RERUN_LOG.txt -- re-runs of 14 October 2026 (Python 3.13.5, sympy 1.14.0, numpy 2.5.2; macOS, one laptop), made when
the note was written, from an extracted copy of the folder reproducibility/ as packed for the archive (before this
log and the running times in README.md were added).

Part 1.  Output of 'sh run_quick.sh <python>' (wall-clock time 164 s, one process at a time).
"identical": the new output is byte-identical to the recorded file after removing the fields that record
running times (patterns such as '(0.4s)', '[15s]', 'setup 4s', 'total time 28.4s', 'running time 12.8 s').

== writing_stage
identical    writing_stage/check_note_output.txt
== original
identical    original/outputs/out_expected_p4.txt
identical    original/outputs/out_expected_p6.txt
identical    original/outputs/out_expected_p8.txt
identical    original/outputs/out_direct_d1_p4.txt
identical    original/outputs/out_direct_d1_p6.txt
identical    original/outputs/out_p4_all_n.txt
identical    original/outputs/out_p6_n25.txt
identical    original/outputs/out_p8_n10.txt
identical    original/outputs/out_p10_n6.txt
identical    original/outputs/out_thresholds.txt
identical    original/outputs/out_bounds_check.txt
== verification_run_A
identical    verification_run_A/outputs/out_va_phi.txt
identical    verification_run_A/outputs/out_va_exact_selftest.txt
identical    verification_run_A/outputs/out_va_direct_d1.txt
identical    verification_run_A/outputs/out_va_Elat_small.txt
identical    verification_run_A/outputs/out_va_exact_p6_n25.txt
identical    verification_run_A/outputs/out_va_exact_p8_n10.txt
identical    verification_run_A/outputs/out_va_exact_p10_n6.txt
identical    verification_run_A/outputs/out_va_thresholds.txt
identical    verification_run_A/outputs/out_va_remark8.txt
identical    verification_run_A/outputs/out_va_values.txt
== verification_run_B
identical    verification_run_B/outputs/direct_d1_p4.txt
identical    verification_run_B/outputs/direct_d1_p6.txt
identical    verification_run_B/outputs/direct_d1_p8.txt
identical    verification_run_B/outputs/direct_d1_p10.txt
identical    verification_run_B/outputs/compare_d1.txt
identical    verification_run_B/outputs/compare_dD.txt
identical    verification_run_B/outputs/expansion_p4_n14.txt
identical    verification_run_B/outputs/cert_p6_n25.txt
identical    verification_run_B/outputs/cert_p8_n10.txt
identical    verification_run_B/outputs/cert_p10_n6.txt
identical    verification_run_B/outputs/thresholds_B.txt
identical    verification_run_B/outputs/thm4_highp.txt
identical    verification_run_B/outputs/negctl_B.txt

Part 2.  The three slow exact programs of Step 2 of Theorem 1.4 for p = 4, n <= 120, run separately (three
processes in parallel, nice -n 10), each inside its folder, output redirected to a new file:
  original/scripts:             p4_exact_all_n.py 120 4                              user time 107 s
  verification_run_A/programs:  va_exact_big.py 4 120 1,2,3,4                        user time  94 s
  verification_run_B/programs:  cert_B.py 4 120 1,2,3,4 --detail <new file>          user time 126 s
                                (without the folder of pickled phi tables used by run B; the program computes
                                the 118,616 classes itself)

identical    original/outputs/out_p4_d4_n120.txt
identical    verification_run_A/outputs/out_va_exact_p4_n120.txt
identical    verification_run_B/outputs/cert_p4_n120.txt
identical    verification_run_B/outputs/cert_p4_n120_detail.txt

Not re-run here: the Monte Carlo and quadrature programs (original/scripts/mc_check_d2.py; run A's va_mc_dd.py
and va_alt_reading_d1.py; run B's lemma1_and_mc.py, mc_d1_B.py, quad_Edisc_p_B.py), run B's scans (cert_B.py
with d up to 40) and phi_dump.py/hand_values_B.py.  Run B re-ran all programs of original/, including
mc_check_d2.py with the same seeds, and obtained every recorded output again
(verification_run_B/outputs/claim_rerun_compare.txt).
