Second same-parameter case run, explicitly authorized to repair recording/captions. Total case executions now2; no third run. Same48x48, bounds(16,4,32,28), W12/A.02/nu.1/rho1, initial state, guard values and1200 coarse/2400 fine steps. Only example record/description/caption changes; all other baseline files unchanged.
Original run source saved byte-for-byte at research/fixed_patch_jet_20261013/fixed_patch_jet.py, SHA256 f27cd067fd7b9a7c6d7fab19fa471fe0617ee67889ae3931048557cc5659181b. Original PNG/stdout/summary/results/manifest retained unmodified. Historical source/doc wording is retained as evidence, not silently rewritten.
New stdout.txt is unmodified directly redirected merged stdout/stderr; 25 complete JSON samples exactly match output/summary.json.samples at0,50,...1200. Records include t, macro Q_in/out=sum(rho*vy), node-weighted leaf mass/change, density deviation, sampled maximum Mach and two centerline probes at16.25(fine)/32.5(coarse) with explicit native-column arithmetic mean. Only three macro-field snapshots retained for plots. Native-cell plots inspected, on-site clipping preserved; new captions note flow may cross patch sides and nozzle remains coarse.
Measured exit0, elapsed6.702868s, sampled maximum Mach0.0341615929682, sampled rho range[0.992553949356,1.010591745377]. Discrete leaf mass2208.00001640618 ->2202.30112983007, delta -5.69888657611. Final macro Q_in0.127690201625, Q_out0.0306652393192. Three shared snapshot scalar diagnostics and fluxes are identical to first run. Do not infer steady state or population-budget closure from these observations. Sampling every50 steps is not a bound on every-step maxima.
Scale estimates use continuous peak A=.02, not mean inlet velocity; on-site H=NY-3=45. W/A600, H/A2250, W^2/nu1440, H^2/nu20250; acoustic crossing~77.94 is not a transient-decay time.
Live precheck UUID/occupancy/VRAM/compute-apps/RAM in before.json. Real production compilation, private scratch/plot cache, no256MiB cap. Peak sampled group RSS281206784 bytes; minimum RAM63706316800 bytes, minimum GPU free15826MiB; combined task temporary/related evidence peak2192956 bytes (<2GB). Own temporary runner, compilation scratch and plot cache removed. No environment/system/dependency/DynamisLab edits, commit or push.
Read-only agent confirmed no numerical/guard change and no backend/kernel/helper modification. AST syntax passed before execution; no other tests or numerical cases run.
