Fix to order of module discovery

[akamat10]

Type of Changes

	Type
✓	🐛 Bug fix

Description

• Addresses the issues with module path discovery identified on Inconsistent and sometimes incorrect false positive of no-member error pylint#9883
• Updated tests to reflect current python behavior when calling a script. Details of the expected behavior described here

[akamat10]

One thing this fix needs to consider is how the order of module discovery changes based on whether a user calls:

python3.12 pkg/script.py

vs

python3.12 -m pkg.script

The difference in behavior is described here in a stackoverflow post.

The test fixes to test_nonregr_absolute_import case will now match the former case but perhaps some thought needs to be given to the latter case and have some test cases around it.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.02%. Comparing base (a389ef7) to head (04fbce5).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2531      +/-   ##
==========================================
+ Coverage   92.98%   93.02%   +0.03%     
==========================================
  Files          93       93              
  Lines       11049    11052       +3     
==========================================
+ Hits        10274    10281       +7     
+ Misses        775      771       -4     

Flag	Coverage Δ
linux	92.89% <100.00%> (+0.03%)	⬆️
pypy	93.02% <100.00%> (+0.03%)	⬆️
windows	93.00% <100.00%> (+0.03%)	⬆️

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

Files with missing lines	Coverage Δ
astroid/nodes/_base_nodes.py	94.31% <100.00%> (+0.06%)	⬆️
astroid/nodes/scoped_nodes/scoped_nodes.py	92.79% <100.00%> (ø)

... and 2 files with indirect coverage changes

[Pierre-Sassoulas]

Sorry for not reviewing this sooner, this is the kind of change that are hard to assess/test and it's frightening 😅

[akamat10]

No need to be sorry. I should reevaluate this given it has been a while since I have looked at this one.