Replace threadpool with Rust's Standard Library for Parallel Execution

[notJoon]

Description

In the existing run_parallel, the threadpool package was used to handle parallel tasks. However, this package was only used in this function, and it has been determined that the standard library of Rust can adequately address these needs. Therefore, modifications have been made to no longer use this library.

To replicate the functionality previously provided by threadpool, the following changes have been made:

1. For each GrammarConfiguration instance, std::thread::spawn is used to create a separate thread.
2. std::sync::mpsc::channel is used to collect the results of tasks executed in each thread.
3. The process waits until all threads have finished and then collects the results from each thread.

[pascalkuthe]

Std has a threadpool too. Spawning threads manually isn't ideal

[archseer]

Is there a threadpool in std? I wasn't able to find one. Either way, the reason I didn't implement this with threads directly in the first place is that with a pool we can limit concurrency to a fixed amount of threads. We don't want to build all 100+ grammars at once concurrently

[pascalkuthe]

Is there a threadpool in std? I wasn't able to find one. Either way, the reason I didn't implement this with threads directly in the first place is that with a pool we can limit concurrency to a fixed amount of threads. We don't want to build all 100+ grammars at once concurrently

i meant the scope thread api. You can create 1 scope/thread per core, each with a reference to a channel and then pull from that channel. That is what ingore does internally (its more fancy these days as they recently added workstealing but unitl recently it just used a simple channel). So the API is not quite equivalent but its pretty close (just a couple lok exta)