Skip to content

RelationalAI/object_store_ffi

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

object_store_ffi

A Rust library that defines an (unofficial) C API for the object_store crate (docs.rs).

Note

This library is not a part of object_store or the Apache Arrow project.

To build it you need to have rust and cargo installed, and then run cargo build.

Pre-built binaries are available via the Julia package object_store_ffi_jll.jl. The primary user of object_store_ffi is the Julia package RustyObjectStore.jl.

Releasing

New releases of object_store_ffi are made by creating and pushing a tag with the new version (format: v[0-9]+.[0-9]+.[0-9]+). This will dispatch a release workflow that will build the assets and create the GitHub Release.

git tag -a v0.3.0 -m "New Release"
git push --tags

Each release should increment the version number following Semantic Versioning and should be kept in sync with the version field of Cargo.toml.

Whenever a new release of object_store_ffi is made, we should build binaries and release a new version of object_store_ffi_jll.jl. This is done by making a pull request to Yggdrasil to update the object_store_ffi build recipe. The PR should usually just update the GitSource commit hash to point to the new release, and to update the version number. Merging such a PR will automatically build the binaries for the supported platforms and publish them to the object_store_ffi_jll.jl repository as part of a new release of object_store_ffi_jll.jl.

Local build with BinaryBuilder.jl

If you want to try building object_store_ffi_jll.jl locally, this can be done using BinaryBuilder.jl. You will need a local copy of the object_store_ffi build recipe build_tarballs.jl, and a Julia environment with BinaryBuilder.jl installed.

For example, clone Yggdrasil and navigate to the object_store_ffi directory:

$ git clone git@github.com:JuliaPackaging/Yggdrasil.git
$ cd Yggdrasil/O/object_store_ffi/

Create a Julia environment with BinaryBuilder installed:

$ julia --project=build_env -e 'using Pkg; Pkg.add("BinaryBuilder")'

You can then run the build_tarballs.jl script:

$ julia --project=build_env build_tarballs.jl aarch64-apple-darwin --deploy=local --verbose --debug

See build_tarballs.jl --help for supported arguments.

See the BinaryBuilder documentation for more details.

Note BinaryBuilder requires Docker if running on MacOS or Windows.