> Version pinning is a bad practice because it slows deployment of security updates.
How is that not true for stdlibs?
> It also only works for end applications, as libraries cannot reasonably pin their own dependencies.
Really?
> Version pinning is a bad practice because it slows deployment of security updates.
How is that not true for stdlibs?
> It also only works for end applications, as libraries cannot reasonably pin their own dependencies.
Really?
And Rust ints can overflow. It's a tradeoff.
>> Version pinning is a bad practice because it slows deployment of security updates.
> How is that not true for stdlibs?
Because, if without pinning, the package manager always gets the latest available version and installs it for all applications on the system once and for all. When pinning dependencies, system administrators need to wait for every single application to update its pin.
There are provisions to do math that won't overflow or that can saturate. For example https://doc.rust-lang.org/std/primitive.u32.html#method.chec...
> Really?
At least not usefully if types from those dependencies appear on the API surface.