Testnet go checklist and bench-log: the pin's known-failed gate form fails as it must (10^10 on every payload and reward); form 1's remaining lines

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

docs/bench-log.md

@@ -2742,6 +2742,7 @@ The go object as landed (the 0.3.24 node pin 5b673577: the Devnet 3 object commi
 | the UTXO side, last 30 blocks | identical on both nodes; 30 payload subsidies at the 18-decimal schedule, 0 wrong; 80/20 exact on 20 of 20 single-payee coinbases |
 | the execution layer | 512 rewards against the schedule at each blue block's own DAA, 0 wrong; the sum of the rewards equals eth_getBalance on both nodes: 4,097,208,879,629,629,629,622 wei = 4,097.20887963 IGN |
 | the start-up lines | both heads on digest `b2e856ed...`; no public seed dialled |
+| the known-failed form | the 8-decimal schedule against the same chain shape (420 s, 286 blocks): payloads wrong 30 of 30, rewards wrong 285 of 285, each by exactly 10^10; the sinks, tips and balances still equal between the nodes; FAIL |
 
 The fast-time line of the object on the same pair (six keys, 10:33 to 10:43 UK): 15 of 16 checks true (the stamped object byte 7 on every mined
 block, class v4 at rung 0, the first lock at DAA 142 and lock 19 at the end, the bonus 72/28 on the silent key's 91 blocks and 80/20 on the

docs/plans/testnet-go.md

@@ -172,7 +172,7 @@ Mac: 27 passed (the crate is not in the PC build inputs). The packager's self-te
 | 9d | Windows: `igneum-prove-host` beside every Windows node (verification from genesis); PC 1's node refuses to start without it | the release engineer (the 0.4.0 package) | NOT DONE: to check in the 0.4.0 Windows installer |
 | 9j | The two-node 18-decimal gate on the FINAL binaries (fork e6dd3afd, igneumd 58f96049..., igneum-miner 5625caee...; `tools/fleet/base-unit-gate.sh` 72b02b48, the GPU form on a RunPod RTX A5000 rented and run by the fleet lane): PASS at 15:19 UK, 7 October 2026: 322 chain blocks in 600 s, both nodes on one sink and one exec tip (0x142), the 30 inspected payloads all the 18-decimal schedule, 491 segment rewards against the schedule at each blue block's own DAA with 0 wrong, eth_getBalance 3,929.11134259 IGN (3,929,111,342,592,592,592,602 wei) equal on both nodes and to the sum of the rewards; both node heads print `Base unit: 10^18`, the digest `4fbb2152...` and `igneumd/2.1.0-e6dd3afd`, node B's exec line the genesis `01294fd3...`. The known-failed form (the 8-decimal schedule, 420 s) FAILED as it must at 15:27 UK: payloads wrong 30 of 30, rewards wrong 305 of 305. The logs: ~/igneum-fleet/tn-gate/out/ on the Mac | the fleet lane ran it, this lane reads it | PASS (the cut's closing gate) ON THE 7 OCTOBER RE-CUT; RE-RUNS ON THE GO OBJECT (the ruling of 19:2x UK) |
 | 9h | The two-node 18-decimal gate on the re-cut binaries (`tools/fleet/base-unit-gate.sh` 72b02b48, the GPU form: RunPod RTX 3090, driver 580.65.06, the hive CUDA worker, rented and run by the fleet lane): PASS at 10:08 UK, 7 October 2026, on igneumd 8a6f1f56... and igneum-miner 726cf290... (fork commit f1717419): 577 blocks in 600 s, 421 chain blocks, both nodes on one sink and one exec tip (0x1a5), the 30 inspected coinbase payloads all the 18-decimal schedule (day-0 block 10 IGN = 10^19 base units), 80/20 exact on 18 of 18 single-payee coinbases, 577 segment rewards checked against the schedule at each blue block's own DAA with 0 wrong, eth_getBalance 4,617.53654629 IGN (4,617,536,546,296,296,296,298 wei) equal on both nodes and to the sum of the rewards; both node heads carry `Base unit: 10^18`, the digest `80af8aa1...`, the ladder active at rung 0, fees v1 from DAA 0 and proof verification from DAA 0 under the pinned ids. The first run of the same form (09:4x UK) read 165 of 561 rewards one ramp step low: the gate's check 5 still priced every reward at the chain block's DAA, fixed in 72b02b48 (each blue block its own DAA, the per-block rule). On the testnet object itself no voter exists inside a run (the 30-day window), so this line shows the layout, the switches and the identity without the bonus; the devnet-suffix line (9g) is the proof of the identity under the bonus | the fleet lane ran it, this lane reads it | PASS; the known-failed form (GATE_EXPECT_DECIMALS=8, the 8-decimal schedule against the same chain, 420 s) FAILED as it must at 10:15 UK: payload subsidies wrong 30 of 30, every segment reward wrong; the pod destroyed on this lane's done line ON THE 7 OCTOBER RE-CUT; RE-RUNS ON THE GO OBJECT (the ruling of 19:2x UK) |
-| 9p | The pin's first pod run (8 October 2026, 09:45 to 10:00 UK, RunPod 3090 2rls7gfwii2d3g, the fleet hand): the fast-time line on 5b673577's pair read 14 of 16 checks true (start lines, one digest, class v4 at rung 0 ten epochs, 0 rejected, one sink on five nodes, the first lock at DAA 144, the silent key paid 72/28 on 127 blocks and the voters 80/20 on 318, the bridge identity exact for all five, a leave accepted and effective within the delay, no pause after it) and two false that were the harness's: the object-byte check expected 0 where the pin stamps byte 7 from genesis (the Devnet 3 shape), and "finality active at the end" sat on scenario arithmetic (one silent key of five at a quarter of the window held over a third of a 120-block window on a shared pod CPU, and the rule paused as designed); fixed at testnet-genesis-2 940d91de (the stamped byte read from the node's line; six keys). The two-node gate was void on that pod (its GPU dead for CUDA, cuInit 999: a host fault, the pod retired), and it showed base-unit-gate.sh dialling the public seeds (refused on the digest, harmless), fixed at 940d91de with `--nodnsseed`. Both gates re-run on pod 2 (tn-gate-0324b, RunPod A5000 secure yiu8x5dnh01g2y, driver 580.173.02, rented 10:29 UK after four 3090 hosts failed cuInit at rent; the hive 0.3.24 CUDA worker at 41.65 MH/s), the tree at 940d91de: the two-node 18-decimal gate form 1 PASS at 10:43 UK (340 chain blocks, 30 payloads at the 18-decimal schedule, 80/20 exact on 20 of 20, 512 rewards against the schedule at each blue block's own DAA with 0 wrong, eth_getBalance 4,097.20887963 IGN = 4,097,208,879,629,629,629,622 wei equal on both nodes and to the rewards' sum, both heads on `b2e856ed...`, no seed dialled); the fast-time line 15 of 16 true (six keys: the stamped byte 7 on 587 blocks and no ladder bits, class v4 at rung 0 ten epochs, 0 rejected, one sink on six nodes, the first lock at DAA 142 and lock 19 at the run's end, the silent key paid 72/28 on 91 blocks and the voters 80/20 on 352, the bridge identity exact for all six, a leave accepted and effective, no pause after it) with the one false check the harness's own stop order (its final read came after the voters had exited with the run; fixed at b8074151: the miners outlive the run and the read is taken as it ends); form 2 (the 8-decimal known-failed form) and the object line's re-run on b8074151 owed | the fleet hand runs, this lane reads | form 1 PASS; 15 of 16 with the harness fault fixed; form 2 and the re-run OWED |
+| 9p | The pin's first pod run (8 October 2026, 09:45 to 10:00 UK, RunPod 3090 2rls7gfwii2d3g, the fleet hand): the fast-time line on 5b673577's pair read 14 of 16 checks true (start lines, one digest, class v4 at rung 0 ten epochs, 0 rejected, one sink on five nodes, the first lock at DAA 144, the silent key paid 72/28 on 127 blocks and the voters 80/20 on 318, the bridge identity exact for all five, a leave accepted and effective within the delay, no pause after it) and two false that were the harness's: the object-byte check expected 0 where the pin stamps byte 7 from genesis (the Devnet 3 shape), and "finality active at the end" sat on scenario arithmetic (one silent key of five at a quarter of the window held over a third of a 120-block window on a shared pod CPU, and the rule paused as designed); fixed at testnet-genesis-2 940d91de (the stamped byte read from the node's line; six keys). The two-node gate was void on that pod (its GPU dead for CUDA, cuInit 999: a host fault, the pod retired), and it showed base-unit-gate.sh dialling the public seeds (refused on the digest, harmless), fixed at 940d91de with `--nodnsseed`. Both gates re-run on pod 2 (tn-gate-0324b, RunPod A5000 secure yiu8x5dnh01g2y, driver 580.173.02, rented 10:29 UK after four 3090 hosts failed cuInit at rent; the hive 0.3.24 CUDA worker at 41.65 MH/s), the tree at 940d91de: the two-node 18-decimal gate form 1 PASS at 10:43 UK (340 chain blocks, 30 payloads at the 18-decimal schedule, 80/20 exact on 20 of 20, 512 rewards against the schedule at each blue block's own DAA with 0 wrong, eth_getBalance 4,097.20887963 IGN = 4,097,208,879,629,629,629,622 wei equal on both nodes and to the rewards' sum, both heads on `b2e856ed...`, no seed dialled); the fast-time line 15 of 16 true (six keys: the stamped byte 7 on 587 blocks and no ladder bits, class v4 at rung 0 ten epochs, 0 rejected, one sink on six nodes, the first lock at DAA 142 and lock 19 at the run's end, the silent key paid 72/28 on 91 blocks and the voters 80/20 on 352, the bridge identity exact for all six, a leave accepted and effective, no pause after it) with the one false check the harness's own stop order (its final read came after the voters had exited with the run; fixed at b8074151: the miners outlive the run and the read is taken as it ends); form 2 (the 8-decimal known-failed form, 420 s, 10:44 to 10:51 UK) FAILED as it must: 286 blocks, the sinks, the exec tips (0xbb) and the balances equal on both nodes, 30 of 30 payloads and 285 of 285 rewards wrong by exactly 10^10 (the 18-decimal chain read against the 8-decimal schedule), both heads on `b2e856ed...`; form 1's remaining lines: 513 blocks, exec tips 0x154 on both, exit 0. The object line's re-run on b8074151 owed | the fleet hand runs, this lane reads | form 1 PASS, form 2 FAIL as it must; the object line's re-run OWED |
 | 9g | The fast-time line of the object (`infra/fast-time/testnet-object.mjs`, a devnet-suffix network on igneum-build-1 at the 60x profile: 5 nodes, four voting keys and one `--no-vote` key, every switch of `testnet-object.json` from genesis, the fast-time finality profile with leave delay 60, the testnet schedule at the devnet's unit): PASS at 11:12 UK, 7 October 2026, run 10 on the archive binaries (igneumd 57ad7dc2..., fork 5c25c1fb): the start-up lines and one digest on every node; class v4 at rung 0 (27 passes) through ten epochs, no ladder bits, the object byte 0 on 551 blocks; 0 rejected, one sink on five nodes; the first lock at DAA 142 (172 s); the silent key's 87 blocks paid the bonus split on the UTXO side (72 percent of the subsidy at each block's own DAA, 720,118,333 sompi at DAA 142 against 800,131,481 for a voter) and the voters' 320 blocks the plain split; the bridge identity exact for all five miners over the chain (UTXO payments x 10^10 = execution credits less the tip's own: 96,832,949,992 sompi against 968,329,499,920,000,000,000 wei for voter a, and so on); a signed leave accepted at DAA 212, its key out of the voter count (5 to 4) at DAA 351, no pause longer than one second after it, finality active at the end at lock 18. The road to it: run 1 (pre-fix binaries) showed the executor crediting 80 percent to the silent key while the coinbase paid 72 (the N7 finding), run 7 the side blues credited one ramp step high when the per-block switch was left out of the profile (the N8 shape), run 9 every check green but the harness's own pause reading (one-second reason flickers at each new lock, corrected to consecutive seconds). On the testnet object itself the weight window is 30 days, so this line is the proof of the bonus and the identity before the go | this lane | PASS; the known-failed forms are on record from the same morning: run 1 on the pre-fix binaries (the silent key credited 80 percent on the EVM side against 72 on the UTXO side, every other check green) and run 7 on the fixed binaries with the per-block switch left out of the profile (every miner's credit above its payments by the side-blue steps). A binary that carries the per-block field without the bonus fix does not exist (the pre-fix binaries refuse an override naming the field), so no third form runs ON THE 7 OCTOBER RE-CUT; RE-RUNS ON THE GO OBJECT (the ruling of 19:2x UK) |
 | 9n | THE LAST STEP BEFORE THE GO (the coordinator's planned step, 22:4x UK, 7 October 2026): the object above holds the class v5 floor at never because Devnet 3's v5 crossing has not happened; the ruling is class v5 at 0 on the testnet IF that crossing reads clean (at DAA 68,400, about 12:54 BST on 8 October after the floor's last re-cut; the go not before the evening of 8 October). After the crossing's read the node lane re-cuts the testnet object once more with `program_class_v5_activation_daa` 0 (a new digest, the genesis unchanged), this table's digest row and the runbook's `--digest` are updated from the new binary's print, and the build-server lane re-arms the seeds on the pull-path dry run with the new binaries and sha. That re-cut object is the one the go uses; the object above is not placed on a seed if the crossing reads clean. If the crossing does not read clean, the object above stands and the v5 floor stays at never | the node lane re-cuts, this lane the table, the build-server lane re-arms | PLANNED, after the Devnet 3 v5 crossing's read |
 | 9o | Keyed payout addresses (the fleet lane's defect line of 8 October 2026, 09:4x UK, from Devnet 3: `tools/fleet/fleet.py` line 83 wrote a random payout address with no key for every rented box, so 578,000 IGN of Devnet 3 rewards by 22:28 UK sat on addresses nobody can spend from, and the load generator ran on the 432 IGN dev-fee key; the fleet's keyed-throwaway fix lands 8 October): before the go, every payout address in the testnet object and in anything the seeds or the first miner run has a key held by a named holder (the seeds mine nothing and name no payout; the first miner of step 10 pays the app's own keyed wallet or a named key; the dev-fee address is the app's). The faucet: the testnet genesis carries one IGN to OP_FALSE (unspendable) and no allocation, so a faucet is funded by the first keyed miner's blocks, never at genesis; a genesis allocation would be the project lead's word and a re-cut of the hash; the chain-side faucet or a funded key for load tests is logged for the 0.3.25 line (the node lane, 09:4x UK) | the fleet lane (its fix), the miner-community lead (the first miner's key), the project lead (any allocation) | OPEN |