Fix memory leak in secondary, cleanup code.

README.md

@@ -4,6 +4,7 @@ Replicated in-memory database and file store.
 
 ## TODO
 
+* [ ] mdb: Tests for using `nil` snapshots ?
 * [ ] mdb: tests for sanitize and validate functions
 * [ ] Test: lib/wal iterator w/ corrupt file (random corruptions)
 * [ ] Test: lib/wal io.go

go.mod

@@ -1,9 +1,10 @@
 module git.crumpington.com/public/jldb
 
-go 1.21.1
+go 1.22
 
 require (
 	github.com/google/btree v1.1.2
+	go.uber.org/goleak v1.3.0
 	golang.org/x/net v0.15.0
 	golang.org/x/sys v0.12.0
 )

go.sum

@@ -1,6 +1,16 @@
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
+github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/google/btree v1.1.2 h1:xf4v41cLI2Z6FxbKm+8Bu+m8ifhj15JuZ9sa0jZCMUU=
 github.com/google/btree v1.1.2/go.mod h1:qOPhT0dTNdNzV6Z/lhRX0YXUafgPLFUh+gZMl761Gm4=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/stretchr/testify v1.8.0 h1:pSgiaMZlXftHpm5L7V1+rVB+AZJydKsMxsQBIJw4PKk=
+github.com/stretchr/testify v1.8.0/go.mod h1:yNjHg4UonilssWZ8iaSj1OCr/vHnekPRkoO+kdMU+MU=
+go.uber.org/goleak v1.3.0 h1:2K3zAYmnTNqV73imy9J1T3WC+gmCePx2hEGkimedGto=
+go.uber.org/goleak v1.3.0/go.mod h1:CoHD4mav9JJNrW/WLlf7HGZPjdw8EucARQHekz1X6bE=
 golang.org/x/net v0.15.0 h1:ugBLEUaxABaB5AJqW9enI0ACdci2RUd4eP51NTBvuJ8=
 golang.org/x/net v0.15.0/go.mod h1:idbUs1IY1+zTqbi8yxTbhexhEEk5ur9LInksu6HrEpk=
 golang.org/x/sys v0.12.0 h1:CM0HF96J0hcLAwsHPJZjfdNzs0gftsLfgKt57wWHJ0o=
 golang.org/x/sys v0.12.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
+gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

lib/errs/error.go

@@ -8,26 +8,27 @@ import (
 )
 
 type Error struct {
-	msg        string
-	code       int64
+	Code       int64
+	Msg        string
+	StackTrace string
+
 	collection string
 	index      string
-	stackTrace string
 	err        error // Wrapped error
 }
 
 func NewErr(code int64, msg string) *Error {
 	return &Error{
-		msg:  msg,
-		code: code,
+		Code: code,
+		Msg:  msg,
 	}
 }
 
 func (e *Error) Error() string {
 	if e.collection != "" || e.index != "" {
-		return fmt.Sprintf(`[%d] (%s/%s) %s`, e.code, e.collection, e.index, e.msg)
+		return fmt.Sprintf(`[%d] (%s/%s) %s`, e.Code, e.collection, e.index, e.Msg)
 	} else {
-		return fmt.Sprintf("[%d] %s", e.code, e.msg)
+		return fmt.Sprintf("[%d] %s", e.Code, e.Msg)
 	}
 }
 
@@ -36,11 +37,15 @@ func (e *Error) Is(rhs error) bool {
 	if !ok {
 		return false
 	}
-	return e.code == e2.code
+	return e.Code == e2.Code
+}
+
+func (e *Error) Unwrap() error {
+	return e.err
 }
 
 func (e *Error) WithErr(err error) *Error {
-	if e2, ok := err.(*Error); ok && e2.code == e.code {
+	if e2, ok := err.(*Error); ok && e2.Code == e.Code {
 		return e2
 	}
 
@@ -49,18 +54,11 @@ func (e *Error) WithErr(err error) *Error {
 	return e2
 }
 
-func (e *Error) Unwrap() error {
-	if e.err != nil {
-		return e.err
-	}
-	return e
-}
-
 func (e *Error) WithMsg(msg string, args ...any) *Error {
 	err := *e
-	err.msg += ": " + fmt.Sprintf(msg, args...)
-	if len(err.stackTrace) == 0 {
-		err.stackTrace = string(debug.Stack())
+	err.Msg += ": " + fmt.Sprintf(msg, args...)
+	if len(err.StackTrace) == 0 {
+		err.StackTrace = string(debug.Stack())
 	}
 	return &err
 }
@@ -78,16 +76,16 @@ func (e *Error) WithIndex(s string) *Error {
 }
 
 func (e *Error) msgTruncacted() string {
-	if len(e.msg) > 255 {
-		return e.msg[:255]
+	if len(e.Msg) > 255 {
+		return e.Msg[:255]
 	}
-	return e.msg
+	return e.Msg
 }
 
 func (e *Error) Write(w io.Writer) error {
 	msg := e.msgTruncacted()
 
-	if err := binary.Write(w, binary.LittleEndian, e.code); err != nil {
+	if err := binary.Write(w, binary.LittleEndian, e.Code); err != nil {
 		return IO.WithErr(err)
 	}
 
@@ -103,7 +101,7 @@ func (e *Error) Read(r io.Reader) error {
 		size uint8
 	)
 
-	if err := binary.Read(r, binary.LittleEndian, &e.code); err != nil {
+	if err := binary.Read(r, binary.LittleEndian, &e.Code); err != nil {
 		return IO.WithErr(err)
 	}
 
@@ -116,6 +114,6 @@ func (e *Error) Read(r io.Reader) error {
 		return IO.WithErr(err)
 	}
 
-	e.msg = string(msgBuf)
+	e.Msg = string(msgBuf)
 	return nil
 }

lib/errs/fmt.go

@@ -10,12 +10,12 @@ func FmtDetails(err error) string {
 
 	var s string
 	if e.collection != "" || e.index != "" {
-		s = fmt.Sprintf(`[%d] (%s/%s) %s`, e.code, e.collection, e.index, e.msg)
+		s = fmt.Sprintf(`[%d] (%s/%s) %s`, e.Code, e.collection, e.index, e.Msg)
 	} else {
-		s = fmt.Sprintf("[%d] %s", e.code, e.msg)
+		s = fmt.Sprintf("[%d] %s", e.Code, e.Msg)
 	}
-	if len(e.stackTrace) != 0 {
-		s += "\n\nStack Trace:\n" + e.stackTrace + "\n"
+	if len(e.StackTrace) != 0 {
+		s += "\n\nStack Trace:\n" + e.StackTrace + "\n"
 	}
 
 	return s

lib/rep/http-handler.go

@@ -12,7 +12,6 @@ const (
 	pathStreamWAL = "stream-wal"
 )
 
-// TODO: Remove this!
 func (rep *Replicator) Handle(w http.ResponseWriter, r *http.Request) {
 	// We'll handle two types of requests: HTTP GET requests for JSON, or
 	// streaming requets for state or wall.

lib/rep/main_test.go

@@ -0,0 +1,11 @@
+package rep
+
+import (
+	"testing"
+
+	"go.uber.org/goleak"
+)
+
+func TestMain(m *testing.M) {
+	goleak.VerifyTestMain(m)
+}

lib/rep/replicator-walgc.go

@@ -15,7 +15,7 @@ func (rep *Replicator) runWALGC() {
 		select {
 		case <-ticker.C:
 			state := rep.getState()
-			before := time.Now().Unix() - rep.conf.WALSegMaxAgeSec
+			before := time.Now().Unix() - rep.conf.WALSegGCAgeSec
 			if err := rep.wal.DeleteBefore(before, state.SeqNum); err != nil {
 				log.Printf("[WAL-GC] failed to delete wal segments: %v", err)
 			}

lib/rep/replicator.go

@@ -36,8 +36,8 @@ type App struct {
 	// SendState: The primary may need to send storage state to a secondary node.
 	SendState func(conn net.Conn) error
 
-	// (1) RecvState: Secondary nodes may need to load state from the primary if the
-	// WAL is too far behind.
+	// (1) RecvState: Secondary nodes may need to load state from the primary if
+	// the WAL is too far behind.
 	RecvState func(conn net.Conn) error
 
 	// (2) InitStorage: Prepare application storage for possible calls to

lib/rep/testapp-harness_test.go

@@ -56,6 +56,7 @@ func (h TestAppHarness) Run(t *testing.T) {
 				WALSegMaxAgeSec: 1,
 				WALSegGCAgeSec:  1,
 			})
+			defer app2.Close()
 
 			val.MethodByName(method.Name).Call([]reflect.Value{
 				reflect.ValueOf(t),

mdb/collection.go

@@ -1,7 +1,6 @@
 package mdb
 
 import (
-	"bytes"
 	"encoding/json"
 	"errors"
 	"hash/crc64"
@@ -25,8 +24,6 @@ type Collection[T any] struct {
 	uniqueIndices []*Index[T]
 
 	ByID *Index[T]
-
-	buf *bytes.Buffer
 }
 
 type CollectionConfig[T any] struct {
@@ -67,7 +64,6 @@ func NewCollection[T any](db *Database, name string, conf *CollectionConfig[T])
 		validate:      conf.Validate,
 		indices:       []*Index[T]{},
 		uniqueIndices: []*Index[T]{},
-		buf:           &bytes.Buffer{},
 	}
 
 	db.addCollection(c.collectionID, c, &collectionState[T]{
@@ -159,6 +155,15 @@ func (c *Collection[T]) Get(tx *Snapshot, id uint64) *T {
 	return c.ByID.Get(tx, item)
 }
 
+func (c *Collection[T]) Has(tx *Snapshot, id uint64) bool {
+	if tx == nil {
+		tx = c.db.Snapshot()
+	}
+	item := new(T)
+	c.setID(item, id)
+	return c.ByID.Has(tx, item)
+}
+
 func (c *Collection[T]) Insert(tx *Snapshot, userItem *T) error {
 	if tx == nil {
 		return c.db.Update(func(tx *Snapshot) error {
@@ -237,6 +242,27 @@ func (c *Collection[T]) update(tx *Snapshot, userItem *T) error {
 	return nil
 }
 
+func (c *Collection[T]) UpdateFunc(tx *Snapshot, id uint64, update func(item *T) error) error {
+	if tx == nil {
+		return c.db.Update(func(tx *Snapshot) error {
+			return c.updateFunc(tx, id, update)
+		})
+	}
+	return c.updateFunc(tx, id, update)
+}
+
+func (c *Collection[T]) updateFunc(tx *Snapshot, id uint64, update func(item *T) error) error {
+	item := c.Get(tx, id)
+	if item == nil {
+		return errs.NotFound
+	}
+	if err := update(item); err != nil {
+		return err
+	}
+	c.setID(item, id) // Don't allow the ID to change.
+	return c.update(tx, item)
+}
+
 func (c *Collection[T]) Upsert(tx *Snapshot, item *T) error {
 	if tx == nil {
 		return c.db.Update(func(tx *Snapshot) error {
@@ -257,6 +283,36 @@ func (c *Collection[T]) upsert(tx *Snapshot, item *T) error {
 	return err
 }
 
+func (c *Collection[T]) UpsertFunc(tx *Snapshot, id uint64, update func(item *T) error) error {
+	if tx == nil {
+		c.db.Update(func(tx *Snapshot) error {
+			return c.upsertFunc(tx, id, update)
+		})
+	}
+	return c.upsertFunc(tx, id, update)
+}
+
+func (c *Collection[T]) upsertFunc(tx *Snapshot, id uint64, update func(item *T) error) error {
+	insert := false
+
+	item := c.Get(tx, id)
+	if item == nil {
+		item = new(T)
+		insert = true
+	}
+
+	if err := update(item); err != nil {
+		return err
+	}
+
+	c.setID(item, id) // Don't allow the ID to change.
+
+	if insert {
+		return c.insert(tx, item)
+	}
+	return c.update(tx, item)
+}
+
 func (c *Collection[T]) Delete(tx *Snapshot, itemID uint64) error {
 	if tx == nil {
 		return c.db.Update(func(tx *Snapshot) error {

mdb/db-rep.go

@@ -99,6 +99,7 @@ func (db *Database) repApply(rec wal.Record) (err error) {
 	}
 	tx.seqNum = rec.SeqNum
 	tx.timestampMS = rec.TimestampMS
+	tx.setReadOnly()
 	db.snapshot.Store(tx)
 	return nil
 }

mdb/db-testcases_test.go

@@ -743,29 +743,25 @@ var testDBTestCases = []DBTestCase{{
 
 			first := true
 			pivot := User{Name: "User1"}
-			db.Users.ByName.AscendAfter(tx, &pivot, func(u *User) bool {
+			for u := range db.Users.ByName.AscendAfter(tx, &pivot) {
 				u.Name += "Mod"
 				if err = db.Users.Update(tx, u); err != nil {
-					return false
+					return err
 				}
 				if first {
 					first = false
-					return true
+					continue
 				}
 
 				prev := db.Users.ByID.Get(tx, &User{ID: u.ID - 1})
 				if prev == nil {
-					err = errors.New("Previous user not found")
-					return false
+					return errors.New("Previous user not found")
 				}
 
 				if !strings.HasSuffix(prev.Name, "Mod") {
-					err = errors.New("Incorrect user name: " + prev.Name)
-					return false
+					return errors.New("Incorrect user name: " + prev.Name)
 				}
-
-				return true
-			})
+			}
 			return nil
 		},
 
@@ -801,29 +797,26 @@ var testDBTestCases = []DBTestCase{{
 			}
 
 			first := true
-			db.Users.ByName.DescendAfter(tx, &User{Name: "User5Mod"}, func(u *User) bool {
+			for u := range db.Users.ByName.DescendAfter(tx, &User{Name: "User5Mod"}) {
 				u.Name = strings.TrimSuffix(u.Name, "Mod")
 				if err = db.Users.Update(tx, u); err != nil {
-					return false
+					return err
 				}
 				if first {
 					first = false
-					return true
+					continue
 				}
 
 				prev := db.Users.ByID.Get(tx, &User{ID: u.ID + 1})
 				if prev == nil {
-					err = errors.New("Previous user not found")
-					return false
+					return errors.New("Previous user not found")
 				}
 
 				if strings.HasSuffix(prev.Name, "Mod") {
-					err = errors.New("Incorrect user name: " + prev.Name)
-					return false
+					return errors.New("Incorrect user name: " + prev.Name)
 				}
+			}
 
-				return true
-			})
 			return nil
 		},
 

mdb/db-testrunner_test.go

@@ -72,7 +72,7 @@ func testRunner_testCase(t *testing.T, testCase DBTestCase) {
 	}
 
 	// TODO: Why is this necessary?
-	time.Sleep(time.Second)
+	//time.Sleep(time.Second)
 	finalStep := testCase.Steps[len(testCase.Steps)-1]
 
 	secondarySnapshot := db2.Snapshot()

mdb/equality_test.go

@@ -1,7 +1,6 @@
 package mdb
 
 import (
-	"fmt"
 	"reflect"
 	"testing"
 )
@@ -20,18 +19,16 @@ func (i Index[T]) AssertEqual(t *testing.T, tx1, tx2 *Snapshot) {
 	}
 
 	errStr := ""
-	i.Ascend(tx1, func(item1 *T) bool {
+	iter := i.Ascend(tx1)
+	for item1 := range iter {
 		item2 := i.Get(tx2, item1)
 		if item2 == nil {
-			errStr = fmt.Sprintf("Indices don't match. %v not found.", item1)
-			return false
+			t.Fatalf("Indices don't match. %v not found.", item1)
 		}
 		if !reflect.DeepEqual(item1, item2) {
-			errStr = fmt.Sprintf("%v != %v", item1, item2)
-			return false
+			t.Fatalf("%v != %v", item1, item2)
 		}
-		return true
-	})
+	}
 
 	if errStr != "" {
 		t.Fatal(errStr)

mdb/index.go

@@ -1,6 +1,7 @@
 package mdb
 
 import (
+	"iter"
 	"unsafe"
 
 	"github.com/google/btree"
@@ -111,32 +112,40 @@ func (i *Index[T]) Max(tx *Snapshot) *T {
 	return nil
 }
 
-func (i *Index[T]) Ascend(tx *Snapshot, each func(*T) bool) {
+func (i *Index[T]) Ascend(tx *Snapshot) iter.Seq[*T] {
 	tx = i.ensureSnapshot(tx)
-	i.btreeForIter(tx).Ascend(func(t *T) bool {
-		return each(i.copy(t))
-	})
+	return func(yield func(*T) bool) {
+		i.btreeForIter(tx).Ascend(func(t *T) bool {
+			return yield(i.copy(t))
+		})
+	}
 }
 
-func (i *Index[T]) AscendAfter(tx *Snapshot, after *T, each func(*T) bool) {
+func (i *Index[T]) AscendAfter(tx *Snapshot, after *T) iter.Seq[*T] {
 	tx = i.ensureSnapshot(tx)
-	i.btreeForIter(tx).AscendGreaterOrEqual(after, func(t *T) bool {
-		return each(i.copy(t))
-	})
+	return func(yield func(*T) bool) {
+		i.btreeForIter(tx).AscendGreaterOrEqual(after, func(t *T) bool {
+			return yield(i.copy(t))
+		})
+	}
 }
 
-func (i *Index[T]) Descend(tx *Snapshot, each func(*T) bool) {
+func (i *Index[T]) Descend(tx *Snapshot) iter.Seq[*T] {
 	tx = i.ensureSnapshot(tx)
-	i.btreeForIter(tx).Descend(func(t *T) bool {
-		return each(i.copy(t))
-	})
+	return func(yield func(*T) bool) {
+		i.btreeForIter(tx).Descend(func(t *T) bool {
+			return yield(i.copy(t))
+		})
+	}
 }
 
-func (i *Index[T]) DescendAfter(tx *Snapshot, after *T, each func(*T) bool) {
+func (i *Index[T]) DescendAfter(tx *Snapshot, after *T) iter.Seq[*T] {
 	tx = i.ensureSnapshot(tx)
-	i.btreeForIter(tx).DescendLessOrEqual(after, func(t *T) bool {
-		return each(i.copy(t))
-	})
+	return func(yield func(*T) bool) {
+		i.btreeForIter(tx).DescendLessOrEqual(after, func(t *T) bool {
+			return yield(i.copy(t))
+		})
+	}
 }
 
 func (i *Index[T]) Count(tx *Snapshot) int {

mdb/index_test.go

@@ -1,9 +1,8 @@
 package mdb
 
 func (i Index[T]) Dump(tx *Snapshot) (l []T) {
-	i.Ascend(tx, func(t *T) bool {
+	for t := range i.Ascend(tx) {
 		l = append(l, *t)
-		return true
-	})
+	}
 	return l
 }

mdb/pfile/freelist.go

@@ -51,6 +51,10 @@ func (f *freeList) Push(pages ...uint64) {
 	}
 }
 
+func (f *freeList) SetNextPage(nextPage uint64) {
+	f.nextPage = nextPage
+}
+
 func (f *freeList) Pop(count int, out []uint64) []uint64 {
 	out = out[:0]
 

mdb/pfile/index.go

@@ -13,14 +13,19 @@ type Index struct {
 }
 
 func NewIndex(f *File) (*Index, error) {
+	firstPage, err := f.pageCount()
+	if err != nil {
+		return nil, err
+	}
+
 	idx := &Index{
-		fList: newFreeList(0),
+		fList: newFreeList(firstPage),
 		aList: *newAllocList(),
 		seen:  map[[2]uint64]struct{}{},
 		mask:  []bool{},
 	}
 
-	err := f.iterate(func(pageID uint64, page dataPage) error {
+	err = f.iterate(func(pageID uint64, page dataPage) error {
 		header := page.Header()
 		switch header.PageType {
 		case pageTypeHead:

mdb/pfile/pagefile.go

@@ -134,6 +134,21 @@ func (pf *File) writePage(page dataPage, id uint64) error {
 // Reading
 // ----------------------------------------------------------------------------
 
+func (pf *File) pageCount() (uint64, error) {
+	fi, err := pf.f.Stat()
+	if err != nil {
+		return 0, errs.IO.WithErr(err)
+	}
+
+	fileSize := fi.Size()
+	if fileSize%pageSize != 0 {
+		return 0, errs.Corrupt.WithMsg("File size isn't a multiple of page size.")
+	}
+
+	maxPage := uint64(fileSize / pageSize)
+	return maxPage, nil
+}
+
 func (pf *File) iterate(each func(pageID uint64, page dataPage) error) error {
 	pf.lock.RLock()
 	defer pf.lock.RUnlock()

mdb/txaggregator.go

@@ -1,92 +0,0 @@
-package mdb
-
-/*
-type txAggregator struct {
-	Stop     chan struct{}
-	Done     *sync.WaitGroup
-	ModChan  chan txMod
-	W        *cswal.Writer
-	Index    *pagefile.Index
-	Snapshot *atomic.Pointer[Snapshot]
-}
-
-func (p txAggregator) Run() {
-	defer p.Done.Done()
-	defer p.W.Close()
-
-	var (
-		tx       *Snapshot
-		mod      txMod
-		rec      cswal.Record
-		err      error
-		toNotify = make([]chan error, 0, 1024)
-	)
-
-READ_FIRST:
-
-	toNotify = toNotify[:0]
-
-	select {
-	case mod = <-p.ModChan:
-		goto BEGIN
-	case <-p.Stop:
-		goto END
-	}
-
-BEGIN:
-
-	tx = p.Snapshot.Load().begin()
-	goto APPLY_MOD
-
-CLONE:
-
-	tx = tx.clone()
-	goto APPLY_MOD
-
-APPLY_MOD:
-
-	if err = mod.Update(tx); err != nil {
-		mod.Resp <- err
-		goto ROLLBACK
-	}
-
-	toNotify = append(toNotify, mod.Resp)
-	goto NEXT
-
-ROLLBACK:
-
-	if len(toNotify) == 0 {
-		goto READ_FIRST
-	}
-
-	tx = tx.rollback()
-	goto NEXT
-
-NEXT:
-
-	select {
-	case mod = <-p.ModChan:
-		goto CLONE
-	default:
-		goto WRITE
-	}
-
-WRITE:
-
-	rec, err = writeChangesToWAL(tx.changes, p.Index, p.W)
-	if err == nil {
-		tx.seqNum = rec.SeqNum
-		tx.updatedAt = rec.CreatedAt
-		tx.setReadOnly()
-		p.Snapshot.Store(tx)
-	}
-
-	for i := range toNotify {
-		toNotify[i] <- err
-	}
-
-	goto READ_FIRST
-
-END:
-}
-*/