dali VS go-plugin

Tangentially related, I've written a barebones assembler for Android .apk files once (strictly speaking, the assembler is for .dex files, but it also comes with a set of tools to package and sign .apk files). It's written mainly in Nim and provides enough primitives to allow creating Java "stubs" for native .so libraries, so that .apk-s can be built in Nim WITHOUT JDK AT ALL. The Android NDK is still kinda needed/useful, though IIRC mainly for access to adb, and especially adb logcat (which you'll need A LOT for debugging if you try to use this contraption).

I'd love to One Day™ Rewrite It In Rust.

The .dex assembler itself is at: https://github.com/akavel/dali — you may like to check out the tests at: https://github.com/akavel/dali/tree/master/tests to see how using it looks like.

An example project with a simple .apk written purely in Nim (NO JDK) is at: https://github.com/akavel/hellomello/tree/flappy (unfortunately, given Nim's poor packaging story, it's most probably already bitrotten to the extent that it can't be quickly and easily built & used out of the box). I recorded a presentation about this for an online Nim conference — see: https://www.youtube.com/watch?v=wr9X5NCwPlI&list=PLxLdEZg8DR...

https://github.com/akavel/dali was one (a fully hand-written assembler for Android .apk files); I managed to write a rudimentary flappy-bird-like prototype in it and did a presentation: https://www.youtube.com/watch?v=wr9X5NCwPlI&list=PLxLdEZg8DR... but on shelf now, didn't get much attention, and I don't feel bad about it. Had some roadblocks, but managed to overcome them, and I'm honestly surprised how the core effort was basically easy to implement and how the formats were open and relatively simple. (The main real issues I had were that debugging via adb logs was tiresome when something was not working.) What was funny about this project was that I started it with basically a thought of: "there will be probably some annoying roadblock at some point that will make it unviable to continue; I accept that and will be ok with stopping once I stumble upon it; but I don't see one clearly from the start [I did some quick initial research how the formats & the bytecode look and they seemed rather simple], and I'm really curious how far I can get if I decide to not think about this possible roadblock". Turns out I was able to get all the way to the end :D

I am looking forward to a mix of both - I am hoping to add a concept called "operator" which would be a go-plugin [1], just like terraform providers, but build backends. So, someone would be able to, say, write a Slack plugin (in Go, or anything over RPC) which sends a message once a build is complete - like Jenkins/GitHub actions, or just scripts that we can reuse like GitLab CI through `modules`.

Perhaps a new registry where we can push custom modules and providers (operators in this case), I'm curious to know about if we have any existing implementations we could reuse for the registry.

[1]: https://github.com/hashicorp/go-plugin

I'd recommend looking into a different approach for plugins such as hashicorp/go-plugin (which uses multiple process PIDs and RPC communication between them) or traefik/yaegi (which implements a Go-compatible scripting language that can be interpreted at runtime and which still supports most Go modules).

You'll never fully protect your code from someone who's dead-set on reverse-engineering it, however, you can use https://github.com/hashicorp/go-plugin or a similar RPC technique, which will let you ship binary plugins and will also be less fragile and janky compared to something made with `-buildmode=plugin`.

The plugin system that hashicorp uses for all their projects works very well. It's essentially a local RPC implementation. https://github.com/hashicorp/go-plugin

can you try older go-plugin versions? The only major change in v1.4.4 was a bugfix for automtls. https://github.com/hashicorp/go-plugin/blob/master/CHANGELOG.md