Re-run of the reproducibility package, 17 October 2026 (while the package was assembled)

Environment: Python 3.13.5 (macOS, Apple silicon); numpy 2.5.2, sympy 1.14.0, scipy 1.18.1, mpmath 1.3.0.
Every program was run with "nice -n 10 python3 -I <program> [argument] > <new output file>" inside its own folder
of an extracted copy of this package, in three lanes run in parallel (at most three processes at a time).
The new outputs were compared with the recorded outputs by cmp. Wall-clock times are rounded.

Program                                              time      comparison with the recorded output
check_note.py                                        0.5 s     identical to check_note_output.txt
original/scripts/01_word_and_certificate.py          0.1 s     identical
original/scripts/02_trace_polynomial.py NEW.txt      6.5 s     identical except the line "polynomials written to ..."
                                                               (it prints the file name given as argument);
                                                               the polynomial file is identical to outputs/02_polynomials.txt
original/scripts/03_numerics_elliptic.py            19.8 s     identical
original/scripts/04_lifted_rotation_numbers.py       8.6 s     identical
original/scripts/05_identities_and_pingpong.py       0.7 s     identical
original/scripts/06_propD_one_elliptic.py            2.4 s     identical
original/scripts/07_second_pass_checks.py            3.3 s     identical
verification_run_A/v1_certificate.py                 0.1 s     identical
verification_run_A/v2_trace_poly.py NEW.txt          1.5 s     identical; polynomial file identical to v2_polynomials.txt
verification_run_A/v3_lifts.py                       0.8 s     identical
verification_run_A/v4_readings.py                   21.8 s     identical
verification_run_A/v5_schottky_exact.py              0.1 s     identical
verification_run_A/v5b_schottky_deeper.py            0.4 s     identical
verification_run_A/v5c_schottky_random_exact.py      3.6 s     identical
verification_run_A/v6_identities.py                  0.4 s     identical
verification_run_A/v7_theoremA_numeric.py            0.1 s     identical
verification_run_A/v8_negative_trace_pair.py         0.4 s     identical
verification_run_B/own_01_word_certificate.py        0.1 s     identical
verification_run_B/own_02_trace_polynomial.py        3.2 s     identical; the file own_02_polynomials.txt it writes is
                                                               identical to own_02_polynomials_recorded.txt
verification_run_B/own_02b_h_sign_map.py             0.6 s     identical
verification_run_B/own_03_identities.py              0.3 s     identical
verification_run_B/own_04_theoremA_attack.py         2.0 s     identical
verification_run_B/own_05_propD_and_rotation.py     54.4 s     identical
verification_run_B/own_06_negative_trace_cases.py    2.2 s     identical
verification_run_B/own_06b_pm1_certificates.py       0.6 s     identical
verification_run_B/own_07_lower_sheet_and_pingpong.py 2.0 s    identical
verification_run_B/own_08_kapovich_degree.py         0.1 s     identical

The outputs in verification_run_B/claim_rerun_outputs/ (the second verification run's re-run of original/scripts)
are identical to original/outputs/, except the line "polynomials written to ..." of 02_trace_polynomial.out.

All programs exited with code 0.

Second re-run, 17 October 2026, by the verification run on the final text (programs in independent_run_2/)

The archive source.zip of the package (before independent_run_2/ was added) was extracted into a new folder and
every program was run there with "nice -n 10 python -I <program> [argument]" inside its own folder, four
processes at a time (Python 3.13.5; numpy 2.5.2, sympy 1.14.0, scipy 1.18.1, mpmath 1.3.0). All 28 programs exited
with code 0, and every output is identical to the recorded one, except the line "polynomials written to ..." of
original/scripts/02_trace_polynomial.py; the polynomial files written by original/scripts/02,
verification_run_A/v2 and verification_run_B/own_02 are identical to the recorded ones. Details and wall-clock
times: independent_run_2/rerun_report.txt.

The four programs of independent_run_2/ were run when they were written and again from their copies in the
package; the outputs are identical (run2_01_exact.py 1.0 s, run2_02_tracepoly.py 0.3 s, run2_03_theoremA.py
0.3 s, run2_04_mpmath_rotation.py 2.7 s; the last one needs mpmath).
