[LanguageWorkers]Implement function reload

We are too aggressive about restarting the language worker processes (for example, on re-loading user code). Can lead to confusing log messages about `node exited with code -1`

-----
First reported: https://github.com/Azure/azure-functions-core-tools/issues/689
_Per @blueelvis:_
Hi!

Host Version - `2.0.12050.0`
Running I am starting the function and after starting, I get the following in the logs -

```
[02-09-2018 13:43:04] Generating 4 job function(s)
[02-09-2018 13:43:04] Found the following functions:
[02-09-2018 13:43:04] Host.Functions.damage-assessment-bot
[02-09-2018 13:43:04] Host.Functions.messages
[02-09-2018 13:43:04] Host.Functions.some-function
[02-09-2018 13:43:04] Host.Functions.test
[02-09-2018 13:43:04]
[02-09-2018 13:43:04] Host initialized (86ms)
[02-09-2018 13:43:04] Host started (88ms)
[02-09-2018 13:43:04] Job host started
[02-09-2018 13:43:04] Stopping JobHost
[02-09-2018 13:43:04] Job host stopped
info: Host.General[0]
Script host restarted.
[02-09-2018 13:43:04] Language Worker Process exited.
[02-09-2018 13:43:04] node exited with code -1
.
[02-09-2018 13:43:04] Worker 6cdf2495-5568-40ee-ac26-a973a2423e99 connecting on 127.0.0.1:22499 []
[02-09-2018 13:43:05] Path must be a string. Received undefined
[02-09-2018 13:43:09] Host lock lease acquired by instance ID '000000000000000000000000571B63F6'.
```

The Function starts working fine but any idea why am I getting that the job host was stopped and node exited with code -1?

Start by tracing the LanguageWorkers behavior during user-code reloads and compare it with the restart and log sequence described in the issue and linked report. Done means function reload no longer unnecessarily restarts language workers or produces misleading process-exit messages.

Written by the indexing model from the issue text.