Robert Bragg

Okey, trying to look in to this to see if we can get this landed... I've added it to the 0.21 milestone too to hopefully not lose track of resolving...

Just for reference; smoke testing this on Windows then `cargo test` did actually Just work for me even without adding my jdk's `bin/` directory to the `PATH` explicitly (which I...

Okey, I've rebased this work and then made two follow up changes: 1. I just kept the new dynamic loading in favor of the previous build-time linking. (including some corresponding...

> This issue was tricky to test, and often showed for users as a JVM crash that pointed to failing to load the garbage collector, as on their path/java_home they...

I started merging PRs with API changes for 0.21 today and am hoping to land this PR too but I'll let it sit for a few days in case @argv-minus-one...

> You might be right with documentation of libloading. > > I do think the issue is something for `jni-rs` to handle, rather than leaving it to the user, as...

Logging on Android can be handled using a crate such as [android_logger](https://crates.io/crates/android_logger) E.g. add this to your deps: `android_logger = "0.11.0"` and initialize something like: ```rust android_logger::init_once(android_logger::Config::default().with_min_level(log::Level::Info)); ``` You don't...

thanks for the offer here @jrose-signal I'll try and get a chance to look at this more closely - it certainly looks like it could be a good option. I...

Since it relates to the `call_*_method` interface; something else that I wonder here is whether it could be nicer to have separate call APIs according to the return type (in...

I suppose my initial instinct is that: - I quite like that the macros @jrose-signal has are relatively simple (although more fancy things would be possible with a proc macro...