Re-run of the package of version 1 (release/source.zip, sha256 b1f66619ca8333539268cb501a97bef4493676c94784524a13cf89963450fe89),
extracted into a fresh scratch folder; Python 3.13.5, one process each, nice -n 10; outputs compared with the recorded files of the archive.

check_note.py                                         identical (2667 checks, 0 failures, ALL CHECKS PASSED)
first_version/check_A5_witness.py                     identical
first_version/check_psl2_witness.py                   identical
first_version/check_zsigmondy_choices.py              identical
stage2_run_B: witness_psl2.py 4 5 7 8 9 11 13 16      identical up to running times (out/w_small.txt)
stage2_run_B: witness_psl2.py 17 19 23 25 27 29 31 32 identical up to running times (out/w_mid.txt)
stage2_run_B: obstruction_psl2.py 5 7 9 11 13         identical up to running times (out/obs_small.txt)
stage2_run_B: obstruction_psl2.py 17 19 23 25 27 29 31 identical up to running times (out/obs_mid.txt)
stage2_run_B: check_suzuki8.py                        identical up to running times (out/sz8.txt)
stage2_run_B: negative_controls.py                    identical (out/negative_controls.txt)
stage2_run_B: test_toolkit.py                         prints "toolkit tests passed"
stage2_run_B: heights_groups.py (the 11 groups of hg1) identical up to running times (out/hg1.txt)
stage2_run_B: zsig_arith.py (sympy)                   identical in the three lines it prints; the recorded out/zsig_arith.txt has one more
                                                      line (has_ppd(2,6)= False ...), which the program does not print: it comes from a
                                                      separate call of the function has_ppd (to be documented in independent_runs/README.md)
stage2_run_A: my_obstruction_check.py                 identical up to the final line "exit=0" appended by the run
stage2_run_A: my_zsigmondy_check.py (sympy)           identical
stage2_run_A: rerun_psl2_output.txt, rerun_zsig_output.txt agree with the first-version outputs
stage1_run_B: scripts/run_witness.py A5 A5_wrong_inner identical up to running times (out/witness_A5.txt)
stage1_run_B: scripts/run_tests.py S5 SL25 GL25 2^4:A5perm ASL24 2^4:S5perm A5xS4 ASL25 A5xA5 A5wrC2 A5A5diag2
                                                      identical up to running times and the class representative attaining a maximum
                                                      (hash ordering); all values agree (out/tests_small_medium.txt)
Not re-run: the slow programs (witness_psl2.py for 37..128, obstruction_psl2.py 41..73, heights_groups.py hg2/hg3,
wreath_tests.py, stage1_run_B break_*.py, run_extra.py, stage2_run_A my_psl2_defcheck.py and my_sz8_check.py).
