Better support Hardhat tests in --parallel mode

[frangio]

In Parallel mode, each thread gets its own Hardhat Network instance, yet they all have the same chain id so they use the same manifest file. I don't think this will cause errors with the manifest file because due to being a development network we basically ignore the contents if they don't match expectations, but it will for sure cause unnecessary contention and lead to errors such as NomicFoundation/hardhat#3093.

This should probably be fixed along with #465, by detecting different instances and using a different manifest file for each.

[frangio]

Note: There is currently no built-in way to uniquely identify Hardhat Network instances. We could use process.pid as a workaround untill that exists.

[Ramarti]

+1

  2) Forta
       "before all" hook for "check deployment":
     Error: Lock file is already being held
      at /home/somedude/workspace/forta-token/node_modules/proper-lockfile/lib/lockfile.js:68:47
      at callback (node_modules/graceful-fs/polyfills.js:306:20)
      at FSReqCallback.oncomplete (node:fs:199:5)

[frangio]

There is now a way to implement this with the hardhat_metadata RPC method: NomicFoundation/hardhat#3382.

Haven't found docs but here's the type that will be returned by the RPC method:

https://github.com/NomicFoundation/hardhat/blob/5cecb9692ca8bfa4a57a592e3233b24336da6e1c/packages/hardhat-core/src/internal/core/jsonrpc/types/output/metadata.ts