test: Add wallet reorg test during assumeutxo background sync #34535

b-l-u-e commented at 9:23 PM on February 8, 2026: contributor

This PR adds a test that verifies wallet balance consistency when a chain reorg occurs while assumeutxo background validation is in progress

motivation currently test file tests wallet functionality with snapshots but there is no test covering what happens when reorg occurs during background validation phase.

so what it tests

wallet state remains consistent through reorg
balance updates when coins are orphaned

DrahtBot added the label Tests on Feb 8, 2026

DrahtBot commented at 9:24 PM on February 8, 2026: contributor

The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

Reviews

See the guideline for information on the review process.

Type	Reviewers
Concept ACK	yuvicc, aditrajj
Stale ACK	Bortlesboat

If your review is incorrectly listed, please copy-paste <code></code> into the comment that the bot should ignore.

b-l-u-e marked this as ready for review on Feb 9, 2026

in test/functional/wallet_assumeutxo.py:156 in 3a8f776299

 151 | +        # Build an alternative chain on n0 that excludes the post snapshot wallet payment
 152 | +        fork_point = SNAPSHOT_BASE_HEIGHT
 153 | +        orig_height = n0.getblockcount()
 154 | +        orig_chainwork = int(n0.getblockchaininfo()['chainwork'], 16)
 155 | +        self.disconnect_nodes(n3.index, n0.index)
 156 | +        n0.invalidateblock(n0.getblockhash(fork_point + 1))

yuvicc commented at 5:04 AM on February 16, 2026:

Instead of using invalidateblock directly, we could leverage fork-based reorg testing here so that it matches the real re-org scenario. For more see mempool_updatefromblock.py test.

b-l-u-e commented at 7:55 PM on February 16, 2026:

i made changes to use fork-based reorg , thanks for the tip

yuvicc commented at 5:05 AM on February 16, 2026: contributor

Concept ACK

b-l-u-e force-pushed on Feb 16, 2026

DrahtBot added the label CI failed on Feb 16, 2026

DrahtBot removed the label CI failed on Feb 16, 2026

in test/functional/wallet_assumeutxo.py:145 in 75368100aa outdated

 140 | +        payment_addr = w_reorg.getnewaddress()
 141 | +        pay_amount = 5
 142 | +        funding_wallet = n0.get_wallet_rpc("w")
 143 | +        txid = funding_wallet.sendtoaddress(payment_addr, pay_amount)
 144 | +        self.generate(n0, nblocks=2, sync_fun=self.no_op)
 145 | +        # Ensure the assumeutxo node receives the blocks with the payment

Bortlesboat commented at 10:12 PM on March 15, 2026:

nit: 180s timeout seems overly generous here. At this point the blocks have already been synced via sync_blocks and the payment has confirmations, so gettransaction should return quickly. Would 30-60s be sufficient? Long timeouts can mask real issues by making flaky tests take a long time to fail rather than failing fast.

b-l-u-e commented at 8:28 PM on April 27, 2026:

thanks i will update to 60s

Bortlesboat commented at 10:13 PM on March 15, 2026: none

ACK 75368100aa36ce136039b734c317354a04ef783b. Ran wallet_assumeutxo.py locally on Ubuntu 24.04, all tests pass including the new reorg-during-background-sync test.

DrahtBot requested review from yuvicc on Mar 15, 2026

aditrajj commented at 7:25 AM on March 20, 2026: none

Concept ACK

in test/functional/wallet_assumeutxo.py:120 in 75368100aa

 115 | +
 116 | +        # Restart pruned node fresh to ensure snapshot has more work than active chainstate
 117 | +        self.stop_node(n3.index)
 118 | +        rmtree(n3.chain_path)
 119 | +        self.start_node(n3.index, extra_args=self.extra_args[n3.index])
 120 | +        n3.setmocktime(n0.getblockheader(n0.getbestblockhash())['time'])

maflcko commented at 7:54 AM on March 20, 2026:

I think according to #34439, rmtree should be avoided, if possible.

I think this is also trivial to avoid with a hacky hack:

Start the node with -reindex -mocktime=123 (This will reject blocks and headers, and delete them?)
The n3.setmocktime call here will then recover the mocktime, and the test should continue normall

However, I haven't tried if this hack actually works and if the test passes

b-l-u-e commented at 9:55 PM on April 27, 2026:

Tried the approach and it works thanks for the hack. I updated the test and passes locally

<details> <summary>test output</summary>

2026-04-27T21:47:05.297236Z TestFramework (INFO): PRNG seed is: 8502276312754656986
2026-04-27T21:47:05.347981Z TestFramework (INFO): Initializing test directory /tmp/bitcoin_func_test_jjdqbo8r
2026-04-27T21:47:06.603434Z TestFramework (INFO): -- Testing assumeutxo
2026-04-27T21:47:06.605792Z TestFramework (INFO): Creating a UTXO snapshot at height 299
2026-04-27T21:47:07.147304Z TestFramework (INFO): Loading snapshot into second node from /tmp/bitcoin_func_test_jjdqbo8r/node0/regtest/utxos.dat
2026-04-27T21:47:07.185135Z TestFramework (INFO): Backup from the snapshot height can be loaded during background sync
2026-04-27T21:47:07.190968Z TestFramework (INFO): Backup from before the snapshot height can't be loaded during background sync
2026-04-27T21:47:07.194877Z TestFramework (INFO): Backup from the snapshot height can be loaded during background sync (pruned node)
2026-04-27T21:47:07.230742Z TestFramework (INFO): Backup from before the snapshot height can't be loaded during background sync (pruned node)
2026-04-27T21:47:07.240374Z TestFramework (INFO): Test loading descriptors during background sync
2026-04-27T21:47:07.257714Z TestFramework (INFO): Test that rescanning blocks from before the snapshot fails when blocks are not available from the background sync yet
2026-04-27T21:47:07.258696Z TestFramework (INFO): Unload wallets and sync node up to height 359
2026-04-27T21:47:07.487422Z TestFramework (INFO): Verify node state during background sync
2026-04-27T21:47:07.626065Z TestFramework (INFO): Ensuring wallet can be restored from a backup that was created before the snapshot height
2026-04-27T21:47:07.651290Z TestFramework (INFO): Check balance of a wallet that is active during snapshot completion
2026-04-27T21:47:08.939432Z TestFramework (INFO): Ensuring descriptors can be loaded after background sync
2026-04-27T21:47:08.954958Z TestFramework (INFO): Ensuring wallet can't be restored from a backup that was created before the pruneheight (pruned node)
2026-04-27T21:47:09.210299Z TestFramework (INFO): Ensuring wallet can be restored from a backup that was created at the pruneheight (pruned node)
2026-04-27T21:47:09.223229Z TestFramework (INFO): Ensure wallet balances stay consistent across a reorg during assumeutxo background sync
2026-04-27T21:47:11.899506Z TestFramework (INFO): Stopping nodes
2026-04-27T21:47:12.057854Z TestFramework (INFO): Cleaning up /tmp/bitcoin_func_test_jjdqbo8r on exit
2026-04-27T21:47:12.057971Z TestFramework (INFO): Tests successful

sedited commented at 2:35 PM on April 27, 2026: contributor

@b-l-u-e There are a bunch of unresolved review comments here. Can you respond to them?

b-l-u-e commented at 7:31 PM on April 27, 2026: contributor

@b-l-u-e There are a bunch of unresolved review comments here. Can you respond to them?

thanks for the reminder

b-l-u-e force-pushed on Apr 27, 2026

b-l-u-e requested review from Bortlesboat on Apr 27, 2026

b-l-u-e requested review from maflcko on Apr 27, 2026

Bortlesboat commented at 6:24 AM on April 28, 2026: none

ACK ad99afd3da255703547e00f1fe8e4bd60a403c75. Re-reviewed the changes since my previous ACK 75368100aa36ce136039b734c317354a04ef783b, including the switch from removing the pruned node chain directory to the -reindex/-mocktime path and the shorter transaction confirmation timeout. Built current tip on WSL2 Ubuntu 24.04 with wallet support enabled and ran test/functional/wallet_assumeutxo.py; the test passed.

test: Add wallet reorg test during assumeutxo background sync

Signed-off-by: b-l-u-e <8102260+blue@users.noreply.github.com>

e42bff60df

in test/functional/wallet_assumeutxo.py:163 in ad99afd3da

 158 | +        tip_time = n0.getblockheader(n0.getbestblockhash())['time']
 159 | +        fork_blocks = []
 160 | +        for i in range((orig_height - fork_point) + 2):
 161 | +            block = create_block(
 162 | +                hashprev=fork_block_hash,
 163 | +                coinbase=create_coinbase(height=fork_point + 1 + i),

maflcko commented at 6:51 AM on April 28, 2026:

                height=fork_point + 1 + i,

nit, I think this should work as well on current master, but I haven't tried

b-l-u-e force-pushed on Apr 28, 2026

sedited requested review from maflcko on May 2, 2026