messy first version of a properly working depth-first graph method, which allows the pool to work as expected. Many more tests need to be added, and there still is a problem with shutdown as sometimes it won't kill all threads, mainly because the process came up with worker threads started, which cannot be

1 files changed, 7 insertions, 1 deletions

diff --git a/lib/git/async/util.py b/lib/git/async/util.py
index 00d0dbab..b7750b0b 100644
--- a/lib/git/async/util.py
+++ b/lib/git/async/util.py
@@ -206,7 +206,6 @@ class AsyncQueue(deque):
 			return old
 		finally:
 			self.mutex.release()
-			
 			# if we won't receive anymore items, inform the getters
 			if not state:
 				self.not_empty.notify_all()
@@ -222,6 +221,13 @@ class AsyncQueue(deque):
 
 	def put(self, item, block=True, timeout=None):
 		self.mutex.acquire()
+		# NOTE: we explicitly do NOT check for our writable state
+		# Its just used as a notification signal, and we need to be able 
+		# to continue writing to prevent threads ( easily ) from failing
+		# to write their computed results, which we want in fact
+		# NO: we want them to fail and stop processing, as the one who caused
+		# the channel to close had a reason and wants the threads to 
+		# stop on the task as soon as possible
 		if not self._writable:
 			self.mutex.release()
 			raise ReadOnly