aboutsummaryrefslogtreecommitdiff
path: root/src/script/standard.cpp
blob: 61017380612cc496d4518f68bfd597eff42086d8 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
// Copyright (c) 2009-2010 Satoshi Nakamoto
// Copyright (c) 2009-2021 The Bitcoin Core developers
// Distributed under the MIT software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.

#include <script/standard.h>

#include <crypto/sha256.h>
#include <hash.h>
#include <pubkey.h>
#include <script/interpreter.h>
#include <script/script.h>
#include <util/strencodings.h>

#include <string>

typedef std::vector<unsigned char> valtype;

CScriptID::CScriptID(const CScript& in) : BaseHash(Hash160(in)) {}
CScriptID::CScriptID(const ScriptHash& in) : BaseHash(static_cast<uint160>(in)) {}

ScriptHash::ScriptHash(const CScript& in) : BaseHash(Hash160(in)) {}
ScriptHash::ScriptHash(const CScriptID& in) : BaseHash(static_cast<uint160>(in)) {}

PKHash::PKHash(const CPubKey& pubkey) : BaseHash(pubkey.GetID()) {}
PKHash::PKHash(const CKeyID& pubkey_id) : BaseHash(pubkey_id) {}

WitnessV0KeyHash::WitnessV0KeyHash(const CPubKey& pubkey) : BaseHash(pubkey.GetID()) {}
WitnessV0KeyHash::WitnessV0KeyHash(const PKHash& pubkey_hash) : BaseHash(static_cast<uint160>(pubkey_hash)) {}

CKeyID ToKeyID(const PKHash& key_hash)
{
    return CKeyID{static_cast<uint160>(key_hash)};
}

CKeyID ToKeyID(const WitnessV0KeyHash& key_hash)
{
    return CKeyID{static_cast<uint160>(key_hash)};
}

WitnessV0ScriptHash::WitnessV0ScriptHash(const CScript& in)
{
    CSHA256().Write(in.data(), in.size()).Finalize(begin());
}

std::string GetTxnOutputType(TxoutType t)
{
    switch (t) {
    case TxoutType::NONSTANDARD: return "nonstandard";
    case TxoutType::PUBKEY: return "pubkey";
    case TxoutType::PUBKEYHASH: return "pubkeyhash";
    case TxoutType::SCRIPTHASH: return "scripthash";
    case TxoutType::MULTISIG: return "multisig";
    case TxoutType::NULL_DATA: return "nulldata";
    case TxoutType::WITNESS_V0_KEYHASH: return "witness_v0_keyhash";
    case TxoutType::WITNESS_V0_SCRIPTHASH: return "witness_v0_scripthash";
    case TxoutType::WITNESS_V1_TAPROOT: return "witness_v1_taproot";
    case TxoutType::WITNESS_UNKNOWN: return "witness_unknown";
    } // no default case, so the compiler can warn about missing cases
    assert(false);
}

static bool MatchPayToPubkey(const CScript& script, valtype& pubkey)
{
    if (script.size() == CPubKey::SIZE + 2 && script[0] == CPubKey::SIZE && script.back() == OP_CHECKSIG) {
        pubkey = valtype(script.begin() + 1, script.begin() + CPubKey::SIZE + 1);
        return CPubKey::ValidSize(pubkey);
    }
    if (script.size() == CPubKey::COMPRESSED_SIZE + 2 && script[0] == CPubKey::COMPRESSED_SIZE && script.back() == OP_CHECKSIG) {
        pubkey = valtype(script.begin() + 1, script.begin() + CPubKey::COMPRESSED_SIZE + 1);
        return CPubKey::ValidSize(pubkey);
    }
    return false;
}

static bool MatchPayToPubkeyHash(const CScript& script, valtype& pubkeyhash)
{
    if (script.size() == 25 && script[0] == OP_DUP && script[1] == OP_HASH160 && script[2] == 20 && script[23] == OP_EQUALVERIFY && script[24] == OP_CHECKSIG) {
        pubkeyhash = valtype(script.begin () + 3, script.begin() + 23);
        return true;
    }
    return false;
}

/** Test for "small positive integer" script opcodes - OP_1 through OP_16. */
static constexpr bool IsSmallInteger(opcodetype opcode)
{
    return opcode >= OP_1 && opcode <= OP_16;
}

/** Retrieve a minimally-encoded number in range [min,max] from an (opcode, data) pair,
 *  whether it's OP_n or through a push. */
static std::optional<int> GetScriptNumber(opcodetype opcode, valtype data, int min, int max)
{
    int count;
    if (IsSmallInteger(opcode)) {
        count = CScript::DecodeOP_N(opcode);
    } else if (IsPushdataOp(opcode)) {
        if (!CheckMinimalPush(data, opcode)) return {};
        try {
            count = CScriptNum(data, /* fRequireMinimal = */ true).getint();
        } catch (const scriptnum_error&) {
            return {};
        }
    } else {
        return {};
    }
    if (count < min || count > max) return {};
    return count;
}

static bool MatchMultisig(const CScript& script, int& required_sigs, std::vector<valtype>& pubkeys)
{
    opcodetype opcode;
    valtype data;

    CScript::const_iterator it = script.begin();
    if (script.size() < 1 || script.back() != OP_CHECKMULTISIG) return false;

    if (!script.GetOp(it, opcode, data)) return false;
    auto req_sigs = GetScriptNumber(opcode, data, 1, MAX_PUBKEYS_PER_MULTISIG);
    if (!req_sigs) return false;
    required_sigs = *req_sigs;
    while (script.GetOp(it, opcode, data) && CPubKey::ValidSize(data)) {
        pubkeys.emplace_back(std::move(data));
    }
    auto num_keys = GetScriptNumber(opcode, data, required_sigs, MAX_PUBKEYS_PER_MULTISIG);
    if (!num_keys) return false;
    if (pubkeys.size() != static_cast<unsigned long>(*num_keys)) return false;

    return (it + 1 == script.end());
}

std::optional<std::pair<int, std::vector<Span<const unsigned char>>>> MatchMultiA(const CScript& script)
{
    std::vector<Span<const unsigned char>> keyspans;

    // Redundant, but very fast and selective test.
    if (script.size() == 0 || script[0] != 32 || script.back() != OP_NUMEQUAL) return {};

    // Parse keys
    auto it = script.begin();
    while (script.end() - it >= 34) {
        if (*it != 32) return {};
        ++it;
        keyspans.emplace_back(&*it, 32);
        it += 32;
        if (*it != (keyspans.size() == 1 ? OP_CHECKSIG : OP_CHECKSIGADD)) return {};
        ++it;
    }
    if (keyspans.size() == 0 || keyspans.size() > MAX_PUBKEYS_PER_MULTI_A) return {};

    // Parse threshold.
    opcodetype opcode;
    std::vector<unsigned char> data;
    if (!script.GetOp(it, opcode, data)) return {};
    if (it == script.end()) return {};
    if (*it != OP_NUMEQUAL) return {};
    ++it;
    if (it != script.end()) return {};
    auto threshold = GetScriptNumber(opcode, data, 1, (int)keyspans.size());
    if (!threshold) return {};

    // Construct result.
    return std::pair{*threshold, std::move(keyspans)};
}

TxoutType Solver(const CScript& scriptPubKey, std::vector<std::vector<unsigned char>>& vSolutionsRet)
{
    vSolutionsRet.clear();

    // Shortcut for pay-to-script-hash, which are more constrained than the other types:
    // it is always OP_HASH160 20 [20 byte hash] OP_EQUAL
    if (scriptPubKey.IsPayToScriptHash())
    {
        std::vector<unsigned char> hashBytes(scriptPubKey.begin()+2, scriptPubKey.begin()+22);
        vSolutionsRet.push_back(hashBytes);
        return TxoutType::SCRIPTHASH;
    }

    int witnessversion;
    std::vector<unsigned char> witnessprogram;
    if (scriptPubKey.IsWitnessProgram(witnessversion, witnessprogram)) {
        if (witnessversion == 0 && witnessprogram.size() == WITNESS_V0_KEYHASH_SIZE) {
            vSolutionsRet.push_back(std::move(witnessprogram));
            return TxoutType::WITNESS_V0_KEYHASH;
        }
        if (witnessversion == 0 && witnessprogram.size() == WITNESS_V0_SCRIPTHASH_SIZE) {
            vSolutionsRet.push_back(std::move(witnessprogram));
            return TxoutType::WITNESS_V0_SCRIPTHASH;
        }
        if (witnessversion == 1 && witnessprogram.size() == WITNESS_V1_TAPROOT_SIZE) {
            vSolutionsRet.push_back(std::move(witnessprogram));
            return TxoutType::WITNESS_V1_TAPROOT;
        }
        if (witnessversion != 0) {
            vSolutionsRet.push_back(std::vector<unsigned char>{(unsigned char)witnessversion});
            vSolutionsRet.push_back(std::move(witnessprogram));
            return TxoutType::WITNESS_UNKNOWN;
        }
        return TxoutType::NONSTANDARD;
    }

    // Provably prunable, data-carrying output
    //
    // So long as script passes the IsUnspendable() test and all but the first
    // byte passes the IsPushOnly() test we don't care what exactly is in the
    // script.
    if (scriptPubKey.size() >= 1 && scriptPubKey[0] == OP_RETURN && scriptPubKey.IsPushOnly(scriptPubKey.begin()+1)) {
        return TxoutType::NULL_DATA;
    }

    std::vector<unsigned char> data;
    if (MatchPayToPubkey(scriptPubKey, data)) {
        vSolutionsRet.push_back(std::move(data));
        return TxoutType::PUBKEY;
    }

    if (MatchPayToPubkeyHash(scriptPubKey, data)) {
        vSolutionsRet.push_back(std::move(data));
        return TxoutType::PUBKEYHASH;
    }

    int required;
    std::vector<std::vector<unsigned char>> keys;
    if (MatchMultisig(scriptPubKey, required, keys)) {
        vSolutionsRet.push_back({static_cast<unsigned char>(required)}); // safe as required is in range 1..20
        vSolutionsRet.insert(vSolutionsRet.end(), keys.begin(), keys.end());
        vSolutionsRet.push_back({static_cast<unsigned char>(keys.size())}); // safe as size is in range 1..20
        return TxoutType::MULTISIG;
    }

    vSolutionsRet.clear();
    return TxoutType::NONSTANDARD;
}

bool ExtractDestination(const CScript& scriptPubKey, CTxDestination& addressRet)
{
    std::vector<valtype> vSolutions;
    TxoutType whichType = Solver(scriptPubKey, vSolutions);

    switch (whichType) {
    case TxoutType::PUBKEY: {
        CPubKey pubKey(vSolutions[0]);
        if (!pubKey.IsValid())
            return false;

        addressRet = PKHash(pubKey);
        return true;
    }
    case TxoutType::PUBKEYHASH: {
        addressRet = PKHash(uint160(vSolutions[0]));
        return true;
    }
    case TxoutType::SCRIPTHASH: {
        addressRet = ScriptHash(uint160(vSolutions[0]));
        return true;
    }
    case TxoutType::WITNESS_V0_KEYHASH: {
        WitnessV0KeyHash hash;
        std::copy(vSolutions[0].begin(), vSolutions[0].end(), hash.begin());
        addressRet = hash;
        return true;
    }
    case TxoutType::WITNESS_V0_SCRIPTHASH: {
        WitnessV0ScriptHash hash;
        std::copy(vSolutions[0].begin(), vSolutions[0].end(), hash.begin());
        addressRet = hash;
        return true;
    }
    case TxoutType::WITNESS_V1_TAPROOT: {
        WitnessV1Taproot tap;
        std::copy(vSolutions[0].begin(), vSolutions[0].end(), tap.begin());
        addressRet = tap;
        return true;
    }
    case TxoutType::WITNESS_UNKNOWN: {
        WitnessUnknown unk;
        unk.version = vSolutions[0][0];
        std::copy(vSolutions[1].begin(), vSolutions[1].end(), unk.program);
        unk.length = vSolutions[1].size();
        addressRet = unk;
        return true;
    }
    case TxoutType::MULTISIG:
    case TxoutType::NULL_DATA:
    case TxoutType::NONSTANDARD:
        return false;
    } // no default case, so the compiler can warn about missing cases
    assert(false);
}

namespace {
class CScriptVisitor
{
public:
    CScript operator()(const CNoDestination& dest) const
    {
        return CScript();
    }

    CScript operator()(const PKHash& keyID) const
    {
        return CScript() << OP_DUP << OP_HASH160 << ToByteVector(keyID) << OP_EQUALVERIFY << OP_CHECKSIG;
    }

    CScript operator()(const ScriptHash& scriptID) const
    {
        return CScript() << OP_HASH160 << ToByteVector(scriptID) << OP_EQUAL;
    }

    CScript operator()(const WitnessV0KeyHash& id) const
    {
        return CScript() << OP_0 << ToByteVector(id);
    }

    CScript operator()(const WitnessV0ScriptHash& id) const
    {
        return CScript() << OP_0 << ToByteVector(id);
    }

    CScript operator()(const WitnessV1Taproot& tap) const
    {
        return CScript() << OP_1 << ToByteVector(tap);
    }

    CScript operator()(const WitnessUnknown& id) const
    {
        return CScript() << CScript::EncodeOP_N(id.version) << std::vector<unsigned char>(id.program, id.program + id.length);
    }
};
} // namespace

CScript GetScriptForDestination(const CTxDestination& dest)
{
    return std::visit(CScriptVisitor(), dest);
}

CScript GetScriptForRawPubKey(const CPubKey& pubKey)
{
    return CScript() << std::vector<unsigned char>(pubKey.begin(), pubKey.end()) << OP_CHECKSIG;
}

CScript GetScriptForMultisig(int nRequired, const std::vector<CPubKey>& keys)
{
    CScript script;

    script << nRequired;
    for (const CPubKey& key : keys)
        script << ToByteVector(key);
    script << keys.size() << OP_CHECKMULTISIG;

    return script;
}

bool IsValidDestination(const CTxDestination& dest) {
    return dest.index() != 0;
}

/*static*/ TaprootBuilder::NodeInfo TaprootBuilder::Combine(NodeInfo&& a, NodeInfo&& b)
{
    NodeInfo ret;
    /* Iterate over all tracked leaves in a, add b's hash to their Merkle branch, and move them to ret. */
    for (auto& leaf : a.leaves) {
        leaf.merkle_branch.push_back(b.hash);
        ret.leaves.emplace_back(std::move(leaf));
    }
    /* Iterate over all tracked leaves in b, add a's hash to their Merkle branch, and move them to ret. */
    for (auto& leaf : b.leaves) {
        leaf.merkle_branch.push_back(a.hash);
        ret.leaves.emplace_back(std::move(leaf));
    }
    /* Lexicographically sort a and b's hash, and compute parent hash. */
    if (a.hash < b.hash) {
        ret.hash = (HashWriter{HASHER_TAPBRANCH} << a.hash << b.hash).GetSHA256();
    } else {
        ret.hash = (HashWriter{HASHER_TAPBRANCH} << b.hash << a.hash).GetSHA256();
    }
    return ret;
}

void TaprootSpendData::Merge(TaprootSpendData other)
{
    // TODO: figure out how to better deal with conflicting information
    // being merged.
    if (internal_key.IsNull() && !other.internal_key.IsNull()) {
        internal_key = other.internal_key;
    }
    if (merkle_root.IsNull() && !other.merkle_root.IsNull()) {
        merkle_root = other.merkle_root;
    }
    for (auto& [key, control_blocks] : other.scripts) {
        scripts[key].merge(std::move(control_blocks));
    }
}

void TaprootBuilder::Insert(TaprootBuilder::NodeInfo&& node, int depth)
{
    assert(depth >= 0 && (size_t)depth <= TAPROOT_CONTROL_MAX_NODE_COUNT);
    /* We cannot insert a leaf at a lower depth while a deeper branch is unfinished. Doing
     * so would mean the Add() invocations do not correspond to a DFS traversal of a
     * binary tree. */
    if ((size_t)depth + 1 < m_branch.size()) {
        m_valid = false;
        return;
    }
    /* As long as an entry in the branch exists at the specified depth, combine it and propagate up.
     * The 'node' variable is overwritten here with the newly combined node. */
    while (m_valid && m_branch.size() > (size_t)depth && m_branch[depth].has_value()) {
        node = Combine(std::move(node), std::move(*m_branch[depth]));
        m_branch.pop_back();
        if (depth == 0) m_valid = false; /* Can't propagate further up than the root */
        --depth;
    }
    if (m_valid) {
        /* Make sure the branch is big enough to place the new node. */
        if (m_branch.size() <= (size_t)depth) m_branch.resize((size_t)depth + 1);
        assert(!m_branch[depth].has_value());
        m_branch[depth] = std::move(node);
    }
}

/*static*/ bool TaprootBuilder::ValidDepths(const std::vector<int>& depths)
{
    std::vector<bool> branch;
    for (int depth : depths) {
        // This inner loop corresponds to effectively the same logic on branch
        // as what Insert() performs on the m_branch variable. Instead of
        // storing a NodeInfo object, just remember whether or not there is one
        // at that depth.
        if (depth < 0 || (size_t)depth > TAPROOT_CONTROL_MAX_NODE_COUNT) return false;
        if ((size_t)depth + 1 < branch.size()) return false;
        while (branch.size() > (size_t)depth && branch[depth]) {
            branch.pop_back();
            if (depth == 0) return false;
            --depth;
        }
        if (branch.size() <= (size_t)depth) branch.resize((size_t)depth + 1);
        assert(!branch[depth]);
        branch[depth] = true;
    }
    // And this check corresponds to the IsComplete() check on m_branch.
    return branch.size() == 0 || (branch.size() == 1 && branch[0]);
}

TaprootBuilder& TaprootBuilder::Add(int depth, const CScript& script, int leaf_version, bool track)
{
    assert((leaf_version & ~TAPROOT_LEAF_MASK) == 0);
    if (!IsValid()) return *this;
    /* Construct NodeInfo object with leaf hash and (if track is true) also leaf information. */
    NodeInfo node;
    node.hash = (HashWriter{HASHER_TAPLEAF} << uint8_t(leaf_version) << script).GetSHA256();
    if (track) node.leaves.emplace_back(LeafInfo{script, leaf_version, {}});
    /* Insert into the branch. */
    Insert(std::move(node), depth);
    return *this;
}

TaprootBuilder& TaprootBuilder::AddOmitted(int depth, const uint256& hash)
{
    if (!IsValid()) return *this;
    /* Construct NodeInfo object with the hash directly, and insert it into the branch. */
    NodeInfo node;
    node.hash = hash;
    Insert(std::move(node), depth);
    return *this;
}

TaprootBuilder& TaprootBuilder::Finalize(const XOnlyPubKey& internal_key)
{
    /* Can only call this function when IsComplete() is true. */
    assert(IsComplete());
    m_internal_key = internal_key;
    auto ret = m_internal_key.CreateTapTweak(m_branch.size() == 0 ? nullptr : &m_branch[0]->hash);
    assert(ret.has_value());
    std::tie(m_output_key, m_parity) = *ret;
    return *this;
}

WitnessV1Taproot TaprootBuilder::GetOutput() { return WitnessV1Taproot{m_output_key}; }

TaprootSpendData TaprootBuilder::GetSpendData() const
{
    assert(IsComplete());
    assert(m_output_key.IsFullyValid());
    TaprootSpendData spd;
    spd.merkle_root = m_branch.size() == 0 ? uint256() : m_branch[0]->hash;
    spd.internal_key = m_internal_key;
    if (m_branch.size()) {
        // If any script paths exist, they have been combined into the root m_branch[0]
        // by now. Compute the control block for each of its tracked leaves, and put them in
        // spd.scripts.
        for (const auto& leaf : m_branch[0]->leaves) {
            std::vector<unsigned char> control_block;
            control_block.resize(TAPROOT_CONTROL_BASE_SIZE + TAPROOT_CONTROL_NODE_SIZE * leaf.merkle_branch.size());
            control_block[0] = leaf.leaf_version | (m_parity ? 1 : 0);
            std::copy(m_internal_key.begin(), m_internal_key.end(), control_block.begin() + 1);
            if (leaf.merkle_branch.size()) {
                std::copy(leaf.merkle_branch[0].begin(),
                          leaf.merkle_branch[0].begin() + TAPROOT_CONTROL_NODE_SIZE * leaf.merkle_branch.size(),
                          control_block.begin() + TAPROOT_CONTROL_BASE_SIZE);
            }
            spd.scripts[{leaf.script, leaf.leaf_version}].insert(std::move(control_block));
        }
    }
    return spd;
}

std::optional<std::vector<std::tuple<int, CScript, int>>> InferTaprootTree(const TaprootSpendData& spenddata, const XOnlyPubKey& output)
{
    // Verify that the output matches the assumed Merkle root and internal key.
    auto tweak = spenddata.internal_key.CreateTapTweak(spenddata.merkle_root.IsNull() ? nullptr : &spenddata.merkle_root);
    if (!tweak || tweak->first != output) return std::nullopt;
    // If the Merkle root is 0, the tree is empty, and we're done.
    std::vector<std::tuple<int, CScript, int>> ret;
    if (spenddata.merkle_root.IsNull()) return ret;

    /** Data structure to represent the nodes of the tree we're going to build. */
    struct TreeNode {
        /** Hash of this node, if known; 0 otherwise. */
        uint256 hash;
        /** The left and right subtrees (note that their order is irrelevant). */
        std::unique_ptr<TreeNode> sub[2];
        /** If this is known to be a leaf node, a pointer to the (script, leaf_ver) pair.
         *  nullptr otherwise. */
        const std::pair<CScript, int>* leaf = nullptr;
        /** Whether or not this node has been explored (is known to be a leaf, or known to have children). */
        bool explored = false;
        /** Whether or not this node is an inner node (unknown until explored = true). */
        bool inner;
        /** Whether or not we have produced output for this subtree. */
        bool done = false;
    };

    // Build tree from the provided branches.
    TreeNode root;
    root.hash = spenddata.merkle_root;
    for (const auto& [key, control_blocks] : spenddata.scripts) {
        const auto& [script, leaf_ver] = key;
        for (const auto& control : control_blocks) {
            // Skip script records with nonsensical leaf version.
            if (leaf_ver < 0 || leaf_ver >= 0x100 || leaf_ver & 1) continue;
            // Skip script records with invalid control block sizes.
            if (control.size() < TAPROOT_CONTROL_BASE_SIZE || control.size() > TAPROOT_CONTROL_MAX_SIZE ||
                ((control.size() - TAPROOT_CONTROL_BASE_SIZE) % TAPROOT_CONTROL_NODE_SIZE) != 0) continue;
            // Skip script records that don't match the control block.
            if ((control[0] & TAPROOT_LEAF_MASK) != leaf_ver) continue;
            // Skip script records that don't match the provided Merkle root.
            const uint256 leaf_hash = ComputeTapleafHash(leaf_ver, script);
            const uint256 merkle_root = ComputeTaprootMerkleRoot(control, leaf_hash);
            if (merkle_root != spenddata.merkle_root) continue;

            TreeNode* node = &root;
            size_t levels = (control.size() - TAPROOT_CONTROL_BASE_SIZE) / TAPROOT_CONTROL_NODE_SIZE;
            for (size_t depth = 0; depth < levels; ++depth) {
                // Can't descend into a node which we already know is a leaf.
                if (node->explored && !node->inner) return std::nullopt;

                // Extract partner hash from Merkle branch in control block.
                uint256 hash;
                std::copy(control.begin() + TAPROOT_CONTROL_BASE_SIZE + (levels - 1 - depth) * TAPROOT_CONTROL_NODE_SIZE,
                          control.begin() + TAPROOT_CONTROL_BASE_SIZE + (levels - depth) * TAPROOT_CONTROL_NODE_SIZE,
                          hash.begin());

                if (node->sub[0]) {
                    // Descend into the existing left or right branch.
                    bool desc = false;
                    for (int i = 0; i < 2; ++i) {
                        if (node->sub[i]->hash == hash || (node->sub[i]->hash.IsNull() && node->sub[1-i]->hash != hash)) {
                            node->sub[i]->hash = hash;
                            node = &*node->sub[1-i];
                            desc = true;
                            break;
                        }
                    }
                    if (!desc) return std::nullopt; // This probably requires a hash collision to hit.
                } else {
                    // We're in an unexplored node. Create subtrees and descend.
                    node->explored = true;
                    node->inner = true;
                    node->sub[0] = std::make_unique<TreeNode>();
                    node->sub[1] = std::make_unique<TreeNode>();
                    node->sub[1]->hash = hash;
                    node = &*node->sub[0];
                }
            }
            // Cannot turn a known inner node into a leaf.
            if (node->sub[0]) return std::nullopt;
            node->explored = true;
            node->inner = false;
            node->leaf = &key;
            node->hash = leaf_hash;
        }
    }

    // Recursive processing to turn the tree into flattened output. Use an explicit stack here to avoid
    // overflowing the call stack (the tree may be 128 levels deep).
    std::vector<TreeNode*> stack{&root};
    while (!stack.empty()) {
        TreeNode& node = *stack.back();
        if (!node.explored) {
            // Unexplored node, which means the tree is incomplete.
            return std::nullopt;
        } else if (!node.inner) {
            // Leaf node; produce output.
            ret.emplace_back(stack.size() - 1, node.leaf->first, node.leaf->second);
            node.done = true;
            stack.pop_back();
        } else if (node.sub[0]->done && !node.sub[1]->done && !node.sub[1]->explored && !node.sub[1]->hash.IsNull() &&
                   (HashWriter{HASHER_TAPBRANCH} << node.sub[1]->hash << node.sub[1]->hash).GetSHA256() == node.hash) {
            // Whenever there are nodes with two identical subtrees under it, we run into a problem:
            // the control blocks for the leaves underneath those will be identical as well, and thus
            // they will all be matched to the same path in the tree. The result is that at the location
            // where the duplicate occurred, the left child will contain a normal tree that can be explored
            // and processed, but the right one will remain unexplored.
            //
            // This situation can be detected, by encountering an inner node with unexplored right subtree
            // with known hash, and H_TapBranch(hash, hash) is equal to the parent node (this node)'s hash.
            //
            // To deal with this, simply process the left tree a second time (set its done flag to false;
            // noting that the done flag of its children have already been set to false after processing
            // those). To avoid ending up in an infinite loop, set the done flag of the right (unexplored)
            // subtree to true.
            node.sub[0]->done = false;
            node.sub[1]->done = true;
        } else if (node.sub[0]->done && node.sub[1]->done) {
            // An internal node which we're finished with.
            node.sub[0]->done = false;
            node.sub[1]->done = false;
            node.done = true;
            stack.pop_back();
        } else if (!node.sub[0]->done) {
            // An internal node whose left branch hasn't been processed yet. Do so first.
            stack.push_back(&*node.sub[0]);
        } else if (!node.sub[1]->done) {
            // An internal node whose right branch hasn't been processed yet. Do so first.
            stack.push_back(&*node.sub[1]);
        }
    }

    return ret;
}

std::vector<std::tuple<uint8_t, uint8_t, CScript>> TaprootBuilder::GetTreeTuples() const
{
    assert(IsComplete());
    std::vector<std::tuple<uint8_t, uint8_t, CScript>> tuples;
    if (m_branch.size()) {
        const auto& leaves = m_branch[0]->leaves;
        for (const auto& leaf : leaves) {
            assert(leaf.merkle_branch.size() <= TAPROOT_CONTROL_MAX_NODE_COUNT);
            uint8_t depth = (uint8_t)leaf.merkle_branch.size();
            uint8_t leaf_ver = (uint8_t)leaf.leaf_version;
            tuples.push_back(std::make_tuple(depth, leaf_ver, leaf.script));
        }
    }
    return tuples;
}