Not possible to restart internal go server #107
Description
We've seen an issue where for some reason the internal go server started by hrequests.cffi stops unexpectedly. We subsequently see Connection refused errors for every request.
As a workaround, we intended to try to stop and then relaunch the go server, however this is not currently possible as we can't re-register the http handlers:
panic: http: multiple registrations for /request
goroutine 20 [running]:
[github.com/bogdanfinn/fhttp.(*ServeMux).Handle(0x110cd15c0](http://github.com/bogdanfinn/fhttp.(*ServeMux).Handle(0x110cd15c0), {0x11081e09e, 0x8}, {0x1109ba948, 0x1109b77a0})
/go/pkg/mod/github.com/bogdanfinn/fhttp@v0.5.29/server.go:2473 +0x1f0
[github.com/bogdanfinn/fhttp.(*ServeMux).HandleFunc(...)](http://github.com/bogdanfinn/fhttp.(*ServeMux).HandleFunc(...))
/go/pkg/mod/github.com/bogdanfinn/fhttp@v0.5.29/server.go:2510
[github.com/bogdanfinn/fhttp.HandleFunc(...)](http://github.com/bogdanfinn/fhttp.HandleFunc(...))
/go/pkg/mod/github.com/bogdanfinn/fhttp@v0.5.29/server.go:2522
main.startServer({0x116e9aa18, 0x5})
/source/server.go:203 +0xb8
created by main.StartServer in goroutine 17
/source/server.go:216 +0x68
ideally there would be
- some internal mechanism to restart the server on failure
- logs which are propagated to python for starting/stopping the server
- a way in userland to restart the server without errors