stop depending on rand
Reply Retweet Označi sa "sviđa mi se"