Skip to content

VirtusLab/libsodiumjni

 
 

Repository files navigation

libsodiumjni

Minimal JNI bindings for libsodium

This project provides JNI bindings for some libsodium methods.

It's used by Scala CLI to encode repository secrets about to be uploaded to GitHub via the GitHub API. For now, it only exposes libsodium methods that Scala CLI needs.

The motivation for writing libsodiumjni, rather than using any of the existing JNA-based bindings, is mainly using JNI rather than JNA to interface with libsodium, as JNI is supported from GraalVM native images, while JNA is not.

Build

libsodiumjni is built with Mill.

Compile everything with

$ ./mill __.compile

The Mill build automatically compile the C bindings, and generates a .so / .dylib / .dll. These files are also built on the CI, and pushed to Maven Central (see below).

Artifacts

libsodiumjni is pushed to Maven Central. Add it in a Mill module with

def ivyDeps = super.ivyDeps() ++ Seq(
  ivy"io.github.alexarchambault.tmp.libsodiumjni:libsodiumjni:0.0.2"
)

Both static and dynamic libraries are also pushed as is to Maven Central (see the .so / .dylib / .dll and .a / .lib files here for example). From the command line, you can fetch those files with commands such as

$ cs fetch --intransitive io.github.alexarchambault.tmp.libsodiumjni:libsodiumjni:0.0.2,classifier=x86_64-pc-win32,ext=dll,type=dll -A dll

Adding support for new libsodium new methods

Typical workflow:

  • add a native method in SodiumApi.java
  • run ./mill __.compile, this should add a corresponding declaration in libsodiumjni_internal_SodiumApi.h
  • add an implementation for that declaration in libsodiumjni_internal_SodiumApi.c in the same directory
  • run ./mill __.compile to ensure that the new C code compiles fine
  • use the newly added native method in Sodium.java say
  • optionally, add tests for it in SodiumTests.java
  • ensure those tests pass when running ./mill __.test

About

Minimal JNI bindings for libsodium

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Scala 62.7%
  • Java 12.4%
  • Shell 8.5%
  • Batchfile 8.2%
  • C 8.2%