feat(math): un-mirror priority_metric (#2571 resolved) + Q2 prev-tick group-votes by jucor · Pull Request #2645 · compdemocracy/polis

jucor · 2026-07-22T00:51:10Z

Restore the real priority_metric branching formula in BOTH engine modes:
Clojure HEAD passes a real boolean since #2611 (conversation.clj:686, merged
2026-07-18), so the all-49 #1961 bug-mirror (#2571) is no longer the parity
behavior — the correct formula is.

Q2 (CLOJURE_QUIRKS): Clojure's :comment-priorities node SHADOWS its
current-tick group-votes input with (:group-votes conv) — the PREVIOUS
tick's stored value (conversation.clj:658). clojure-legacy mode now consumes
the captured prev-tick group-votes ({} on the first tick == Clojure's nil);
improved mode keeps the current-tick read. The current tick's group-votes
are stored on self.group_votes each tick (both modes, like self.pca) as the
in-memory analogue of Clojure persisting :group-votes in math_main. Restart
seam (from_dict/poller reload of group-votes) deferred to the poller
equivalence phase — noted in journal.

Stale-mirror test sites reworked: the designed xfail on
test_priority_metric_non_meta_squared is harvested (formula test now gates);
TestD12CommentPriorities now pins varied (non-all-49) Python output +
coverage instead of the mirror signature; the exact-value comparison in
test_legacy_clojure_regression.py::test_comment_priorities is xfailed for
ALL variants (reference blobs predate the Clojure #2611 fix — value parity
is validated by the H-B replay battery until blob regen with a fixed
generator).

TDD: tests/test_priority_unmirror.py — RED observed (formula tests fail on
the mirror; prev-tick flow + group_votes storage fail on old wiring), 7
GREEN after. Targeted: 13 passed (unmirror+float-serialization), 21 passed
(D12/priority subset), 4 xfailed (legacy regression value comparisons).

commit-id:a85047a7

Stack:

⚠️ Part of a stack created by spr. Do not merge manually using the UI - doing so may have unexpected results.

… group-votes Restore the real priority_metric branching formula in BOTH engine modes: Clojure HEAD passes a real boolean since #2611 (conversation.clj:686, merged 2026-07-18), so the all-49 #1961 bug-mirror (#2571) is no longer the parity behavior — the correct formula is. Q2 (CLOJURE_QUIRKS): Clojure's :comment-priorities node SHADOWS its current-tick group-votes input with (:group-votes conv) — the PREVIOUS tick's stored value (conversation.clj:658). clojure-legacy mode now consumes the captured prev-tick group-votes ({} on the first tick == Clojure's nil); improved mode keeps the current-tick read. The current tick's group-votes are stored on self.group_votes each tick (both modes, like self.pca) as the in-memory analogue of Clojure persisting :group-votes in math_main. Restart seam (from_dict/poller reload of group-votes) deferred to the poller equivalence phase — noted in journal. Stale-mirror test sites reworked: the designed xfail on test_priority_metric_non_meta_squared is harvested (formula test now gates); TestD12CommentPriorities now pins varied (non-all-49) Python output + coverage instead of the mirror signature; the exact-value comparison in test_legacy_clojure_regression.py::test_comment_priorities is xfailed for ALL variants (reference blobs predate the Clojure #2611 fix — value parity is validated by the H-B replay battery until blob regen with a fixed generator). TDD: tests/test_priority_unmirror.py — RED observed (formula tests fail on the mirror; prev-tick flow + group_votes storage fail on old wiring), 7 GREEN after. Targeted: 13 passed (unmirror+float-serialization), 21 passed (D12/priority subset), 4 xfailed (legacy regression value comparisons). commit-id:a85047a7

jucor marked this pull request as draft July 22, 2026 00:52