Support native Python3 in Ghidra with Pyhidra

[mahaloz]

Using Pyhidra, we should be able to run 99% of all libbs code inside Ghidra natively, which will give us a HUGE speed improvement.

Tasks

• [x] Update the auto-detection and make the Ghidra backend configurable: https://github.com/binsync/libbs/blob/9f97a51e3f84955bd45f7b5929d4a2dad1760e5c/libbs/api/decompiler_interface.py#L657
• [x] Rewrite all uses of bridge outside of the API wrapper: https://github.com/binsync/libbs/blob/9f97a51e3f84955bd45f7b5929d4a2dad1760e5c/libbs/decompilers/ghidra/compat/ghidra_api.py#L15 We must now wrap any type of remote eval.
• [x] Update headless running to match Pyhdria style
• [x] Test BinSync on Pyhidra
• [x] Find a way to deal with remote eval code like: https://github.com/binsync/libbs/blob/9f97a51e3f84955bd45f7b5929d4a2dad1760e5c/libbs/decompilers/ghidra/interface.py#L463 Maybe we can just make these special cases?

[chf0x]

As part of this refactoring, please add the capability to load an existing project without re-running the ghidra analysis on it (in headless mode). Since we have many projects cached, this would save a significant amount of time

[mahaloz]

@chf0x would you happen to know if there is a way to start a normal Ghidra headless instance without re-running analysis? If not, it's fine I'll find it eventually :).

[chf0x]

You mean with pyhidra, right? You can just set analyze=False.

with pyhidra.open_program("file.exe", analyze=False) as flat_api:
    do_stuff

To open a project, specify the project_name and project_location, and binary_path = None

https://github.com/dod-cyber-crime-center/pyhidra/blob/c878e91b53498f65f2eb0255e22189a6d172917c/pyhidra/core.py#L178