Allow all commands with the `on` breakpoint hook

[sivaplaysmC]

Allowing any command to be executed on a brekpoint hit, will allow for better usage. One example i can think of is having a empty function called _debugBreak(), and whenever _debugBreak() is hit, it should automatically execute the 'next' command - effectively stopping right after _debugBreak.

I also do not see any reason as to why only a limited set of commands can be executed when a breakpoint is hit.

[Lslightly]

allow python script might be good. https://lldb.llvm.org/use/python.html

[aarzilli]

This is effectively implemented by #4170 (although the very specific use case described here is better accomplished using runtime.Breakpoint).