Ensure all sub-processes are joined before killing them.

Without waiting for them to run their handlers, and killing them right
away, some may not have the "time" to write(2) their prints.

It typically happen on GH workflows.

Author: JulienPalard

sphinxlint.py

@@ -560,6 +560,8 @@ def main(argv=None):
     else:
         with multiprocessing.Pool() as pool:
             results = pool.map(configured_check_file, todo)
+            pool.close()
+            pool.join()
 
     count = reduce(Counter.__add__, results)