Use asyncio.create_subprocess instead of subprocess.run

[marczalik]

One sentence summary of this PR (This should go in the CHANGELOG!)
Use asyncio.create_subprocess_execinstead of blocking subprocess.run calls in packers/unpackers.

Link to Related Issue(s)
#53

Please describe the changes in your request.
See above

Anyone you think should look at this, specifically?
@rbs-jacob

[rbs-jacob]

I made a few comments, but the comments generally need to be applied across all of the places that were switched from subprocess to asyncio.

[rbs-jacob]

Here are some additional places that were missed that should probably also be converted:

jacob@jacob ofrak % rg -i subprocess | rg -v 'asyncio|test|patch_maker|build_image|ipynb|\.md|ofrak_ghidra'
./ofrak_core/ofrak/model/component_model.py:import subprocess
./ofrak_core/ofrak/model/component_model.py:            retcode = subprocess.call(
./ofrak_core/ofrak/model/component_model.py:                stdout=subprocess.DEVNULL,
./ofrak_core/ofrak/model/component_model.py:                stderr=subprocess.DEVNULL,
./ofrak_core/ofrak/component/abstract.py:from subprocess import CalledProcessError
./ofrak_core/ofrak/component/abstract.py:            raise ComponentSubprocessError(e)
./ofrak_core/ofrak/component/abstract.py:class ComponentSubprocessError(RuntimeError):
./ofrak_core/ofrak/core/apk.py:import subprocess
./ofrak_core/ofrak/core/apk.py:            retcode = subprocess.call(
./ofrak_core/ofrak/core/apk.py:                stdout=subprocess.DEVNULL,
./ofrak_core/ofrak/core/apk.py:                stderr=subprocess.DEVNULL,
./ofrak_core/ofrak/core/tar.py:import subprocess
./ofrak_core/ofrak/core/tar.py:                subprocess.run(command, check=True, capture_output=True)
./ofrak_core/ofrak/core/squashfs.py:import subprocess
./ofrak_core/ofrak/core/squashfs.py:            result = subprocess.run(
./ofrak_core/ofrak/core/squashfs.py:                stdout=subprocess.PIPE,
./ofrak_core/ofrak/core/squashfs.py:                stderr=subprocess.DEVNULL,

[rbs-jacob]

One minor outstanding change

[whyitfor]

@marczalik, here (and in other places in the PR) its seems that there are functional changes.

Before, it would return 0 == retcode. After this PR, any tool that does not have a return code of 0 raises a CalledProcessError. Was this change in functionality intentional?

[whyitfor]

I'm asking because this changes the behavior of ComponentExternalTool, making it less usable.

[marczalik]

@whyitfor In general, raising CalledProcessError for the majority of these cases is consistent with previous behavior, where the check=True in subprocess.run() does the same thing. (Here are the docs for reference)

However, you're correct in that this specific case , component_model.py, and 'squashfs.py` line 30 I should not have added these. I will create a bugfix for this.