When using a [custom] RPC client, and upon that client crashing (that is, ungracefully dropping a connection), I am somewhat consistently having bitcoind SEGV.
I'll see if I can't create a simple test case...
edit: not as consistently as I'd hoped
Will re-open if I can reproduce consistently, I can't seem to re-create it suddenly despite having what felt like I had a minimal test case... Debugging woes :+1:
We'd at least need a stack trace from gdb (for the http threads) to troubleshoot this.