[FlyTheElephant1]

Motivation

The temporary 6480-block coinbase maturity window (heights 973440–979919) currently ends at 979920 and releases every coin in that window at once. This change makes the 6480 maturity permanent by setting CoinbaseMaturityLongReleaseHeight to INT_MAX and CoinbaseMaturityLong to 6480. Coins created at or after 973440 mature in 6480 blocks.

The revalidation marker is extended to the same permanent range so nodes that previously certified the temporary window re-check from the start height onward.

Tests

src/test/coinbase_maturity_tests.cpp updated so mainnet asserts the permanent 6480 + max release height while testnet4 and regtest keep their original temporary-window checks.
./build/bin/test_bitcoin --run_test=coinbase_maturity_tests reports no errors.

[geobees]

Tested 235347e (Ubuntu on WSL2, GCC 15.2.0, no GUI). Not an ACK yet, one unit test regression:

Fails: validation_chainstatemanager_tests/chainstatemanager_snapshot_revalidation_waits_for_background
validation_chainstatemanager_tests.cpp(703): check marker.stop_height == 102 has failed [103 != 102]

The same test passes on the base (58398ba, 10 test cases, no errors), so this comes from the PR. Cause: the shared helper AddLongCoinbaseMaturityRevalidationDeployment changed stop_height from ReleaseHeight - 1 to ReleaseHeight, which moves the regtest revalidation range by one block. coinbase_maturity_tests was updated for it (149 -> 150) but this test was not. The mainnet change doesn't need that edit, since INT_MAX - 1 is already a valid stop height.

Verified locally: reverting the stop_height line in chainparams.cpp to ReleaseHeight - 1 makes validation_chainstatemanager_tests pass again; coinbase_maturity_tests then only needs its three stop_height expectations reverted (mainnet INT_MAX - 1, testnet4 158110, regtest 149).

Passes: build; coinbase_maturity_tests; unit suite 145/146 (only the failure above); functional: feature_chainstate_revalidation, mempool_long_coinbase_maturity, mempool_spend_coinbase.

Nit: with the release height at INT_MAX, rpc/blockchain.cpp (~line 1967) returns early, so the long-maturity window info (height_end, maturity) appears to stop being reported on mainnet. Tools reading it will no longer see it.

Happy to re-test once the test is fixed.

[geobees]

Re-tested 98f8ea8: the chainparams revert fixes validation_chainstatemanager_tests (passes now). One assertion left in coinbase_maturity_tests.cpp:48 (testnet4): test_deployment.stop_height == test_consensus.CoinbaseMaturityLongReleaseHeight fails [158110 != 158111]. It needs - 1 like the mainnet check above it.

[geobees]

Re-tested eb6e889 (Ubuntu on WSL2, GCC 15.2.0, no GUI): the earlier failures are fixed.

• build OK
• coinbase_maturity_tests: pass
• validation_chainstatemanager_tests: pass
• unit suite: 146/146
• functional: feature_chainstate_revalidation, mempool_long_coinbase_maturity, mempool_spend_coinbase pass

Only nit left: with ReleaseHeight at INT_MAX, rpc/blockchain.cpp (~line 1967) returns early, so height_end / maturity appear to stop being reported on mainnet.

[geobees]

Tested c705c3a on top of 58398ba.

• Build OK, ctest 146/146
• coinbase_maturity_tests, feature_chainstate_revalidation, mempool_long_coinbase_maturity, mempool_spend_coinbase: pass
• rpc_blockchain.py (v1 and v2 transport): FAIL. eb6e889 passes.

Dropping the early return in LongCoinbaseMaturityDescPushBack makes networks that never configure the deployment (e.g. regtest) report:
long_coinbase_maturity {height: 2147483647, coinbase_start_height: 2147483647, maturity: 100, active: false}
but _test_getdeploymentinfo expects deployments: {}.

Adding this line at the top of the function makes rpc_blockchain.py pass on both transports:
if (params.CoinbaseMaturityLongEnforceHeight == std::numeric_limits::max()) return;
Mainnet still reports the entry, and the height_end guard stays as is. I did not re-run the full suite with that line.

[geobees]

Tested 4bd9f45 on top of 58398ba.

As pushed it does not build: in LongCoinbaseMaturityDescPushBack the new guard
if (params.CoinbaseMaturityLongEnforceHeight == std::numeric_limits<int>::max()) return;
comes before const Consensus::Params& params{chainman.GetConsensus()};
-> src/rpc/blockchain.cpp:1966: error: 'params' was not declared in this scope

With those two lines swapped (declaration first, then the guard) I get:

• build OK
• ctest 146/146
• rpc_blockchain.py (v1 and v2 transport), feature_chainstate_revalidation, mempool_long_coinbase_maturity, mempool_spend_coinbase: all pass

So the logic looks right...it just needs the lines reordered.

[geobees]

Tested b525aec

Ubuntu on WSL2, GCC 15.2, no GUI, on top of 58398ba.

• build OK
• ctest 146/146 (incl. coinbase_maturity_tests, validation_chainstatemanager_tests)
• functional: rpc_blockchain (v1+v2 transport), feature_chainstate_revalidation, mempool_long_coinbase_maturity, mempool_spend_coinbase: pass

Earlier heads had a validation_chainstatemanager_tests failure (stop_height), an rpc_blockchain failure on unconfigured networks (c705c3a) and a compile error (4bd9f45); all fixed here. I did not test combining with #468.

[FlyTheElephant1]

./build/bin/test_bitcoin
Running 765 test cases...

*** No errors detected

I got the same result as you on the ctest.

Debian bookworm.

[geobees]

Tested ACK b525aec (details above).

[chainboarder]

Concept NACK

[mynpub]

Concept NACK

[ghastlyxbt]

Concept NACK.

Too small CBM to be effective.