fix: fix perf bug in mox compile (#159)

* fix compile perf bug

fix a perf bug in mox compile. `compile_` was returning the deployer,
which pickles it to the process's stdout. however, this is not necessary
as the result is not used.

on my machine, this brings warm compile of 36 contracts from 5s to 2.2s.

* fix signature

* fix regression -- inspect module uses compile_

add a wrapper for use by multiprocessing which discards the result.

moccasin/commands/compile.py

@@ -115,7 +115,7 @@ def compile_project(
     with multiprocessing.Pool(n_cpus) as pool:
         for contract_path in contracts_to_compile:
             res = pool.apply_async(
-                compile_,
+                compile_noret,
                 (contract_path, build_folder),
                 dict(is_zksync=is_zksync, write_data=write_data),
             )
@@ -150,7 +150,7 @@ def compile_(
     compiler_args: dict | None = None,
     is_zksync: bool = False,
     write_data: bool = False,
-) -> VyperDeployer | None:
+) -> VyperDeployer | VVMDeployer:
     logger.debug(f"Compiling contract {contract_path}")
 
     # Getting the compiler Data
@@ -209,3 +209,8 @@ def compile_(
     logger.debug(f"Done compiling {contract_name}")
 
     return deployer
+
+# discard the result of the compilation so that we don't need to pickle it
+# between processes
+def compile_noret(*args, **kwargs) -> None:
+    compile_(*args, **kwargs)