make -C v4 post79, run twice on 0a0be0ca's tree (2026-11-20), checking #190.
- v4/system/post_cases.c and docs/v4.0.0/POST79.md: the same on both runs, and the same as committed.
- docs/v4.0.0/POST.md: not the same. POST.md-run1 and POST.md-run2 are the two; run1-vs-run2.diff
shows five cells of the "v3 printed" column that differ. Three are addresses (FIND.user_word,
>BODY.basic, >BODY.variable), despite setarch -R; two are INFER-VARIANCE@.after_run and
INFER-EARLY-EXIT@.after_run, which come of the run's history. POST.md-run2.diff is run 2 against
the committed copy. POST.md was then put back as committed.