Improve the performance of `kompile`

[Stevengre]

kompile is quite slow, and it doesn’t seem to speed up even increased memory using -Xmx. And I found no way for concurrent kompilation. This has made debugging a bug very time-consuming.

[ehildenb]

@Stevengre do you have examples it's slow on? We've made a lot of performance improvements in Kompile, and it also is able to do parallel compilations of parts of the pipeline, but without an example we can't optimize your specific use case. If it's a private codebase, maybe we could get access out of band and see what is making it slow with a profiler.

[Stevengre]

It's currently a private repo, but it will be open-sourced in the future. I'm in the process of refactoring right now, and it seems like Kompile is running much faster. I suspect there might have been issues with the previous syntax definitions and rule implementations. If the problem persists after the refactor, we can explore further optimizations.