Sort out various issues with the chip standalone server code

[bzbarsky-apple]

Problem

I was looking over #2042 and I had a few questions I could not find answers to:

1. Why was chip-tool the right place to place the standalone server? It does not seem to share any code with the existing chip-tool command-line client...
2. Why was various code in src/app commented out instead of making sure that the new chip server is linked with -ffunction-sections, which would have presumably fixed the link issues encountered?
3. Where did the generated code for the new server come from? It's missing some modifications (e.g. replacing boolean with bool in the signature of halSleepCallback) that are needed to make the generated silicon labs code compile.

Proposed Solution

Sort out the above.

@erjiaqing @andy31415 @woody-apple @jelderton

[issue-label-bot]

Issue Label Bot is not confident enough to auto-label this issue. See dashboard for more details.

[bzbarsky-apple]

@bhaskar-apple @rwalker-apple @sagar-apple

[bzbarsky-apple]

Why was chip-tool the right place to place the standalone server? It does not seem to share any code with the existing chip-tool command-line client...

And in particular, I believe the medium-term plan is to include generated client code for various clusters in chip-tool (e.g. onoff client), which is not the same generated code that the standalone server wants, presumably. So at that point they will be fighting over the gen directory or we'll have two separate gen directories or something...

[bzbarsky-apple]

This code has been moved, looks like, in #2365