Optimize _compute_mro()

[jacobtylerwalls]

	Type
✓	🔨 Refactoring

Description

• Optimize clean_duplicates_mro()
• Short-circuit in _compute_mro() for "builtins.object"

Performance improvement is more noticeable in projects making heavy use of inheritance. For the project music21 from the pylint primer:

Benchmark

from pylint.lint import Run
Run(['music21'], exit=False)  # using project's own rcfile

main

452811674 function calls (383785563 primitive calls) in 161.566 seconds
ncalls  tottime  percall  cumtime  percall filename:lineno(function)
...
427785/97707    0.934    0.000   10.187    0.000 scoped_nodes.py:2808(_compute_mro)

PR

450362344 function calls (381157337 primitive calls) in 157.018 seconds
ncalls  tottime  percall  cumtime  percall filename:lineno(function)
...
427785/97707    0.692    0.000    8.160    0.000 scoped_nodes.py:2808(_compute_mro)

That's ~2s saved out of ~160s, or 1.25%.

[codecov]

Codecov Report

Merging #2187 (81a4d9b) into main (feafcb8) will increase coverage by 0.02%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2187      +/-   ##
==========================================
+ Coverage   92.61%   92.63%   +0.02%     
==========================================
  Files          94       94              
  Lines       10814    10818       +4     
==========================================
+ Hits        10015    10021       +6     
+ Misses        799      797       -2     

Flag	Coverage Δ
linux	92.39% <100.00%> (+0.02%)	⬆️
pypy	87.59% <100.00%> (+0.02%)	⬆️
windows	92.22% <100.00%> (+0.02%)	⬆️

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

Impacted Files	Coverage Δ
astroid/nodes/scoped_nodes/scoped_nodes.py	91.96% <100.00%> (+0.02%)	⬆️

... and 1 file with indirect coverage changes

[DanielNoord]

astroid/astroid/nodes/scoped_nodes/scoped_nodes.py

Line 2806 in feafcb8

yield from baseobj.bases

This makes this SuccesfulInferenceResult although I wonder if ClassDef.bases could be narrowed down. I'd be okay with making this ClassDef for now and fixing later if necessary, but note that this will add one more mypy warning after _inferred_bases gets its signature typed.

[jacobtylerwalls]

Yep. However, Pyright also says it could be ClassDef due to this:

astroid/astroid/nodes/scoped_nodes/scoped_nodes.py

Lines 2803 to 2804 in feafcb8

	if not baseobj.hide:
	yield baseobj

Let's wait until we're ready to type this all together.

[DanielNoord]

Yeah of course, the bases line just broadens the type (perhaps/probably incorrectly)