From 457d38b19b73ad061d9563e5b01d4afab8bed5bc Mon Sep 17 00:00:00 2001
From: Christian Zangl <laktak@cdak.net>
Date: Thu, 19 Nov 2020 11:24:15 +0100
Subject: [PATCH] option for workers

---
 README.md      |  7 +++++--
 chkbit/main.py | 30 +++++++++++++++++++++++++-----
 setup.py       |  2 +-
 3 files changed, 31 insertions(+), 8 deletions(-)

diff --git a/README.md b/README.md
index 0d9ee26..d6ba483 100644
--- a/README.md
+++ b/README.md
@@ -45,18 +45,19 @@ chkbit will
 Run `chkbit PATH` to verify only.
 
 ```
-usage: chkbit.py [-h] [-u] [-f] [-i] [-q] [-v] [PATH [PATH ...]]
+usage: chkbit.py [-h] [-u] [-f] [-i] [-w N] [-q] [-v] [PATH [PATH ...]]
 
 Checks the data integrity of your files. See https://github.com/laktak/chkbit-py
 
 positional arguments:
-  PATH
+  PATH                directories to check
 
 optional arguments:
   -h, --help          show this help message and exit
   -u, --update        update indices (without this chkbit will only verify files)
   -f, --force         force update of damaged items
   -i, --verify-index  verify files in the index only (will not report new files)
+  -w N, --workers N   number of workers to use, default=5
   -q, --quiet         quiet, don't show progress/information
   -v, --verbose       verbose output
 
@@ -71,6 +72,8 @@ Status codes:
   EXC: internal exception
 ```
 
+chkbit is set to use only 5 workers by default so it will not slow your system to a crawl. You can specify a higher number to make it a lot faster (requires about 128kB of memory per worker).
+
 ## Repair
 
 chkbit cannot repair damage, its job is simply to detect it.
diff --git a/chkbit/main.py b/chkbit/main.py
index 3b3a025..4cdc401 100644
--- a/chkbit/main.py
+++ b/chkbit/main.py
@@ -51,7 +51,10 @@ class Main:
             epilog=STATUS_CODES,
             formatter_class=argparse.RawDescriptionHelpFormatter,
         )
-        parser.add_argument("PATH", nargs="*")
+
+        parser.add_argument(
+            "paths", metavar="PATH", type=str, nargs="*", help="directories to check"
+        )
 
         parser.add_argument(
             "-u",
@@ -71,6 +74,16 @@ class Main:
             help="verify files in the index only (will not report new files)",
         )
 
+        parser.add_argument(
+            "-w",
+            "--workers",
+            metavar="N",
+            action="store",
+            type=int,
+            default=5,
+            help="number of workers to use, default=5",
+        )
+
         parser.add_argument(
             "-q",
             "--quiet",
@@ -84,7 +97,7 @@ class Main:
         self.args = parser.parse_args()
         self.verbose = self.args.verbose
         self.quiet = self.args.quiet
-        if not self.args.PATH:
+        if not self.args.paths:
             parser.print_help()
 
     def _res_worker(self):
@@ -98,15 +111,22 @@ class Main:
     def process(self):
 
         self.res_queue = queue.Queue()
+
+        # the todo queue is used to distribute the work
+        # to the index threads
         todo_queue = queue.Queue()
 
-        for path in self.args.PATH:
+        # put the initial paths into the queue
+        for path in self.args.paths:
             todo_queue.put(path)
 
+        # start indexing
         workers = [
-            IndexThread(idx, self.args, self.res_queue, todo_queue) for idx in range(5)
+            IndexThread(idx, self.args, self.res_queue, todo_queue)
+            for idx in range(self.args.workers)
         ]
 
+        # log the results from the workers
         res_worker = threading.Thread(target=self._res_worker)
         res_worker.daemon = True
         res_worker.start()
@@ -142,7 +162,7 @@ class Main:
 def main():
     try:
         m = Main()
-        if m.args.PATH:
+        if m.args.paths:
             m.process()
             m.print_result()
     except KeyboardInterrupt:
diff --git a/setup.py b/setup.py
index b08bc8b..f41117a 100644
--- a/setup.py
+++ b/setup.py
@@ -11,7 +11,7 @@ with open(os.path.join(os.path.dirname(__file__), "README.md"), encoding="utf-8"
 
 setup(
     name="chkbit",
-    version="2.1.0",
+    version="2.1.1",
     url="https://github.com/laktak/chkbit-py",
     author="Christian Zangl",
     author_email="laktak@cdak.net",