posborne / meta-rust-bin

Yocto layer for installing Rust toolchain from pre-built binaries

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

meta-rust-bin

An OpenEmebdded/Yocto layer providing pre-built toolchains for the Rust programming language.

Basic Example

A basic class for cargo-based executables is provided. The following is a simple recipe that builds the gpio-utils crate from a branch tagged with the version ${PV}:

inherit cargo

SUMMARY = "GPIO Utilities"
HOMEPAGE = "git://github.com/rust-embedded/gpio-utils"
LICENSE = "MIT"

SRC_URI = "https://github.com/rust-embedded/gpio-utils.git;tag=${PV}"
S = "${WORKDIR}/git"

LIC_FILES_CHKSUM = "file://LICENSE-MIT;md5=935a9b2a57ae70704d8125b9c0e39059"

As you can see, there is almost no overhead introduced from the cargo class beyond simply inheriting it. The cargo class adds the appropriate Rust dependencies as well as default compile and install steps.

Features

Currently supported:

  • Rust 1.11
  • x86 (32 and 64-bit), ARM (32 and 64-bit) build systems.
  • All Linux architectures that Rust itself supports (Multiple flavors of: x86, ARM, PPC, and MIPS)
  • Statically-linked libstd, dynamically-linked system libraries (libc, libm, etc)

Future:

  • Building and installing dev and staticdev packages (i.e. allow build and install of static and dynamic library builds).
  • Debug builds with separated debug info to allow gdbserver usage.
  • Running Rust/Cargo on target.
  • Vendoring of Cargo dependencies (to better play with the Yocto offline build model).
  • Use of a shared libstd across all Rust packages on a target system (provides space savings).
  • Total static linking using MUSL.

Advanced Features

Specifying Cargo Features

Because Yocto is primarily used for embedded development, it is likely that projects will have differing features based on whether the crate is run on the hardware or in development on a PC. Cargo features can be easily specified by adding a space-separated list of CARGO_FEATURES to the recipe:

CARGO_FEATURES = "feature1 feature2"

Using Components Individually

Although the cargo class is the easiest way to use this layer, the components it provides may also be used directly. To add the Rust compiler plus target and host standard libraries to the environment, depend on or install rust-bin. To manually install cargo depend on or install cargo-bin.

Note that while there is nothing explicitly preventing the installation of Rust on the target, it hasn't been tested and is likely not to work. Pull requests are welcome!

Pre-built vs. Compiled

This layer exists as a tradeoff against other options, e.g. the meta-rust project. Both exist to satisfy different requirements.

Because this layer uses the upstream compiled versions of Rust and Cargo, it will never be able to support architectures or options not supported by the Rust team itself.

Also, because this layer uses pre-built Rust standard libraries, it is possible that the standard libraries provided with this layer will be less efficient than code produced by a custom-compiled standard library.

However, using pre-built tools has advantages:

  • Updating the layer to a new version of Rust is as easy as updating checksums and file names, so new versions of Rust are available quickly.
  • In almost all modern systems, it is faster to download the binaries than it is to download source and build the Rust toolchain from scratch.
  • Compatability across multiple versions of Yocto is maximized since only basic, stable recipe features are used.
  • Trivial support for all architectures supported by upstream Rust.

Copyright

Copyright (c) 2016, the meta-rust-bin authors.

Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
http://www.apache.org/license/LICENSE-2.0> or the MIT license
<LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
option.  This file may not be copied, modified, or distributed
except according to those terms.

About

Yocto layer for installing Rust toolchain from pre-built binaries

License:Apache License 2.0


Languages

Language:BitBake 61.2%Language:PHP 38.8%