[rust-dev] Deprecating rustpkg
catamorphism at gmail.com
Thu Jan 30 11:26:54 PST 2014
On Wed, Jan 29, 2014 at 1:50 PM, Tony Arcieri <bascule at gmail.com> wrote:
> On Wed, Jan 29, 2014 at 1:13 PM, Jan Niklas Hasse <jhasse at fastmail.fm>
>> I'm very shocked and also very much against removing it. I've just
>> started with Rust a few months ago and actually rustpkg was the thing
>> that I liked the most. I know it's buggy and unfinished, but at least
>> the idea looked wonderful to me.
> Just my 2c: I'd argue the complete opposite.
> Package managers are things that are often created quite early in the
> lifecycle of a language, and their design mistakes can haunt users of the
> language and potentially doom the language itself. Once a package manager is
> entrenched, typically there's no turning back, and you are left forever
> trying to fix the mess you started with.
> Moreover, there's the complex relationships of package managers, dependency
> resolvers, and secure software update infrastructures to consider. I think
> rustpkg was written in the absence of these considerations. I think
> successful packaging/software update systems consider all these problems
> up-front, so the MVP is at least future proof when it comes time to consider
> things like, say, security.
In the interest of documentation (especially for the benefit of
whatever person or people end up writing the next package manager),
can you explain what specific architectural issues in rustpkg preclude
the addition of (better?) dependency resolution and secure software
> If we aren't certain that rustpkg is a solid foundation, now is the time to
> scrap it and build a better one. Otherwise we'll be stuck with it forever.
> Tony Arcieri
> Rust-dev mailing list
> Rust-dev at mozilla.org
Tim Chevalier * http://catamorphism.org/ * Often in error, never in doubt
"If you are silent about your pain, they'll kill you and say you enjoyed it."
-- Zora Neale Hurston
More information about the Rust-dev