As we prepare for alpha release of JS plugins, it feels like we can remove the warning logged to stderr.
```
WARNING: JS plugins are experimental and not subject to semver.
Breaking changes are possible while JS plugins support is under development.
```
They do still remain experimental, but I don't think we need to shout that quite as loudly now.
It'd be better if the path of the specific invalid config were printed when this happened, in case I'm using nested configs, but for now this improves things a bit, at least.
A couple of changes to how plugin names are dealt with:
### Package names
Don't use package name if the specifier used to load the plugin is a relative path.
If your project is called `super-server`, and you load a plugin from `../plugins/no-foo.js`, that plugin shouldn't be called `super-server` - that's the name of the project, not the plugin.
Now it throws an error, telling you to give the plugin a name.
### Ensure `context.id` matches the plugin name
Previously `context.id` might not match the actual plugin name (it could include e.g. `eslint-plugin-` prefix, for example). Now it always matches.
This required re-implementing `normalize_plugin_name` on JS side too.
### Tests
Add a bunch of tests for every combination I could imagine of `plugin.meta.name`, package name, and alias. Those tests make sure `context.id` is same as the plugin name on Rust side.
### Incomplete
There's 2 things remaining to get naming completely correct:
1. Support plugin names with slashes in them e.g. `@scope/eslint-plugin-name` (#14557).
2. It's possible for the same plugin to have different names in different places, because you can alias the same plugin twice with different names in a config, and an override. Plugin names (in `context.id`, and in `// oxlint-disable` comments) should reflect the plugin name *as it is in the config which enabled that rule*.
I'm leaving those 2 for now.
overlookmotel