Speed up `glob.glob()` by reducing number of system calls made

[barneygale]

It should be possible to speed up glob.glob() using a few strategies developed in pathlib.

Linked PRs

• GH-116380: Speed up glob.glob() by removing some system calls #116392
• GH-116380: Add tests for glob.glob0() and glob.glob1(). #117305
• GH-116380: Support dir_fd argument in glob._Globber. #117737
• GH-116380: Support include_hidden argument in glob._Globber. #117738
• GH-116380: Move pathlib globbing implementation into pathlib._glob #118562
• GH-116380: Revert move of pathlib globbing code to pathlib._glob #118678
• GH-116380: Move pathlib-specific code from glob to pathlib._abc. #120011
• GH-116380: Revert addition of test_iglob_iter_close test #130710
• Revert "GH-116380: Speed up glob.[i]glob() by making fewer system calls. (#116392)" #130743
• GH-116380: Speed up glob.[i]glob() by making fewer system calls (take 2) #137474