Aetheria/tests
Repository files (latest commit first)
Filename Latest commit message Latest commit date
Meta Vens fd79201358 Fire control 12.4 fix batch F5: measure the steady state, not the first window
LaneAndSilhouetteAllocateNothingGivenAPooledBuffer failed in 1 of 12 full-suite
runs (1232 bytes over 1000 Lane calls) and in 0 of 100 runs of the test alone
or its class. The allocation is a one-off, not a per-call cost: under suite
load the runtime's background tier-up is delayed by other tests' JIT activity,
so first-use allocation lands inside the single measured window.

The rule stays: Lane and buffered Silhouette allocate nothing. The test now
measures up to 20 rounds of 1000 calls and requires the smallest round to be
zero. A per-call allocation appears in every round and still fails it.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-29 04:17:43 +02:00
..
Aetheria.Shared.Tests Fire control 12.4 fix batch F5: measure the steady state, not the first window 2026-09-29 04:17:43 +02:00
mutation_tests_addressables_cut.py Un-clobber the item-provenance mutation harness; own file for this cut 2026-09-18 05:07:23 +02:00
mutation_tests_shield_panel_cut1.py Add the Cut 1 batchmode probe and its mutation harness 2026-09-18 06:08:24 +02:00
mutation_tests_shield_panel_cut2.py Fix ShieldManager's missing-envelope fallback regression 2026-09-18 06:45:07 +02:00
mutation_tests_shield_panel_cut3.py Add the Cut 3 batchmode probe and its mutation harness 2026-09-18 13:51:44 +02:00