Closed paddor closed 4 months ago
Do you have a code example in C that reproduces this issue?
Sorry, I don't.
Ok, if you don't mind then I'm closing this. We can't substract relevant info from this info. I'm also not sure why zsock_events is called that often. That's not the way to do it IMHO.
zsock_events is called often because it's in an event loop that checks readability using the FD, which is unreliable. After being notified about readability/writability from the FD, you still have to check the socket's events to be sure.
It's okay, it's not an urgent problem / not affecting me these days. If it returns, I'll try to come up with a code example in C.
I'm using ZMQ v4.3.4 and CZMQ v4.2.1 with the CZTop binding (FFI classes are generated in this project). While testing a ping functionality in my app, I ran the ping in a tight loop and noticed that CZMQ sometimes segfaults in
zsock_events()
after a few seconds (a few thousand pings).Relevant stacktraces from Ruby: