handle_close / ReceiveQueue race condition

[PeterSurda]

Came in bitmessage chan:

There is a race condition between the Asyncore thread and a ReceiveQueue thread where the Asyncore thread calls AdvancedDispatcher.handle_close (setting read_buf to an empty bytearray) between a ReceiveQueue thread ensuring the appropriate number of bytes in read_buf and calling the appropriate state_* method.

---

Hmm, I thought this was fixed. Let me check.

Peter

Hi all,

Off the top of your head, any suggestions about the error below? Is it normal?

Bitmessage v.0.6 branch, Tor proxy, log level DEBUG.

May 03 09:56:31 Asyncore tcp.py@57 Outbound proxy connection to 85.114.135.102:8444 May 03 09:56:31 ReceiveQueue_0 receivequeuethread.py@60 Error processing Traceback (most recent call last): File "/opt/PyBitmessage/src/network/receivequeuethread.py", line 50, in run connection.process() File "/opt/PyBitmessage/src/network/advanceddispatcher.py", line 72, in process if not cmd(): File "/opt/PyBitmessage/src/network/socks5.py", line 42, in state_auth_1 ret = struct.unpack('BB', self.read_buf[:2]) error: unpack requires a string argument of length 2