Refactor to move methods to where they belong

[emilsoman]

Changes :

1. Moved rbkit_tracer.c to rbkit_server.c because the file contains not only code for tracing, but all other C functions in Rbkit server.
2. Renamed Rbkit::Profiler to Rbkit::Server because the methods inside the class were responsible for starting/stopping the server and processing commands from the client.
3. Moved all server related methods from Rbkit module to Rbkit::Server class as instance methods. I think this move makes the code easier to reason about, and keeps the user's API (Rbkit module methods) minimal and less confusing. This will also contain all new commands inside the Server class which will help implement Implement support for enabling and disabling of live object creation events #106 in a cleaner way

[iffyuva]

this changeset looks good to me.

[iffyuva]

i have a question, not entirely related to this PR.
why does start_profiling return an instance of server? what is the use of returning an instance of server?

[emilsoman]

Previously, you could do things like Rbkit.status or Rbkit.send_objectspace_dump . These methods don't really make sense because there's no indication that a server has been started to send the status or object space dump. I think it makes more sense to put these methods on an instance of Rbkit::Server, so it would be:

server = Rbkit.start_profiling # or Rbkit.start_server
server.status
server.send_objectspace_dump
server.stop # or Rbkit.stop_server

I don't expect users to use these APIs since they'll be using the GUI to send messages that trigger these methods, still it provides a way to do it from code. We are using this a lot inside our tests.

[iffyuva]

okie, got it