Refactor TargetPath and add support for Python 3.12

[Schamper]

This refactors TargetPath into a structure that's hopefully a little easier to maintain for future Python versions. TargetPath implementations are now split into different files, one for each Python version. Each file has specific changes needed to make it work. Each new release of Python still requires a new patch though.

I've also fixed some small filesystem and other test related bugs I ran into to get this all working.

As I'm writing this, I'm also starting to wonder at what point it might be a better idea to just... subclass pathlib.Path but literally override everything. I'm sure that has it's own set of challenges though.

Depends on fox-it/flow.record#91

[codecov]

Codecov Report

Attention: 148 lines in your changes are missing coverage. Please review.

Comparison is base (011323d) 74.18% compared to head (93d456e) 74.76%.

Files	Patch %	Lines
dissect/target/helpers/compat/path_39.py	88.38%	36 Missing ⚠️
dissect/target/helpers/compat/path_310.py	89.28%	30 Missing ⚠️
dissect/target/helpers/compat/path_311.py	88.41%	30 Missing ⚠️
dissect/target/helpers/compat/path_common.py	81.06%	25 Missing ⚠️
dissect/target/helpers/compat/path_312.py	89.75%	21 Missing ⚠️
dissect/target/filesystem.py	73.33%	4 Missing ⚠️
dissect/target/helpers/fsutil.py	92.85%	1 Missing ⚠️
dissect/target/helpers/polypath.py	97.22%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #494      +/-   ##
==========================================
+ Coverage   74.18%   74.76%   +0.58%     
==========================================
  Files         263      269       +6     
  Lines       21222    22064     +842     
==========================================
+ Hits        15744    16497     +753     
- Misses       5478     5567      +89     

Flag	Coverage Δ
unittests	74.76% <88.16%> (+0.58%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pyrco]

I noticed the 3.12 implementation still uses a "flavour" class insteasd of a module for TargetPath._flavour. I guess that is the easiest solution for now given dynamic separators?

I guess we should also tag&release flow.record and bump the dependency here to have the tests work.

[Schamper]

I noticed the 3.12 implementation still uses a "flavour" class insteasd of a module for TargetPath._flavour. I guess that is the easiest solution for now given dynamic separators?

Yeah, unfortunately because we use dynamic separators we need some degree of a "flavour" class. I was thinking about a different approach with static pre-created flavours/PurePath classes for each combination (similar to how there's a PosixPurePath and WindowsPurePath. That would also allow us to use __slots__ again. I'll give that a try later to see how well that works.

I guess we should also tag&release flow.record and bump the dependency here to have the tests work.

Yes, I'll bump the version in this PR already while we do that.