Add middleware loader

[Schamper]

Consider this a proposal/experiment for a middleware loader, which should facilitate making loaders like #1456 and #1214 easier. I've adjusted the VBK loader as test (which also immediately serves as a unit test).

@qmadev I remember you looked a bit at VBK too, besides the middleware loader itself, does the VBK loader change make sense to you?

[codecov]

Codecov Report

❌ Patch coverage is 89.28571% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 81.09%. Comparing base (54808a5) to head (8e48c16).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
dissect/target/loader.py	86.66%	2 Missing ⚠️
dissect/target/loaders/vbk.py	92.30%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1589      +/-   ##
==========================================
+ Coverage   81.02%   81.09%   +0.06%     
==========================================
  Files         400      401       +1     
  Lines       34965    34988      +23     
==========================================
+ Hits        28332    28372      +40     
+ Misses       6633     6616      -17     

Flag	Coverage Δ
unittests	81.09% <89.28%> (+0.06%)	⬆️

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.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[codspeed-hq]

Merging this PR will not alter performance

✅ 12 untouched benchmarks

---

_{Comparing middleware-loader (8e48c16) with main (0e9dfb4)}

[qmadev]

This is the only change that does not make sense to me right now. Why are we not iterating candidates and loading the first "valid" one anymore?

[qmadev]

It's been a while since I looked at VBK and I don't have access to any data anymore, so I might just be forgetting something.

[Schamper]

This current setup doesn't really allow it because we can only return a single path, and I can't really think of a nicer way to allow that way, except for changing prepare to be an iterator instead of simply returning a path. I can maybe imagine a few more usecases where that could be useful, so maybe it's worth changing to that. Perhaps the reviewer can weigh in too.

But if I understand the logic in the VBK loader correctly, the candidates list should contain paths only if the globs for either vmx or vmcx succeeded, and otherwise the multiloader path. I don't expect there to be a valid way to create a VBK that contains both a vmx and a vmcx file.

[qmadev]

Allright, makes sense to me.

[qmadev]

Other than that, it looks good to me. Seems like this would make sure you never have to look for a new loader within a loader anymore.