test: fix coinbases built on arbitrary parents #36109

pull l0rinc wants to merge 5 commits into bitcoin:master from l0rinc:l0rinc/test-build-valid-fork-blocks changing 3 files +75 −18
  1. l0rinc commented at 12:17 AM on August 28, 2026: contributor

    Problem: Test helpers that attach active-tip templates to arbitrary parents retain the template's subsidy. Existing reorg tests can pass with underpaid coinbases. A fork crossing a halving can overpay and fail with bad-cb-amount. At heights 1 through 16, encoding only the height leaves the coinbase input script too short and triggers bad-cb-length.

    Fix: Share parent-dependent reconstruction between BuildChain() and MinerTestingSetup::Block(), deriving difficulty and coinbase fields from the actual parent. Pad rebuilt coinbase input scripts. Require coinbase-only templates so replacing the reward with the subsidy cannot discard transaction fees.

  2. DrahtBot added the label Tests on Aug 28, 2026
  3. DrahtBot commented at 12:17 AM on August 28, 2026: contributor

    <!--e57a25ab6845829454e8d69fc972939a-->

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

    <!--006a51241073e994b41acfe9ec718e94-->

    External sites

    <!--021abf342d371248e50ceaed478a90ca-->

    Reviews

    See the guideline and AI policy for information on the review process.

    Type Reviewers
    ACK maflcko, jeanpablojp
    Stale ACK Sjors

    If your review is incorrectly listed, please copy-paste <code>&lt;!--meta-tag:bot-skip--&gt;</code> into the comment that the bot should ignore.

    <!--174a7506f384e20aa4161008e828411d-->

    Conflicts

    Reviewers, this pull request conflicts with the following ones:

    • #36792 <sub><img src="https://drahtbot.space/ack_count/bitcoin/bitcoin/36792.svg"></sub> (validation: avoid partial override of non-valid validation state by optout21)
    • #36598 <sub><img src="https://drahtbot.space/ack_count/bitcoin/bitcoin/36598.svg"></sub> ([Demo] validation: remove AssumeUTXO by l0rinc)
    • #36521 <sub><img src="https://drahtbot.space/ack_count/bitcoin/bitcoin/36521.svg"></sub> (Keep witness stacks in serialized form in CScriptWitness by sipa)
    • #35570 <sub><img src="https://drahtbot.space/ack_count/bitcoin/bitcoin/35570.svg"></sub> (refactor: Change some validation.cpp methods to return BlockValidationState by optout21)

    If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.

    <!--5faf32d7da4f0f540f40219e4f7537a3-->

  4. jeanpablojp commented at 6:27 PM on September 8, 2026: contributor

    Concept ACK

    Built each commit and ran the unit tests, with a before and after of both cases against master.

  5. in src/test/util/mining.cpp:89 in 6052be3135
      84 | +        const int height{parent.nHeight + 1};
      85 | +        CMutableTransaction tx_coinbase{*block.vtx.at(0)};
      86 | +        tx_coinbase.nLockTime = static_cast<uint32_t>(parent.nHeight);
      87 | +        // Include a dummy OP_0 so low-height coinbase input scripts meet the minimum length
      88 | +        tx_coinbase.vin.at(0).scriptSig = CScript{} << height << OP_0;
      89 | +        tx_coinbase.vout.at(0).nValue = GetBlockSubsidy(height, params);
    


    jeanpablojp commented at 6:27 PM on September 8, 2026:

    This is an assignment rather than an adjustment, so the template's fees go with it. MinerTestingSetup::Block asks for its template with the mempool on, and with one transaction in it the coinbase reward comes out at 5000001000 on master and 5000000000 here, with the block still carrying that transaction.

    Nothing hits it today, since the mempool is empty at every call. Would .use_mempool = false on that template call be simpler? It makes the site match BuildChain, so the overwrite is exact, and the three suites stay green with it.


    l0rinc commented at 1:18 AM on September 25, 2026:

    Sure, added .use_mempool = false to MinerTestingSetup::Block() and asserted the shared helper receives a coinbase-only block.

  6. in src/test/util/mining.cpp:78 in 6052be3135 outdated
      74 | @@ -75,6 +75,23 @@ std::vector<std::shared_ptr<CBlock>> CreateBlockChain(size_t total_height, const
      75 |      return ret;
      76 |  }
      77 |  
      78 | +void RebuildBlockForParent(CBlock& block, const CBlockIndex& parent, uint32_t time, const Consensus::Params& params)
    


    jeanpablojp commented at 6:27 PM on September 8, 2026:

    Removing the target, subsidy or padding line here leaves blockfilter_index_tests and baseindex_tests green, so nothing on that side covers what the branch changes. The new assertions in MinerTestingSetup::Block catch the subsidy and the padding, and the target has nothing anywhere.

    These two cover all four fields, if they are worth having. The target one calls the helper directly, which is the only way in, since on regtest the recalculated target always equals the template's and no chain can tell the line apart. The coinbase one passes here and fails on master, first on the script length and then at the halving block. Both compile in blockfilter_index_tests.cpp.

    BOOST_FIXTURE_TEST_CASE(rebuild_block_for_parent_target, RegTestingSetup)
    {
        const auto& consensus{Params().GetConsensus()};
        const auto genesis{Params().GenesisBlock()};
        const auto parent_hash{genesis.GetHash()};
        CBlockIndex parent{genesis};
        parent.phashBlock = &parent_hash;
        --parent.nBits; // A valid target, different from the one the template carries.
    
        // Regtest allows min-difficulty blocks, and parent has no pprev, so
        // GetNextWorkRequired returns the parent's target below the threshold and the
        // pow limit above it.
        const auto threshold{uint32_t(2 * consensus.nPowTargetSpacing)};
        for (const auto delay : {1U, threshold, threshold + 1}) {
            CBlock block{genesis};
            RebuildBlockForParent(block, parent, parent.nTime + delay, consensus);
            BOOST_CHECK_EQUAL(block.nBits, delay > threshold ? genesis.nBits : parent.nBits);
        }
    }
    
    BOOST_FIXTURE_TEST_CASE(rebuild_block_for_parent_coinbase, TestChain100Setup)
    {
        const auto& consensus{Params().GetConsensus()};
        auto process = [&](const std::vector<std::shared_ptr<CBlock>>& chain) {
            for (const auto& block : chain) {
                BOOST_CHECK(m_node.chainman->ProcessNewBlock(block, /*force_processing=*/true, /*min_pow_checked=*/true, nullptr));
            }
        };
    
        // A fork rooted below height 17 needs the dummy OP_0, since the BIP34
        // height alone is one byte there (bad-cb-length).
        {
            const CBlockIndex* genesis{WITH_LOCK(cs_main, return m_node.chainman->ActiveChain().Genesis())};
            std::vector<std::shared_ptr<CBlock>> fork;
            BOOST_REQUIRE(BuildChain(m_node, genesis, CScript() << OP_TRUE, 3, fork));
            BOOST_CHECK_EQUAL(fork.front()->vtx[0]->vin[0].scriptSig.size(), 2U);
            process(fork);
        }
    
        // BuildChain only submits headers, so every template keeps being built on
        // the active tip. A chain long enough to pass the halving would otherwise
        // carry the tip's subsidy and fail to connect (bad-cb-amount).
        {
            const CBlockIndex* tip{WITH_LOCK(cs_main, return m_node.chainman->ActiveChain().Tip())};
            const int start{tip->nHeight};
            const int length{consensus.nSubsidyHalvingInterval + 10 - start};
            BOOST_REQUIRE(length > 0);
            std::vector<std::shared_ptr<CBlock>> chain;
            BOOST_REQUIRE(BuildChain(m_node, tip, CScript() << OP_TRUE, length, chain));
            for (int i = 0; i < length; ++i) {
                BOOST_CHECK_EQUAL(chain.at(i)->vtx[0]->vout[0].nValue, GetBlockSubsidy(start + 1 + i, consensus));
            }
            process(chain);
            BOOST_CHECK_EQUAL(WITH_LOCK(cs_main, return m_node.chainman->ActiveChain().Height()), start + length);
        }
    }
    

    l0rinc commented at 1:26 AM on September 25, 2026:

    Thanks! I added similar tests, with the coinbase cases combined into one fork test. I also credited you as a co-author on the commits containing them.

  7. DrahtBot added the label Needs rebase on Sep 24, 2026
  8. l0rinc force-pushed on Sep 25, 2026
  9. l0rinc commented at 1:32 AM on September 25, 2026: contributor

    Rebased to resolve conflicts and applied the review suggestions. The block’s difficulty now uses its actual parent and timestamp, and its coinbase pays the subsidy for its resulting height. I also added focused tests for the target and coinbase fields through the first halving.

  10. DrahtBot removed the label Needs rebase on Sep 25, 2026
  11. l0rinc force-pushed on Oct 1, 2026
  12. l0rinc renamed this:
    test: harden arbitrary-parent block creation
    test: fix coinbases built on arbitrary parents
    on Oct 4, 2026
  13. sedited requested review from jeanpablojp on Nov 12, 2026
  14. sedited removed review request from jeanpablojp on Nov 12, 2026
  15. sedited requested review from Sjors on Nov 12, 2026
  16. sedited requested review from jeanpablojp on Nov 12, 2026
  17. in src/test/util/mining.h:28 in 4e5aeb3535
      23 | @@ -23,6 +24,9 @@ struct NodeContext;
      24 |  /** Create a blockchain, starting from genesis */
      25 |  std::vector<std::shared_ptr<CBlock>> CreateBlockChain(size_t total_height, const CChainParams& params);
      26 |  
      27 | +/** Rebuild a block template for a different parent and timestamp */
      28 | +void RebuildBlockForParent(CBlock& block, const CBlockIndex& parent, uint32_t time);
    


    Sjors commented at 10:33 AM on November 12, 2026:

    In 4e5aeb3535106911f8ca4e0015e70a5ea9c74330 refactor: extract fork block reconstruction nit: SwitchBlockParent(..., uint32_t new_time) might be a better name. The caller's goal is to switch the parent and change the time. The description could then say that the block is rebuilt, and enumerate the side effects (in later commits), like modified subsidy.


    l0rinc commented at 2:02 PM on November 12, 2026:

    Renamed, added description, thanks

  18. in src/test/validation_block_tests.cpp:398 in 4e5aeb3535
     393 | +    const auto parent_hash{genesis.GetHash()};
     394 | +    CBlockIndex parent{genesis};
     395 | +    parent.phashBlock = &parent_hash;
     396 | +    --parent.nBits;
     397 | +
     398 | +    // Regtest uses the parent's target until the min-difficulty delay is exceeded.
    


    Sjors commented at 10:54 AM on November 12, 2026:

    In 4e5aeb3535106911f8ca4e0015e70a5ea9c74330 refactor: extract fork block reconstruction:

    This relies on the 20-minute difficulty reset rule, which only exists on testnet3 and testnet4. Regtest doesn't have difficulty adjustment at all, but technically the same 20-minute reset rule exists. The test exploits this with a synthetic block with an nBits that's impossible in a "real" regtest chain.

    In other words: please document this.

    I'm also worried that it's a bit brittle to rely on, since we could remove the 20-minute reset rule on regtest, but I suppose we're never going to touch that consensus code.


    l0rinc commented at 2:02 PM on November 12, 2026:

    Documented it, thanks

  19. in src/test/util/mining.cpp:81 in 9312297e15 outdated
      77 | @@ -78,12 +78,15 @@ std::vector<std::shared_ptr<CBlock>> CreateBlockChain(size_t total_height, const
      78 |  
      79 |  void RebuildBlockForParent(CBlock& block, const CBlockIndex& parent, uint32_t time, const Consensus::Params& params)
      80 |  {
      81 | +    Assert(block.vtx.size() == 1);
    


    Sjors commented at 10:59 AM on November 12, 2026:

    In 9312297e159c6b3aa5ff0526ef104a4006f8fe5c test: derive fork subsidy from parent: let's document the reason for this restriction, so anyone who needs to relax it will know what to pay attention to.


    l0rinc commented at 2:03 PM on November 12, 2026:

    Done, thanks

  20. in src/test/validation_block_tests.cpp:93 in 9312297e15
      90 |      BOOST_REQUIRE(block_template);
      91 |      auto pblock = std::make_shared<CBlock>(block_template->block);
      92 | -    pblock->hashPrevBlock = prev_hash;
      93 | -    pblock->nTime = ++time;
      94 | +    const auto& parent{*Assert(WITH_LOCK(::cs_main, return m_node.chainman->m_blockman.LookupBlockIndex(prev_hash)))};
      95 | +    RebuildBlockForParent(*pblock, parent, /*time=*/++time, Params().GetConsensus());
    


    Sjors commented at 11:05 AM on November 12, 2026:

    In 9312297e159c6b3aa5ff0526ef104a4006f8fe5c test: derive fork subsidy from parent: this might be a good time to give MinerTestingSetup::Block a quick doc:

    /**
     * Prepare a coinbase-only block extending prev_hash, which need not be the active tip.
     * The coinbase has a unique zero-value output and a reward output spendable with OP_TRUE.
     * Call FinalizeBlock() after any further edits to update commitments, mine the block,
     * and submit its header.
     */
    

    l0rinc commented at 2:03 PM on November 12, 2026:

    Done

  21. Sjors commented at 11:16 AM on November 12, 2026: member

    Concept ACK

    Code review 23a99695efa659ee0a0a290d4e0aa78aff8ddef9, mostly documentation feedback.

    You may find the BlockAssembler::CreateCoinbaseTx helper in #36109 commit 902d836b75d402945ac60936f73ded056347fb57 useful. But I don't mind rebasing that PR.

  22. l0rinc force-pushed on Nov 12, 2026
  23. Sjors commented at 2:47 PM on November 12, 2026: member

    ACK 27240dce14dc0ddde852cf191bba5a230713f1e5

  24. jeanpablojp commented at 10:34 PM on November 17, 2026: contributor

    tACK 27240dce14dc0ddde852cf191bba5a230713f1e5

  25. in src/test/validation_block_tests.cpp:390 in 9fdac77721
     384 | @@ -384,4 +385,20 @@ BOOST_AUTO_TEST_CASE(witness_commitment_index)
     385 |  
     386 |      BOOST_CHECK_EQUAL(GetWitnessCommitmentIndex(pblock), 2);
     387 |  }
     388 | +BOOST_FIXTURE_TEST_CASE(switch_block_parent_coinbase, RegTestingSetup)
     389 | +{
     390 | +    const auto& consensus{Params().GetConsensus()};
    


    maflcko commented at 8:09 AM on November 18, 2026:

    nit in the first commit: Please try to avoid the globals. There is a chainman.GetConsensus()

  26. in src/test/util/mining.h:30 in 7e784f41a0
      23 | @@ -23,6 +24,12 @@ struct NodeContext;
      24 |  /** Create a blockchain, starting from genesis */
      25 |  std::vector<std::shared_ptr<CBlock>> CreateBlockChain(size_t total_height, const CChainParams& params);
      26 |  
      27 | +/**
      28 | + * Switch a block template to a different parent and timestamp.
      29 | + * Rebuild the coinbase height, locktime and input script, then update the merkle root.
      30 | + */
    


    maflcko commented at 8:11 AM on November 18, 2026:

    nit: Use the shorter /// doxygen for new comments? The /** is inconsistent with the one above and below.


    fanquake commented at 10:03 AM on November 18, 2026:

    Could also just drop the declaration and make the definition static, if the intention is not for this to be re-used elsewhere.


    l0rinc commented at 12:40 PM on November 18, 2026:

    Thanks, switched to ///. The helper is already used from another test file.

  27. in src/test/validation_block_tests.cpp:391 in 7e784f41a0
     384 | @@ -385,6 +385,26 @@ BOOST_AUTO_TEST_CASE(witness_commitment_index)
     385 |  
     386 |      BOOST_CHECK_EQUAL(GetWitnessCommitmentIndex(pblock), 2);
     387 |  }
     388 | +
     389 | +BOOST_FIXTURE_TEST_CASE(switch_block_parent_target, RegTestingSetup)
     390 | +{
     391 | +    const auto& consensus{Params().GetConsensus()};
    


    maflcko commented at 8:13 AM on November 18, 2026:

    same here wrt globals: c.f. fa1b76aeb064b315a3767a8f59836ca18aeb117e

  28. in src/test/validation_block_tests.cpp:99 in 7bb37739db
      96 |      BOOST_REQUIRE(block_template);
      97 |      auto pblock = std::make_shared<CBlock>(block_template->block);
      98 | -    pblock->hashPrevBlock = prev_hash;
      99 | -    pblock->nTime = ++time;
     100 | +    const auto& parent{*Assert(WITH_LOCK(::cs_main, return m_node.chainman->m_blockman.LookupBlockIndex(prev_hash)))};
     101 | +    SwitchBlockParent(*pblock, parent, /*new_time=*/++time, Params().GetConsensus());
    


    maflcko commented at 8:35 AM on November 18, 2026:

    second-to-last commit nit: the global Params() would be good to avoid

  29. in src/test/util/mining.cpp:92 in 27240dce14
      88 | @@ -89,7 +89,8 @@ void SwitchBlockParent(CBlock& block, const CBlockIndex& parent, uint32_t new_ti
      89 |      const int height{parent.nHeight + 1};
      90 |      CMutableTransaction tx_coinbase{*block.vtx.at(0)};
      91 |      tx_coinbase.nLockTime = static_cast<uint32_t>(parent.nHeight);
      92 | -    tx_coinbase.vin.at(0).scriptSig = CScript{} << height;
      93 | +    // Include a dummy OP_0 so low-height coinbase input scripts meet the minimum length
    


    maflcko commented at 8:45 AM on November 18, 2026:

    nit in the last commit: Any reason to reword the comment here? The prior comment was preferable, because it mentioned that in reality this is usually some extranonce space. Also, it had the exact error msg ID, so that a simple git grep can be used.

  30. in src/test/validation_block_tests.cpp:44 in 7bb37739db
      39 | @@ -40,6 +40,12 @@ using kernel::ChainstateRole;
      40 |  
      41 |  namespace validation_block_tests {
      42 |  struct MinerTestingSetup : public RegTestingSetup {
      43 | +    /**
      44 | +     * Prepare a coinbase-only block extending prev_hash, which need not be the active tip.
    


    maflcko commented at 8:51 AM on November 18, 2026:

    nit in 7bb37739db34cba73b375a4a47aaebbeb5f49387: It seems a bit confusing to phrase this what it doesn't need to be, and then have the reader ask if it must not be the active tip and what else it can be. It would be better and shorter to just say what it can be: ..., which may refer to any known prior header. Alternatively, the comment suffix can be removed, because accepting a prev_hash should already clarify that the dev can pass any prev hash here.

    Also, use /// doxygen comment?

  31. maflcko approved
  32. maflcko commented at 8:57 AM on November 18, 2026: member

    Seems rare that a unit test needs the helpers to build such a chain, but this seems fine as a style-cleanup.

    review ACK 27240dce14dc0ddde852cf191bba5a230713f1e5 🐁

    <details><summary>Show signature</summary>

    Signature:

    untrusted comment: signature from minisign secret key on empty file; verify via: minisign -Vm "${path_to_any_empty_file}" -P RWTRmVTMeKV5noAMqVlsMugDDCyyTSbA3Re5AkUrhvLVln0tSaFWglOw -x "${path_to_this_whole_four_line_signature_blob}"
    RUTRmVTMeKV5npGrKx1nqXCw5zeVHdtdYURB/KlyA/LMFgpNCs+SkW9a8N95d+U4AP1RJMi+krxU1A3Yux4bpwZNLvVBKy0wLgM=
    trusted comment: review ACK 27240dce14dc0ddde852cf191bba5a230713f1e5 🐁
    rDyUIRZ7S4ZFIQQOJ8CcWLslusVlHaANsiJzQulEb54zKZ9oiLbdXvrmlGHtXYoZ5VFP+d7R5B6McN2F0x41Cw==
    

    </details>

  33. l0rinc force-pushed on Nov 18, 2026
  34. in src/test/validation_block_tests.cpp:105 in 5c4209eecb outdated
     106 | @@ -101,9 +107,7 @@ std::shared_ptr<CBlock> MinerTestingSetup::Block(const uint256& prev_hash)
     107 |      txCoinbase.vout[0].nValue = 0;
     108 |      txCoinbase.vin[0].scriptWitness.SetNull();
     109 |      // Always pad with OP_0 as dummy extraNonce (also avoids bad-cb-length error for block <=16)
     110 | -    const int prev_height{WITH_LOCK(::cs_main, return m_node.chainman->m_blockman.LookupBlockIndex(prev_hash)->nHeight)};
     111 | -    txCoinbase.vin[0].scriptSig = CScript{} << prev_height + 1 << OP_0;
    


    maflcko commented at 12:14 PM on November 18, 2026:

    5c4209eecb19c5031be7a2ff8ddc99f2f113f367: Two line hunk in the wrong commit?

  35. in src/test/validation_block_tests.cpp:390 in 096875f8f3
     383 | @@ -382,4 +384,21 @@ BOOST_AUTO_TEST_CASE(witness_commitment_index)
     384 |  
     385 |      BOOST_CHECK_EQUAL(GetWitnessCommitmentIndex(pblock), 2);
     386 |  }
     387 | +
     388 | +BOOST_FIXTURE_TEST_CASE(switch_block_parent_coinbase, RegTestingSetup)
     389 | +{
     390 | +    const auto& consensus{m_node.chainman->GetConsensus()};
    


    maflcko commented at 12:15 PM on November 18, 2026:

    nit in 096875f8f31360ddb805af1442c4e63809a46caa: Add and use a chainman reference?

  36. maflcko approved
  37. maflcko commented at 12:16 PM on November 18, 2026: member

    lgtm e07fdd2a11 😼

    <details><summary>Show signature</summary>

    Signature:

    untrusted comment: signature from minisign secret key on empty file; verify via: minisign -Vm "${path_to_any_empty_file}" -P RWTRmVTMeKV5noAMqVlsMugDDCyyTSbA3Re5AkUrhvLVln0tSaFWglOw -x "${path_to_this_whole_four_line_signature_blob}"
    RUTRmVTMeKV5npGrKx1nqXCw5zeVHdtdYURB/KlyA/LMFgpNCs+SkW9a8N95d+U4AP1RJMi+krxU1A3Yux4bpwZNLvVBKy0wLgM=
    trusted comment: lgtm e07fdd2a11 😼
    h6tOltp2lhIhuW7iBerbqk6IjIUSRFjNcTTTJff+T1iLy+0hFplznURJcP9M4vqX5aX2XyCQvoOCuzzJmw5sBw==
    

    </details>

  38. test: characterize fork coinbase fields
    Record the subsidy and input-script length of a chain built from genesis through the first halving.
    The active-tip template reward is retained at the halving, and the unpadded low-height coinbases prevent the fork from connecting.
    
    Co-authored-by: JP <jeanpablo.jp@hotmail.com>
    bf551335be
  39. refactor: extract fork block reconstruction
    Move the existing parent-dependent field rewrites out of `BuildChain()` without changing the generated fields.
    A direct helper test records the copied target at and just above regtest's min-difficulty threshold using a synthetic parent with a different target.
    
    Co-authored-by: JP <jeanpablo.jp@hotmail.com>
    4745be9837
  40. test: derive fork difficulty from parent
    Recalculate `nBits` from the actual parent and child timestamp instead of retaining the active-tip template's target.
    Existing chain-building callers use regtest's minimum difficulty, so their generated targets are unchanged.
    33bbf91e1a
  41. test: derive fork subsidy from parent
    Derive the coinbase subsidy from the actual parent height and share the reconstruction with `MinerTestingSetup::Block()`.
    Existing reorg tests cross halvings and still pass with underpaid coinbases.
    A fork extending across a halving can overpay and fail with `bad-cb-amount`.
    
    Require coinbase-only templates so replacing the reward cannot discard transaction fees.
    5cdbfd5f01
  42. test: pad fork coinbase input scripts
    Append a dummy `OP_0` to rebuilt coinbase input scripts so heights 1 through 16 meet the minimum length.
    The fork from genesis can then connect through the first halving.
    Remove the duplicate padding from `MinerTestingSetup::Block()`.
    b7b762880e
  43. l0rinc force-pushed on Nov 18, 2026
  44. maflcko commented at 12:41 PM on November 18, 2026: member

    review ACK b7b762880e7583a06429f9eca525670a2ee18e5e 🐡

    <details><summary>Show signature</summary>

    Signature:

    untrusted comment: signature from minisign secret key on empty file; verify via: minisign -Vm "${path_to_any_empty_file}" -P RWTRmVTMeKV5noAMqVlsMugDDCyyTSbA3Re5AkUrhvLVln0tSaFWglOw -x "${path_to_this_whole_four_line_signature_blob}"
    RUTRmVTMeKV5npGrKx1nqXCw5zeVHdtdYURB/KlyA/LMFgpNCs+SkW9a8N95d+U4AP1RJMi+krxU1A3Yux4bpwZNLvVBKy0wLgM=
    trusted comment: review ACK b7b762880e7583a06429f9eca525670a2ee18e5e 🐡
    0Zw3K7cJBpNcydZZLKGB1255qgUAAZQATeTIqqLZHAqCJZMnAN7TZ70oV8LUSXKjz/KBfmwumCD2l4P+NjdECg==
    

    </details>

  45. DrahtBot requested review from Sjors on Nov 18, 2026
  46. jeanpablojp commented at 2:01 PM on November 18, 2026: contributor

    ACK b7b762880e7583a06429f9eca525670a2ee18e5e


github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin/bitcoin. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2026-11-18 16:51 UTC

This site is hosted by @0xB10C
More mirrored repositories can be found on mirror.b10c.me