Hi all,
Quick question regarding the Socket PDU block... When I'm using the Socket PDU in "TCP Client" mode, I see the following behavior on the maint-3.7 branch:
Quick question regarding the Socket PDU block... When I'm using the Socket PDU in "TCP Client" mode, I see the following behavior on the maint-3.7 branch:
1) On startup, if a server is not already ready on the desired port, I immediately see a "RuntimeError: Connection Refused" and the flowgraph dies.
2) If a server is available, then disconnects, I see a "terminate called after throwing an instance of boost::system::system_error' what(): End of file Aborted (core dumped)", and the flowgraph dies.
As far as I can tell, this is intended operation of the existing Socket PDU block. Correct?
In an ideal world, my application needs to create a TCP Client that can wait until the server is alive (dropping PDUs if there's no connection), then acquire and reconnect to the server later. I dont have an option to use something like ZMQ; I need to interact with a TCP socket server.
Seems to me like this would be a fairly common issue. Does anyone have any ideas if there's a block, maybe in an OOT repo, that does this already? If I make the corresponding changes to the in-tree Socket PDU, would this be a feature that makes sense to upstream?
Thanks,
EJ
In an ideal world, my application needs to create a TCP Client that can wait until the server is alive (dropping PDUs if there's no connection), then acquire and reconnect to the server later. I dont have an option to use something like ZMQ; I need to interact with a TCP socket server.
Seems to me like this would be a fairly common issue. Does anyone have any ideas if there's a block, maybe in an OOT repo, that does this already? If I make the corresponding changes to the in-tree Socket PDU, would this be a feature that makes sense to upstream?
Thanks,
EJ
No comments:
Post a Comment