merkledb

package
v1.10.6-rc.6 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Aug 25, 2023 License: BSD-3-Clause Imports: 32 Imported by: 0

README

Path Based Merkelized Radix Trie

TODOs

  • Remove special casing around the root node from the physical structure of the hashed tree.
  • Analyze performance impact of needing to skip intermediate nodes when generating range and change proofs
    • Consider moving nodes with values to a separate db prefix
  • Analyze performance of using database snapshots rather than in-memory history
  • Improve intermediate node regeneration after ungraceful shutdown by reusing successfully written subtrees

Introduction

The Merkle Trie is a data structure that allows efficient and secure verification of the contents. It is a combination of a Merkle Tree and a Radix Trie.

The trie contains Merkle Nodes, which store key/value and children information.

Each Merkle Node represents a key path into the trie. It stores the key, the value (if one exists), its ID, and the IDs of its children nodes. The children have keys that contain the current node's key path as a prefix, and the index of each child indicates the next nibble in that child's key. For example, if we have two nodes, Node 1 with key path 0x91A and Node 2 with key path 0x91A4, Node 2 is stored in index 0x4 of Node 1's children (since 0x4 is the first value after the common prefix).

To reduce the depth of nodes in the trie, a Merkle Node utilizes path compression. Instead of having a long chain of nodes each containing only a single nibble of the key, we can "compress" the path by recording additional key information with each of a node's children. For example, if we have three nodes, Node 1 with key path 0x91A, Node 2 with key path 0x91A4, and Node 3 with key path 0x91A5132, then Node 1 has a key of 0x91A. Node 2 is stored at index 0x4 of Node 1's children since 4 is the next nibble in Node 2's key after skipping the common nibbles from Node 1's key. Node 3 is stored at index 0x5 of Node 1's children. Rather than have extra nodes for the remainder of Node 3's key, we instead store the rest of the key (132) in Node 1's children info.

+-----------------------------------+
| Merkle Node                       |
|                                   |
| ID: 0x0131                        |  an id representing the current node, derived from the node's value and all children ids
| Key: 0x91                         |  prefix of the key path, representing the location of the node in the trie
| Value: 0x00                       |  the value, if one exists, that is stored at the key path (pathPrefix + compressedPath)
| Children:                         |  a map of children node ids for any nodes in the trie that have this node's key path as a prefix
|   0: [:0x00542F]                  |  child 0 represents a node with key 0x910 with ID 0x00542F
|   1: [0x432:0xA0561C]             |  child 1 represents a node with key 0x911432 with ID 0xA0561C
|   ...                             |
|   15: [0x9A67B:0x02FB093]         |  child 15 represents a node with key 0x91F9A67B with ID 0x02FB093
+-----------------------------------+

Serialization

Node

Nodes are persisted in an underlying database. In order to persist nodes, we must first serialize them. Serialization is done by the encoder interface defined in codec.go.

The node serialization format is as follows:

+----------------------------------------------------+
| Value existence flag (1 byte)                      |
+----------------------------------------------------+
| Value length (varint) (optional)                   |
+----------------------------------------------------+
| Value (variable length bytes) (optional)           |
+----------------------------------------------------+
| Number of children (varint)                        |
+----------------------------------------------------+
| Child index (varint)                               |
+----------------------------------------------------+
| Child compressed path length (varint)              |
+----------------------------------------------------+
| Child compressed path (variable length bytes)      |
+----------------------------------------------------+
| Child ID (32 bytes)                                |
+----------------------------------------------------+
| Child index (varint)                               |
+----------------------------------------------------+
| Child compressed path length (varint)              |
+----------------------------------------------------+
| Child compressed path (variable length bytes)      |
+----------------------------------------------------+
| Child ID (32 bytes)                                |
+----------------------------------------------------+
|...                                                 |
+----------------------------------------------------+

Where:

  • Value existence flag is 1 if this node has a value, otherwise 0.
  • Value length is the length of the value, if it exists (i.e. if Value existince flag is 1.) Otherwise not serialized.
  • Value is the value, if it exists (i.e. if Value existince flag is 1.) Otherwise not serialized.
  • Number of children is the number of children this node has.
  • Child index is the index of a child node within the list of the node's children.
  • Child compressed path length is the length of the child node's compressed path.
  • Child compressed path is the child node's compressed path.
  • Child ID is the child node's ID.

For each child of the node, we have an additional:

+----------------------------------------------------+
| Child index (varint)                               |
+----------------------------------------------------+
| Child compressed path length (varint)              |
+----------------------------------------------------+
| Child compressed path (variable length bytes)      |
+----------------------------------------------------+
| Child ID (32 bytes)                                |
+----------------------------------------------------+

Note that the Child index are not necessarily sequential. For example, if a node has 3 children, the Child index values could be 0, 2, and 15. However, the Child index values must be strictly increasing. For example, the Child index values cannot be 0, 0, and 1, or 1, 0.

Since a node can have up to 16 children, there can be up to 16 such blocks of children data.

Example

Let's take a look at an example node.

Its byte representation (in hex) is: 0x01020204000210579EB3718A7E437D2DDCE931AC7CC05A0BC695A9C2084F5DF12FB96AD0FA32660E06FFF09845893C4F9D92C4E097FCF2589BC9D6882B1F18D1C2FC91D7DF1D3FCBDB4238

The node's key is empty (its the root) and has value 0x02. It has two children. The first is at child index 0, has compressed path 0x01 and ID (in hex) 0x579eb3718a7e437d2ddce931ac7cc05a0bc695a9c2084f5df12fb96ad0fa3266. The second is at child index 14, has compressed path 0x0F0F0F and ID (in hex) 0x9845893c4f9d92c4e097fcf2589bc9d6882b1f18d1c2fc91d7df1d3fcbdb4238.

+--------------------------------------------------------------------+
| Value existence flag (1 byte)                                      |
| 0x01                                                               |
+--------------------------------------------------------------------+
| Value length (varint) (optional)                                   |
| 0x02                                                               |
+--------------------------------------------------------------------+
| Value (variable length bytes) (optional)                           |
| 0x02                                                               |
+--------------------------------------------------------------------+
| Number of children (varint)                                        |
| 0x04                                                               |
+--------------------------------------------------------------------+
| Child index (varint)                                               |
| 0x00                                                               |
+--------------------------------------------------------------------+
| Child compressed path length (varint)                              |
| 0x02                                                               |
+--------------------------------------------------------------------+
| Child compressed path (variable length bytes)                      |
| 0x10                                                               |
+--------------------------------------------------------------------+
| Child ID (32 bytes)                                                |
| 0x579EB3718A7E437D2DDCE931AC7CC05A0BC695A9C2084F5DF12FB96AD0FA3266 |
+--------------------------------------------------------------------+
| Child index (varint)                                               |
| 0x0E                                                               |
+--------------------------------------------------------------------+
| Child compressed path length (varint)                              |
| 0x06                                                               |
+--------------------------------------------------------------------+
| Child compressed path (variable length bytes)                      |
| 0xFFF0                                                             |
+--------------------------------------------------------------------+
| Child ID (32 bytes)                                                |
| 0x9845893C4F9D92C4E097FCF2589BC9D6882B1F18D1C2FC91D7DF1D3FCBDB4238 |
+--------------------------------------------------------------------+
Node Hashing

Each node must have a unique ID that identifies it. This ID is calculated by hashing the following values:

  • The node's children
  • The node's value digest
  • The node's key

Specifically, we encode these values in the following way:

+----------------------------------------------------+
| Number of children (varint)                        |
+----------------------------------------------------+
| Child index (varint)                               |
+----------------------------------------------------+
| Child ID (32 bytes)                                |
+----------------------------------------------------+
| Child index (varint)                               |
+----------------------------------------------------+
| Child ID (32 bytes)                                |
+----------------------------------------------------+
|...                                                 |
+----------------------------------------------------+
| Value existence flag (1 byte)                      |
+----------------------------------------------------+
| Value length (varint) (optional)                   |
+----------------------------------------------------+
| Value (variable length bytes) (optional)           |
+----------------------------------------------------+
| Key length (varint)                                |
+----------------------------------------------------+
| Key (variable length bytes)                        |
+----------------------------------------------------+

Where:

  • Number of children is the number of children this node has.
  • Child index is the index of a child node within the list of the node's children.
  • Child ID is the child node's ID.
  • Value existence flag is 1 if this node has a value, otherwise 0.
  • Value length is the length of the value, if it exists (i.e. if Value existince flag is 1.) Otherwise not serialized.
  • Value is the value, if it exists (i.e. if Value existince flag is 1.) Otherwise not serialized.
  • Key length is the number of nibbles in this node's key.
  • Key is the node's key.

Note that, as with the node serialization format, the Child index values aren't necessarily sequential, but they are unique and strictly increasing. Also like the node serialization format, there can be up to 16 blocks of children data. However, note that child compressed paths are not included in the node ID calculation.

Once this is encoded, we sha256 hash the resulting bytes to get the node's ID.

Encoding Varints and Bytes

Varints are encoded with binary.PutVarint from the standard library's binary/encoding package. Bytes are encoded by simply copying them onto the buffer.

Design choices

[]byte copying

Nodes contain a []byte which represents its value. This slice should never be edited internally. This allows usage without having to make copies of it for safety. Anytime these values leave the library, for example in Get, GetValue, GetProof, GetRangeProof, etc, they need to be copied into a new slice to prevent edits made outside of the library from being reflected in the DB/TrieViews.

Single node type

A Merkle Node holds the IDs of its children, its value, as well as any path extension. This simplifies some logic and allows all of the data about a node to be loaded in a single database read. This trades off a small amount of storage efficiency (some fields may be nil but are still stored for every node).

Validity

A trieView is built atop another trie, and that trie could change at any point. If it does, all descendants of the trie will be marked invalid before the edit of the trie occurs. If an operation is performed on an invalid trie, an ErrInvalid error will be returned instead of the expected result. When a view is committed, all of its sibling views (the views that share the same parent) are marked invalid and any child views of the view have their parent updated to exclude any committed views between them and the db.

Locking

Database has a RWMutex named lock. Its read operations don't store data in a map, so a read lock suffices for read operations. Database has a Mutex named commitLock. It enforces that only a single view/batch is attempting to commit to the database at one time. lock is insufficient because there is a period of view preparation where read access should still be allowed, followed by a period where a full write lock is needed. The commitLock ensures that only a single goroutine makes the transition from read => write.

A trieView is built atop another trie, which may be the underlying Database or another trieView. It's important to guarantee atomicity/consistency of trie operations. That is, if a view method is executing, the views/database underneath the view shouldn't be changing. To prevent this, we need to use locking.

trieView has a RWMutex named lock that's held when methods that access the trie's structure are executing. It is responsible for ensuring that writing/reading from a trieView or from any ancestor is safe. It also has a RWMutex named validityTrackingLock that is held during methods that change the view's validity, tracking of child views' validity, or of the trieView parent trie. This lock ensures that writing/reading from trieView or any of its descendants is safe. The Commit function also grabs the Database's commitLock lock. This is the only trieView method that modifies the underlying Database. If an ancestor is modified during this time, the commit will error with ErrInvalid.

In some of Database's methods, we create a trieView and call unexported methods on it without locking it. We do so because the exported counterpart of the method read locks the Database, which is already locked. This pattern is safe because the Database is locked, so no data under the view is changing, and nobody else has a reference to the view, so there can't be any concurrent access.

To prevent deadlocks, trieView and Database never acquire the lock of any descendant views that are built atop it. That is, locking is always done from a view down to the underlying Database, never the other way around. The validityTrackingLock goes the opposite way. Views can validityTrackingLock their children, but not their ancestors. Because of this, any function that takes the validityTrackingLock should avoid taking the lock as this will likely trigger a deadlock. Keeping lock solely in the ancestor direction and validityTrackingLock solely in the descendant direction prevents deadlocks from occurring.

Documentation

Overview

Package merkledb is a generated GoMock package.

Index

Constants

View Source 
const (
	DefaultEvictionBatchSize = 100
	RootPath                 = EmptyPath
)
View Source 
const (
	NodeBranchFactor = 16
	HashLength       = 32
)
View Source 
const EmptyPath path = ""

Variables

View Source 
var (
	ErrStartRootNotFound = errors.New("start root is not before end root in history")
	ErrRootIDNotPresent  = errors.New("root id is not present in history")
)
View Source 
var (
	ErrInvalidProof                = errors.New("proof obtained an invalid root ID")
	ErrInvalidMaxLength            = errors.New("expected max length to be > 0")
	ErrNonIncreasingValues         = errors.New("keys sent are not in increasing order")
	ErrStateFromOutsideOfRange     = errors.New("state key falls outside of the start->end range")
	ErrNonIncreasingProofNodes     = errors.New("each proof node key must be a strict prefix of the next")
	ErrExtraProofNodes             = errors.New("extra proof nodes in path")
	ErrDataInMissingRootProof      = errors.New("there should be no state or deleted keys in a change proof that had a missing root")
	ErrNoMerkleProof               = errors.New("empty key response must include merkle proof")
	ErrShouldJustBeRoot            = errors.New("end proof should only contain root")
	ErrNoStartProof                = errors.New("no start proof")
	ErrNoEndProof                  = errors.New("no end proof")
	ErrNoProof                     = errors.New("proof has no nodes")
	ErrProofNodeNotForKey          = errors.New("the provided node has a key that is not a prefix of the specified key")
	ErrProofValueDoesntMatch       = errors.New("the provided value does not match the proof node for the provided key's value")
	ErrProofNodeHasUnincludedValue = errors.New("the provided proof has a value for a key within the range that is not present in the provided key/values")
	ErrInvalidMaybe                = errors.New("maybe is nothing but has value")
	ErrInvalidChildIndex           = fmt.Errorf("child index must be less than %d", NodeBranchFactor)
	ErrNilProofNode                = errors.New("proof node is nil")
	ErrNilValueOrHash              = errors.New("proof node's valueOrHash field is nil")
	ErrNilSerializedPath           = errors.New("serialized path is nil")
	ErrNilRangeProof               = errors.New("range proof is nil")
	ErrNilChangeProof              = errors.New("change proof is nil")
	ErrNilMaybeBytes               = errors.New("maybe bytes is nil")
	ErrNilProof                    = errors.New("proof is nil")
	ErrNilValue                    = errors.New("value is nil")
)
View Source 
var (
	ErrCommitted          = errors.New("view has been committed")
	ErrInvalid            = errors.New("the trie this view was based on has changed, rendering this view invalid")
	ErrOddLengthWithValue = errors.New(
		"the underlying db only supports whole number of byte keys, so cannot record changes with odd nibble length",
	)
	ErrGetPathToFailure = errors.New("GetPathTo failed to return the closest node")
	ErrStartAfterEnd    = errors.New("start key > end key")
	ErrViewIsNotAChild  = errors.New("passed in view is required to be a child of the current view")
	ErrNoValidRoot      = errors.New("a valid root was not provided to the trieView constructor")
)

Functions

func New 

func New(ctx context.Context, db database.Database, config Config) (*merkleDB, error)

New returns a new merkle database.

Types

type ChangeProof 

type ChangeProof struct {
	// If false, the node that created this doesn't have
	// sufficient history to generate a change proof and
	// all other fields must be empty.
	// Otherwise at least one other field is non-empty.
	HadRootsInHistory bool
	// A proof that the smallest key in the requested range does/doesn't
	// exist in the trie with the requested start root.
	// Empty if no lower bound on the requested range was given.
	// Note that this may not be an entire proof -- nodes are omitted if
	// they are also in [EndProof].
	StartProof []ProofNode
	// A proof that the largest key in [KeyValues] and [DeletedKeys]
	// does/doesn't exist in the trie with the requested start root.
	// Empty iff no upper bound on the requested range was given
	// and [KeyValues] and [DeletedKeys] are empty.
	EndProof []ProofNode
	// A subset of key-values that were added, removed, or had their values modified
	// between the requested start root (exclusive) and the requested
	// end root (inclusive).
	// Sorted by increasing key.
	KeyChanges []KeyChange
}

func (*ChangeProof) Empty 

func (proof *ChangeProof) Empty() bool

func (*ChangeProof) ToProto added in v1.10.3 

func (proof *ChangeProof) ToProto() *pb.ChangeProof

func (*ChangeProof) UnmarshalProto added in v1.10.3 

func (proof *ChangeProof) UnmarshalProto(pbProof *pb.ChangeProof) error

type ChangeProofer added in v1.10.3 

type ChangeProofer interface {
	// GetChangeProof returns a proof for a subset of the key/value changes in key range
	// [start, end] that occurred between [startRootID] and [endRootID].
	// Returns at most [maxLength] key/value pairs.
	GetChangeProof(
		ctx context.Context,
		startRootID ids.ID,
		endRootID ids.ID,
		start []byte,
		end []byte,
		maxLength int,
	) (*ChangeProof, error)

	// Returns nil iff all of the following hold:
	//   - [start] <= [end].
	//   - [proof] is non-empty iff [proof.HadRootsInHistory].
	//   - All keys in [proof.KeyValues] and [proof.DeletedKeys] are in [start, end].
	//     If [start] is empty, all keys are considered > [start].
	//     If [end] is empty, all keys are considered < [end].
	//   - [proof.KeyValues] and [proof.DeletedKeys] are sorted in order of increasing key.
	//   - [proof.StartProof] and [proof.EndProof] are well-formed.
	//   - When the keys in [proof.KeyValues] are added to [db] and the keys in [proof.DeletedKeys]
	//     are removed from [db], the root ID of [db] is [expectedEndRootID].
	//
	// Assumes [db.lock] isn't held.
	// This is defined on Database instead of ChangeProof because it accesses
	// database internals.
	VerifyChangeProof(
		ctx context.Context,
		proof *ChangeProof,
		start []byte,
		end []byte,
		expectedEndRootID ids.ID,
	) error

	// CommitChangeProof commits the key/value pairs within the [proof] to the db.
	CommitChangeProof(ctx context.Context, proof *ChangeProof) error
}

type Config 

type Config struct {
	// The number of nodes that are evicted from the cache and written to
	// disk at a time.
	EvictionBatchSize int
	// The number of changes to the database that we store in memory in order to
	// serve change proofs.
	HistoryLength int
	NodeCacheSize int
	// If [Reg] is nil, metrics are collected locally but not exported through
	// Prometheus.
	// This may be useful for testing.
	Reg    prometheus.Registerer
	Tracer trace.Tracer
}

type KeyChange added in v1.10.1 

type KeyChange struct {
	Key   []byte
	Value Maybe[[]byte]
}

type KeyValue 

type KeyValue struct {
	Key   []byte
	Value []byte
}

type Maybe 

type Maybe[T any] struct {
	// contains filtered or unexported fields
}

Maybe T = Some T | Nothing. A data wrapper that allows values to be something [Some T] or nothing Nothing. Maybe is used to wrap types: * That can't be represented by nil. * That use nil as a valid value instead of an indicator of a missing value. For more info see https://en.wikipedia.org/wiki/Option_type

func Clone added in v1.9.12 

func Clone(m Maybe[[]byte]) Maybe[[]byte]

func Nothing 

func Nothing[T any]() Maybe[T]

Returns a new Maybe[T] with no value.

func Some 

func Some[T any](val T) Maybe[T]

Returns a new Maybe[T] with the value val.

func (Maybe[T]) IsNothing 

func (m Maybe[T]) IsNothing() bool

Returns true iff [m] has a value.

func (Maybe[T]) Value 

func (m Maybe[T]) Value() T

Returns the value of [m].

type MerkleDB added in v1.10.2 

type MerkleDB interface {
	database.Database
	Trie
	MerkleRootGetter
	ProofGetter
	ChangeProofer
	RangeProofer
}

type MerkleRootGetter added in v1.10.3 

type MerkleRootGetter interface {
	// GetMerkleRoot returns the merkle root of the Trie
	GetMerkleRoot(ctx context.Context) (ids.ID, error)
}

type MockMerkleDB added in v1.10.2 

type MockMerkleDB struct {
	// contains filtered or unexported fields
}

MockMerkleDB is a mock of MerkleDB interface.

func NewMockMerkleDB added in v1.10.2 

func NewMockMerkleDB(ctrl *gomock.Controller) *MockMerkleDB

NewMockMerkleDB creates a new mock instance.

func (*MockMerkleDB) Close added in v1.10.2 

func (m *MockMerkleDB) Close() error

Close mocks base method.

func (*MockMerkleDB) CommitChangeProof added in v1.10.2 

func (m *MockMerkleDB) CommitChangeProof(arg0 context.Context, arg1 *ChangeProof) error

CommitChangeProof mocks base method.

func (*MockMerkleDB) CommitRangeProof added in v1.10.2 

func (m *MockMerkleDB) CommitRangeProof(arg0 context.Context, arg1 []byte, arg2 *RangeProof) error

CommitRangeProof mocks base method.

func (*MockMerkleDB) Compact added in v1.10.2 

func (m *MockMerkleDB) Compact(arg0, arg1 []byte) error

Compact mocks base method.

func (*MockMerkleDB) Delete added in v1.10.2 

func (m *MockMerkleDB) Delete(arg0 []byte) error

Delete mocks base method.

func (*MockMerkleDB) EXPECT added in v1.10.2 

func (m *MockMerkleDB) EXPECT() *MockMerkleDBMockRecorder

EXPECT returns an object that allows the caller to indicate expected use.

func (*MockMerkleDB) Get added in v1.10.2 

func (m *MockMerkleDB) Get(arg0 []byte) ([]byte, error)

Get mocks base method.

func (*MockMerkleDB) GetChangeProof added in v1.10.2 

func (m *MockMerkleDB) GetChangeProof(arg0 context.Context, arg1, arg2 ids.ID, arg3, arg4 []byte, arg5 int) (*ChangeProof, error)

GetChangeProof mocks base method.

func (*MockMerkleDB) GetMerkleRoot added in v1.10.2 

func (m *MockMerkleDB) GetMerkleRoot(arg0 context.Context) (ids.ID, error)

GetMerkleRoot mocks base method.

func (*MockMerkleDB) GetProof added in v1.10.2 

func (m *MockMerkleDB) GetProof(arg0 context.Context, arg1 []byte) (*Proof, error)

GetProof mocks base method.

func (*MockMerkleDB) GetRangeProof added in v1.10.2 

func (m *MockMerkleDB) GetRangeProof(arg0 context.Context, arg1, arg2 []byte, arg3 int) (*RangeProof, error)

GetRangeProof mocks base method.

func (*MockMerkleDB) GetRangeProofAtRoot added in v1.10.2 

func (m *MockMerkleDB) GetRangeProofAtRoot(arg0 context.Context, arg1 ids.ID, arg2, arg3 []byte, arg4 int) (*RangeProof, error)

GetRangeProofAtRoot mocks base method.

func (*MockMerkleDB) GetValue added in v1.10.2 

func (m *MockMerkleDB) GetValue(arg0 context.Context, arg1 []byte) ([]byte, error)

GetValue mocks base method.

func (*MockMerkleDB) GetValues added in v1.10.2 

func (m *MockMerkleDB) GetValues(arg0 context.Context, arg1 [][]byte) ([][]byte, []error)

GetValues mocks base method.

func (*MockMerkleDB) Has added in v1.10.2 

func (m *MockMerkleDB) Has(arg0 []byte) (bool, error)

Has mocks base method.

func (*MockMerkleDB) HealthCheck added in v1.10.2 

func (m *MockMerkleDB) HealthCheck(arg0 context.Context) (interface{}, error)

HealthCheck mocks base method.

func (*MockMerkleDB) Insert added in v1.10.2 

func (m *MockMerkleDB) Insert(arg0 context.Context, arg1, arg2 []byte) error

Insert mocks base method.

func (*MockMerkleDB) NewBatch added in v1.10.2 

func (m *MockMerkleDB) NewBatch() database.Batch

NewBatch mocks base method.

func (*MockMerkleDB) NewIterator added in v1.10.2 

func (m *MockMerkleDB) NewIterator() database.Iterator

NewIterator mocks base method.

func (*MockMerkleDB) NewIteratorWithPrefix added in v1.10.2 

func (m *MockMerkleDB) NewIteratorWithPrefix(arg0 []byte) database.Iterator

NewIteratorWithPrefix mocks base method.

func (*MockMerkleDB) NewIteratorWithStart added in v1.10.2 

func (m *MockMerkleDB) NewIteratorWithStart(arg0 []byte) database.Iterator

NewIteratorWithStart mocks base method.

func (*MockMerkleDB) NewIteratorWithStartAndPrefix added in v1.10.2 

func (m *MockMerkleDB) NewIteratorWithStartAndPrefix(arg0, arg1 []byte) database.Iterator

NewIteratorWithStartAndPrefix mocks base method.

func (*MockMerkleDB) NewPreallocatedView added in v1.10.2 

func (m *MockMerkleDB) NewPreallocatedView(arg0 int) (TrieView, error)

NewPreallocatedView mocks base method.

func (*MockMerkleDB) NewView added in v1.10.2 

func (m *MockMerkleDB) NewView() (TrieView, error)

NewView mocks base method.

func (*MockMerkleDB) Put added in v1.10.2 

func (m *MockMerkleDB) Put(arg0, arg1 []byte) error

Put mocks base method.

func (*MockMerkleDB) Remove added in v1.10.2 

func (m *MockMerkleDB) Remove(arg0 context.Context, arg1 []byte) error

Remove mocks base method.

func (*MockMerkleDB) VerifyChangeProof added in v1.10.2 

func (m *MockMerkleDB) VerifyChangeProof(arg0 context.Context, arg1 *ChangeProof, arg2, arg3 []byte, arg4 ids.ID) error

VerifyChangeProof mocks base method.

type MockMerkleDBMockRecorder added in v1.10.2 

type MockMerkleDBMockRecorder struct {
	// contains filtered or unexported fields
}

MockMerkleDBMockRecorder is the mock recorder for MockMerkleDB.

func (*MockMerkleDBMockRecorder) Close added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Close() *gomock.Call

Close indicates an expected call of Close.

func (*MockMerkleDBMockRecorder) CommitChangeProof added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) CommitChangeProof(arg0, arg1 interface{}) *gomock.Call

CommitChangeProof indicates an expected call of CommitChangeProof.

func (*MockMerkleDBMockRecorder) CommitRangeProof added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) CommitRangeProof(arg0, arg1, arg2 interface{}) *gomock.Call

CommitRangeProof indicates an expected call of CommitRangeProof.

func (*MockMerkleDBMockRecorder) Compact added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Compact(arg0, arg1 interface{}) *gomock.Call

Compact indicates an expected call of Compact.

func (*MockMerkleDBMockRecorder) Delete added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Delete(arg0 interface{}) *gomock.Call

Delete indicates an expected call of Delete.

func (*MockMerkleDBMockRecorder) Get added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Get(arg0 interface{}) *gomock.Call

Get indicates an expected call of Get.

func (*MockMerkleDBMockRecorder) GetChangeProof added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) GetChangeProof(arg0, arg1, arg2, arg3, arg4, arg5 interface{}) *gomock.Call

GetChangeProof indicates an expected call of GetChangeProof.

func (*MockMerkleDBMockRecorder) GetMerkleRoot added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) GetMerkleRoot(arg0 interface{}) *gomock.Call

GetMerkleRoot indicates an expected call of GetMerkleRoot.

func (*MockMerkleDBMockRecorder) GetProof added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) GetProof(arg0, arg1 interface{}) *gomock.Call

GetProof indicates an expected call of GetProof.

func (*MockMerkleDBMockRecorder) GetRangeProof added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) GetRangeProof(arg0, arg1, arg2, arg3 interface{}) *gomock.Call

GetRangeProof indicates an expected call of GetRangeProof.

func (*MockMerkleDBMockRecorder) GetRangeProofAtRoot added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) GetRangeProofAtRoot(arg0, arg1, arg2, arg3, arg4 interface{}) *gomock.Call

GetRangeProofAtRoot indicates an expected call of GetRangeProofAtRoot.

func (*MockMerkleDBMockRecorder) GetValue added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) GetValue(arg0, arg1 interface{}) *gomock.Call

GetValue indicates an expected call of GetValue.

func (*MockMerkleDBMockRecorder) GetValues added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) GetValues(arg0, arg1 interface{}) *gomock.Call

GetValues indicates an expected call of GetValues.

func (*MockMerkleDBMockRecorder) Has added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Has(arg0 interface{}) *gomock.Call

Has indicates an expected call of Has.

func (*MockMerkleDBMockRecorder) HealthCheck added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) HealthCheck(arg0 interface{}) *gomock.Call

HealthCheck indicates an expected call of HealthCheck.

func (*MockMerkleDBMockRecorder) Insert added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Insert(arg0, arg1, arg2 interface{}) *gomock.Call

Insert indicates an expected call of Insert.

func (*MockMerkleDBMockRecorder) NewBatch added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) NewBatch() *gomock.Call

NewBatch indicates an expected call of NewBatch.

func (*MockMerkleDBMockRecorder) NewIterator added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) NewIterator() *gomock.Call

NewIterator indicates an expected call of NewIterator.

func (*MockMerkleDBMockRecorder) NewIteratorWithPrefix added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) NewIteratorWithPrefix(arg0 interface{}) *gomock.Call

NewIteratorWithPrefix indicates an expected call of NewIteratorWithPrefix.

func (*MockMerkleDBMockRecorder) NewIteratorWithStart added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) NewIteratorWithStart(arg0 interface{}) *gomock.Call

NewIteratorWithStart indicates an expected call of NewIteratorWithStart.

func (*MockMerkleDBMockRecorder) NewIteratorWithStartAndPrefix added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) NewIteratorWithStartAndPrefix(arg0, arg1 interface{}) *gomock.Call

NewIteratorWithStartAndPrefix indicates an expected call of NewIteratorWithStartAndPrefix.

func (*MockMerkleDBMockRecorder) NewPreallocatedView added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) NewPreallocatedView(arg0 interface{}) *gomock.Call

NewPreallocatedView indicates an expected call of NewPreallocatedView.

func (*MockMerkleDBMockRecorder) NewView added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) NewView() *gomock.Call

NewView indicates an expected call of NewView.

func (*MockMerkleDBMockRecorder) Put added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Put(arg0, arg1 interface{}) *gomock.Call

Put indicates an expected call of Put.

func (*MockMerkleDBMockRecorder) Remove added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) Remove(arg0, arg1 interface{}) *gomock.Call

Remove indicates an expected call of Remove.

func (*MockMerkleDBMockRecorder) VerifyChangeProof added in v1.10.2 

func (mr *MockMerkleDBMockRecorder) VerifyChangeProof(arg0, arg1, arg2, arg3, arg4 interface{}) *gomock.Call

VerifyChangeProof indicates an expected call of VerifyChangeProof.

type Proof 

type Proof struct {
	// Nodes in the proof path from root --> target key
	// (or node that would be where key is if it doesn't exist).
	// Must always be non-empty (i.e. have the root node).
	Path []ProofNode
	// This is a proof that [key] exists/doesn't exist.
	Key []byte

	// Nothing if [Key] isn't in the trie.
	// Otherwise the value corresponding to [Key].
	Value Maybe[[]byte]
}

An inclusion/exclustion proof of a key.

func (*Proof) ToProto added in v1.10.3 

func (proof *Proof) ToProto() *pb.Proof

func (*Proof) UnmarshalProto added in v1.10.3 

func (proof *Proof) UnmarshalProto(pbProof *pb.Proof) error

func (*Proof) Verify 

func (proof *Proof) Verify(ctx context.Context, expectedRootID ids.ID) error

Returns nil if the trie given in [proof] has root [expectedRootID]. That is, this is a valid proof that [proof.Key] exists/doesn't exist in the trie with root [expectedRootID].

type ProofGetter added in v1.10.3 

type ProofGetter interface {
	// GetProof generates a proof of the value associated with a particular key,
	// or a proof of its absence from the trie
	GetProof(ctx context.Context, bytesPath []byte) (*Proof, error)
}

type ProofNode 

type ProofNode struct {
	KeyPath SerializedPath
	// Nothing if this is an intermediate node.
	// The value in this node if its length < [HashLen].
	// The hash of the value in this node otherwise.
	ValueOrHash Maybe[[]byte]
	Children    map[byte]ids.ID
}

func (*ProofNode) ToProto added in v1.10.3 

func (node *ProofNode) ToProto() *pb.ProofNode

Assumes [node.Key.KeyPath.NibbleLength] <= math.MaxUint64.

func (*ProofNode) UnmarshalProto added in v1.10.3 

func (node *ProofNode) UnmarshalProto(pbNode *pb.ProofNode) error

type RangeProof 

type RangeProof struct {
	// A proof that the smallest key in the requested range does/doesn't exist.
	// Note that this may not be an entire proof -- nodes are omitted if
	// they are also in [EndProof].
	StartProof []ProofNode

	// If no upper range bound was given, [KeyValues] is empty,
	// and [StartProof] is non-empty, this is empty.
	//
	// If no upper range bound was given, [KeyValues] is empty,
	// and [StartProof] is empty, this is the root.
	//
	// If an upper range bound was given and [KeyValues] is empty,
	// this is a proof for the upper range bound.
	//
	// Otherwise, this is a proof for the largest key in [KeyValues].
	EndProof []ProofNode

	// This proof proves that the key-value pairs in [KeyValues] are in the trie.
	// Sorted by increasing key.
	KeyValues []KeyValue
}

A proof that a given set of key-value pairs are in a trie.

func (*RangeProof) ToProto added in v1.10.3 

func (proof *RangeProof) ToProto() *pb.RangeProof

func (*RangeProof) UnmarshalProto added in v1.10.3 

func (proof *RangeProof) UnmarshalProto(pbProof *pb.RangeProof) error

func (*RangeProof) Verify 

func (proof *RangeProof) Verify(
	ctx context.Context,
	start []byte,
	end []byte,
	expectedRootID ids.ID,
) error

Returns nil iff all the following hold:

  • [start] <= [end].
  • [proof] is non-empty.
  • All keys in [proof.KeyValues] are in the range [start, end]. If [start] is empty, all keys are considered > [start]. If [end] is empty, all keys are considered < [end].
  • [proof.KeyValues] is sorted by increasing key.
  • [proof.StartProof] and [proof.EndProof] are well-formed.
  • One of the following holds: [end] and [proof.EndProof] are empty. [proof.StartProof], [start], [end], and [proof.KeyValues] are empty and [proof.EndProof] is just the root. [end] is non-empty and [proof.EndProof] is a valid proof of a key <= [end].
  • [expectedRootID] is the root of the trie containing the given key-value pairs and start/end proofs.

type RangeProofer added in v1.10.3 

type RangeProofer interface {
	// GetRangeProofAtRoot returns a proof for the key/value pairs in this trie within the range
	// [start, end] when the root of the trie was [rootID].
	GetRangeProofAtRoot(
		ctx context.Context,
		rootID ids.ID,
		start,
		end []byte,
		maxLength int,
	) (*RangeProof, error)

	// CommitRangeProof commits the key/value pairs within the [proof] to the db.
	// [start] is the smallest key in the range this [proof] covers.
	CommitRangeProof(ctx context.Context, start []byte, proof *RangeProof) error
}

type ReadOnlyTrie 

type ReadOnlyTrie interface {
	MerkleRootGetter
	ProofGetter

	// GetValue gets the value associated with the specified key
	// database.ErrNotFound if the key is not present
	GetValue(ctx context.Context, key []byte) ([]byte, error)

	// GetValues gets the values associated with the specified keys
	// database.ErrNotFound if the key is not present
	GetValues(ctx context.Context, keys [][]byte) ([][]byte, []error)

	// GetRangeProof generates a proof of up to maxLength smallest key/values with keys between start and end
	GetRangeProof(ctx context.Context, start, end []byte, maxLength int) (*RangeProof, error)

	database.Iteratee
	// contains filtered or unexported methods
}

type SerializedPath 

type SerializedPath struct {
	NibbleLength int
	Value        []byte
}

SerializedPath contains a path from the trie. The trie branch factor is 16, so the path may contain an odd number of nibbles. If it did contain an odd number of nibbles, the last 4 bits of the last byte should be discarded.

func (SerializedPath) AppendNibble 

func (s SerializedPath) AppendNibble(nibble byte) SerializedPath

func (SerializedPath) Equal 

func (s SerializedPath) Equal(other SerializedPath) bool

func (SerializedPath) HasPrefix 

func (s SerializedPath) HasPrefix(prefix SerializedPath) bool

HasPrefix returns true iff [prefix] is a prefix of [s] or equal to it.

func (SerializedPath) HasStrictPrefix 

func (s SerializedPath) HasStrictPrefix(prefix SerializedPath) bool

Returns true iff [prefix] is a prefix of [s] but not equal to it.

func (SerializedPath) NibbleVal 

func (s SerializedPath) NibbleVal(nibbleIndex int) byte

type Trie 

type Trie interface {
	ReadOnlyTrie

	// Remove will delete a key from the Trie
	Remove(ctx context.Context, key []byte) error

	// NewPreallocatedView returns a new view on top of this Trie with space allocated for changes
	NewPreallocatedView(estimatedChanges int) (TrieView, error)

	// NewView returns a new view on top of this Trie
	NewView() (TrieView, error)

	// Insert a key/value pair into the Trie
	Insert(ctx context.Context, key, value []byte) error
}

type TrieView 

type TrieView interface {
	Trie

	// CommitToDB takes the changes of this trie and commits them down the view stack
	// until all changes in the stack commit to the database
	// Takes the DB commit lock
	CommitToDB(ctx context.Context) error

	// CommitToParent takes changes of this TrieView and commits them to its parent Trie
	// Takes the DB commit lock
	CommitToParent(ctx context.Context) error
	// contains filtered or unexported methods
}

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.