Allow olly to attach to an external process

[eutro]

This PR allows olly to attach to an external process by directory and PID. That is, it implements #41.

I understand that @tmcgilchrist mentioned here that he has already started working on an implementation, I hope I am not stepping on toes with this.

---

The primary non-obvious difference between this and the existing code in launch.ml is that Unix.wait cannot be used to determine if the process is still alive (you can only wait for children), and so Unix.kill pid 0 is used instead. According to the POSIX specification of kill:

If sig is 0 (the null signal), error checking is performed but no signal is actually sent. The null signal can be used to check the validity of pid.

and

The kill() function shall fail if:
[EINVAL]
The value of the sig argument is an invalid or unsupported signal number.
[EPERM]
The process does not have permission to send the signal to any receiving process.
[ESRCH]
No process or process group can be found corresponding to that specified by pid.

So

let alive () =
  try Unix.kill pid 0; true
  with Unix.Unix_error (Unix.ESRCH, _, _) -> false

should work portably on Unix.

Olly currently already doesn't work on Windows due to how Unix handles PIDs (badly). I leave figuring out what to do with that for a later PR.

---

The second major change is how to actually handle this at the command line. Ideally, as mentioned here, we could do something like olly gc-stats --attach pid, while still allowing olly gc-stats /path/to/program.

Cmdliner has no clean way to express this, so I just specify --attach as a normal option and fail (still with a command line parsing error) if none/both are specified.

I also made EXECUTABLE be a sequence of arguments rather than just one space-split one, this makes expressing the above slightly easier. It is also in-line with existing tools like perf record or gdb --args, and allows olly to pass arguments (and execute programs) containing spaces. This part is technically backwards-incompatible, if that's a problem.

[tmcgilchrist]

@eutro thank you for making this work. I've added minor changes to get CI working and added 5.2 builds.
Merging this work under #50