On Wednesday, 8 April 2020 17:21:44 CEST 刘元哲 wrote: > Re-Update: > After compared the different result > between ssh_message_global_request_reply_success > before ssh_channel_open_reverse_forward or after it I > found ssh_channel_listen_forward want a reply, so if I reply it > immediately, the ssh client then said "ssh_global_request: Global request > tcpip-forward success"and can't > receive ssh_channel_open_reverse_forward request. Otherwise if I > choose to reply it after ssh_channel_open_reverse_forward, > the ssh_channel_listen_forward will hang there, continue wait > for the reply.
I've never used that code, we don't have example code nor a test for it. So as always I would argue: Untested code is broken code. I think we need to implement a test to actually fix the code. However I don't have time for this. Contributions in this area are very welcome. See the server testing code Anderson implemented for a start. Best regards, Andreas -- Andreas Schneider a...@cryptomilk.org GPG-ID: 8DFF53E18F2ABC8D8F3C92237EE0FC4DCC014E3D