Merge remote-tracking branch 'boltdb/master' into project-list

# Conflicts:
#	README.md

README.md

@@ -437,7 +437,7 @@ all the keys in a bucket:
 db.View(func(tx *bolt.Tx) error {
 	// Assume bucket exists and has keys
 	b := tx.Bucket([]byte("MyBucket"))
-	
+
 	b.ForEach(func(k, v []byte) error {
 		fmt.Printf("key=%s, value=%s\n", k, v)
 		return nil
@@ -617,7 +617,7 @@ Boltmobiledemo.BoltDB boltDB = Boltmobiledemo.NewBoltDB(path)
 {
     NSURL* URL= [NSURL fileURLWithPath: filePathString];
     assert([[NSFileManager defaultManager] fileExistsAtPath: [URL path]]);
-    
+
     NSError *error = nil;
     BOOL success = [URL setResourceValue: [NSNumber numberWithBool: YES]
                                   forKey: NSURLIsExcludedFromBackupKey error: &error];
@@ -839,6 +839,7 @@ Below is a list of public, open source projects that use Bolt:
 * [Request Baskets](https://github.com/darklynx/request-baskets) - A web service to collect arbitrary HTTP requests and inspect them via REST API or simple web UI, similar to [RequestBin](http://requestb.in/) service
 * [Go Report Card](https://goreportcard.com/) - Go code quality report cards as a (free and open source) service.
 * [Boltdb Boilerplate](https://github.com/bobintornado/boltdb-boilerplate) - Boilerplate wrapper around bolt aiming to make simple calls one-liners.
-* [Storm](https://github.com/asdine/storm) - A simple ORM around BoltDB
+* [lru](https://github.com/crowdriff/lru) - Easy to use Bolt-backed Least-Recently-Used (LRU) read-through cache with chainable remote stores.
+* [Storm](https://github.com/asdine/storm) - A simple ORM around BoltDB.
 
 If you are using Bolt in a project please send a pull request to add it to the list.

bolt_unix.go

@@ -11,7 +11,7 @@ import (
 )
 
 // flock acquires an advisory lock on a file descriptor.
-func flock(f *os.File, exclusive bool, timeout time.Duration) error {
+func flock(db *DB, mode os.FileMode, exclusive bool, timeout time.Duration) error {
 	var t time.Time
 	for {
 		// If we're beyond our timeout then return an error.
@@ -27,7 +27,7 @@ func flock(f *os.File, exclusive bool, timeout time.Duration) error {
 		}
 
 		// Otherwise attempt to obtain an exclusive lock.
-		err := syscall.Flock(int(f.Fd()), flag|syscall.LOCK_NB)
+		err := syscall.Flock(int(db.file.Fd()), flag|syscall.LOCK_NB)
 		if err == nil {
 			return nil
 		} else if err != syscall.EWOULDBLOCK {
@@ -40,8 +40,8 @@ func flock(f *os.File, exclusive bool, timeout time.Duration) error {
 }
 
 // funlock releases an advisory lock on a file descriptor.
-func funlock(f *os.File) error {
+func funlock(db *DB) error {
-	return syscall.Flock(int(f.Fd()), syscall.LOCK_UN)
+	return syscall.Flock(int(db.file.Fd()), syscall.LOCK_UN)
 }
 
 // mmap memory maps a DB's data file.

bolt_unix_solaris.go

@@ -11,7 +11,7 @@ import (
 )
 
 // flock acquires an advisory lock on a file descriptor.
-func flock(f *os.File, exclusive bool, timeout time.Duration) error {
+func flock(db *DB, mode os.FileMode, exclusive bool, timeout time.Duration) error {
 	var t time.Time
 	for {
 		// If we're beyond our timeout then return an error.
@@ -32,7 +32,7 @@ func flock(f *os.File, exclusive bool, timeout time.Duration) error {
 		} else {
 			lock.Type = syscall.F_RDLCK
 		}
-		err := syscall.FcntlFlock(f.Fd(), syscall.F_SETLK, &lock)
+		err := syscall.FcntlFlock(db.file.Fd(), syscall.F_SETLK, &lock)
 		if err == nil {
 			return nil
 		} else if err != syscall.EAGAIN {
@@ -45,13 +45,13 @@ func flock(f *os.File, exclusive bool, timeout time.Duration) error {
 }
 
 // funlock releases an advisory lock on a file descriptor.
-func funlock(f *os.File) error {
+func funlock(db *DB) error {
 	var lock syscall.Flock_t
 	lock.Start = 0
 	lock.Len = 0
 	lock.Type = syscall.F_UNLCK
 	lock.Whence = 0
-	return syscall.FcntlFlock(uintptr(f.Fd()), syscall.F_SETLK, &lock)
+	return syscall.FcntlFlock(uintptr(db.file.Fd()), syscall.F_SETLK, &lock)
 }
 
 // mmap memory maps a DB's data file.

bolt_windows.go

@@ -16,6 +16,8 @@ var (
 )
 
 const (
+	lockExt = ".lock"
+
 	// see https://msdn.microsoft.com/en-us/library/windows/desktop/aa365203(v=vs.85).aspx
 	flagLockExclusive       = 2
 	flagLockFailImmediately = 1
@@ -46,7 +48,16 @@ func fdatasync(db *DB) error {
 }
 
 // flock acquires an advisory lock on a file descriptor.
-func flock(f *os.File, exclusive bool, timeout time.Duration) error {
+func flock(db *DB, mode os.FileMode, exclusive bool, timeout time.Duration) error {
+	// Create a separate lock file on windows because a process
+	// cannot share an exclusive lock on the same file. This is
+	// needed during Tx.WriteTo().
+	f, err := os.OpenFile(db.path+lockExt, os.O_CREATE, mode)
+	if err != nil {
+		return err
+	}
+	db.lockfile = f
+
 	var t time.Time
 	for {
 		// If we're beyond our timeout then return an error.
@@ -62,7 +73,7 @@ func flock(f *os.File, exclusive bool, timeout time.Duration) error {
 			flag |= flagLockExclusive
 		}
 
-		err := lockFileEx(syscall.Handle(f.Fd()), flag, 0, 1, 0, &syscall.Overlapped{})
+		err := lockFileEx(syscall.Handle(db.lockfile.Fd()), flag, 0, 1, 0, &syscall.Overlapped{})
 		if err == nil {
 			return nil
 		} else if err != errLockViolation {
@@ -75,8 +86,11 @@ func flock(f *os.File, exclusive bool, timeout time.Duration) error {
 }
 
 // funlock releases an advisory lock on a file descriptor.
-func funlock(f *os.File) error {
+func funlock(db *DB) error {
-	return unlockFileEx(syscall.Handle(f.Fd()), 0, 1, 0, &syscall.Overlapped{})
+	err := unlockFileEx(syscall.Handle(db.lockfile.Fd()), 0, 1, 0, &syscall.Overlapped{})
+	db.lockfile.Close()
+	os.Remove(db.path+lockExt)
+	return err
 }
 
 // mmap memory maps a DB's data file.

db.go

@@ -93,6 +93,7 @@ type DB struct {
 
 	path     string
 	file     *os.File
+	lockfile *os.File // windows only
 	dataref  []byte // mmap'ed readonly, write throws SEGV
 	data     *[maxMapSize]byte
 	datasz   int
@@ -177,7 +178,7 @@ func Open(path string, mode os.FileMode, options *Options) (*DB, error) {
 	// if !options.ReadOnly.
 	// The database file is locked using the shared lock (more than one process may
 	// hold a lock at the same time) otherwise (options.ReadOnly is set).
-	if err := flock(db.file, !db.readOnly, options.Timeout); err != nil {
+	if err := flock(db, mode, !db.readOnly, options.Timeout); err != nil {
 		_ = db.close()
 		return nil, err
 	}
@@ -379,6 +380,10 @@ func (db *DB) Close() error {
 }
 
 func (db *DB) close() error {
+	if !db.opened {
+		return nil
+	}
+	
 	db.opened = false
 
 	db.freelist = nil
@@ -397,7 +402,7 @@ func (db *DB) close() error {
 		// No need to unlock read-only file.
 		if !db.readOnly {
 			// Unlock the file.
-			if err := funlock(db.file); err != nil {
+			if err := funlock(db); err != nil {
 				log.Printf("bolt.Close(): funlock error: %s", err)
 			}
 		}
@@ -824,8 +829,10 @@ func (db *DB) grow(sz int) error {
 	// Truncate and fsync to ensure file size metadata is flushed.
 	// https://github.com/boltdb/bolt/issues/284
 	if !db.NoGrowSync && !db.readOnly {
-		if err := db.file.Truncate(int64(sz)); err != nil {
+		if runtime.GOOS != "windows" {
-			return fmt.Errorf("file resize error: %s", err)
+			if err := db.file.Truncate(int64(sz)); err != nil {
+				return fmt.Errorf("file resize error: %s", err)
+			}
 		}
 		if err := db.file.Sync(); err != nil {
 			return fmt.Errorf("file sync error: %s", err)

node.go

@@ -463,43 +463,6 @@ func (n *node) rebalance() {
 		target = n.prevSibling()
 	}
 
-	// If target node has extra nodes then just move one over.
-	if target.numChildren() > target.minKeys() {
-		if useNextSibling {
-			// Reparent and move node.
-			if child, ok := n.bucket.nodes[target.inodes[0].pgid]; ok {
-				child.parent.removeChild(child)
-				child.parent = n
-				child.parent.children = append(child.parent.children, child)
-			}
-			n.inodes = append(n.inodes, target.inodes[0])
-			target.inodes = target.inodes[1:]
-
-			// Update target key on parent.
-			target.parent.put(target.key, target.inodes[0].key, nil, target.pgid, 0)
-			target.key = target.inodes[0].key
-			_assert(len(target.key) > 0, "rebalance(1): zero-length node key")
-		} else {
-			// Reparent and move node.
-			if child, ok := n.bucket.nodes[target.inodes[len(target.inodes)-1].pgid]; ok {
-				child.parent.removeChild(child)
-				child.parent = n
-				child.parent.children = append(child.parent.children, child)
-			}
-			n.inodes = append(n.inodes, inode{})
-			copy(n.inodes[1:], n.inodes)
-			n.inodes[0] = target.inodes[len(target.inodes)-1]
-			target.inodes = target.inodes[:len(target.inodes)-1]
-		}
-
-		// Update parent key for node.
-		n.parent.put(n.key, n.inodes[0].key, nil, n.pgid, 0)
-		n.key = n.inodes[0].key
-		_assert(len(n.key) > 0, "rebalance(2): zero-length node key")
-
-		return
-	}
-
 	// If both this node and the target node are too small then merge them.
 	if useNextSibling {
 		// Reparent all child nodes being moved.

tx.go

@@ -5,6 +5,7 @@ import (
 	"io"
 	"os"
 	"sort"
+	"strings"
 	"time"
 	"unsafe"
 )
@@ -202,8 +203,17 @@ func (tx *Tx) Commit() error {
 	// If strict mode is enabled then perform a consistency check.
 	// Only the first consistency error is reported in the panic.
 	if tx.db.StrictMode {
-		if err, ok := <-tx.Check(); ok {
+		ch := tx.Check()
-			panic("check fail: " + err.Error())
+		var errs []string
+		for {
+			err, ok := <-ch
+			if !ok {
+				break
+			}
+			errs = append(errs, err.Error())
+		}
+		if len(errs) > 0 {
+			panic("check fail: " + strings.Join(errs, "\n"))
 		}
 	}
 
@@ -297,12 +307,34 @@ func (tx *Tx) WriteTo(w io.Writer) (n int64, err error) {
 	}
 	defer func() { _ = f.Close() }()
 
-	// Copy the meta pages.
+	// Generate a meta page. We use the same page data for both meta pages.
-	tx.db.metalock.Lock()
+	buf := make([]byte, tx.db.pageSize)
-	n, err = io.CopyN(w, f, int64(tx.db.pageSize*2))
+	page := (*page)(unsafe.Pointer(&buf[0]))
-	tx.db.metalock.Unlock()
+	page.flags = metaPageFlag
+	*page.meta() = *tx.meta
+
+	// Write meta 0.
+	page.id = 0
+	page.meta().checksum = page.meta().sum64()
+	nn, err := w.Write(buf)
+	n += int64(nn)
 	if err != nil {
-		return n, fmt.Errorf("meta copy: %s", err)
+		return n, fmt.Errorf("meta 0 copy: %s", err)
+	}
+
+	// Write meta 1 with a lower transaction id.
+	page.id = 1
+	page.meta().txid -= 1
+	page.meta().checksum = page.meta().sum64()
+	nn, err = w.Write(buf)
+	n += int64(nn)
+	if err != nil {
+		return n, fmt.Errorf("meta 1 copy: %s", err)
+	}
+
+	// Move past the meta pages in the file.
+	if _, err := f.Seek(int64(tx.db.pageSize*2), os.SEEK_SET); err != nil {
+		return n, fmt.Errorf("seek: %s", err)
 	}
 
 	// Copy data pages.

tx_test.go

@@ -570,7 +570,7 @@ func TestTx_CopyFile_Error_Meta(t *testing.T) {
 
 	if err := db.View(func(tx *bolt.Tx) error {
 		return tx.Copy(&failWriter{})
-	}); err == nil || err.Error() != "meta copy: error injected for tests" {
+	}); err == nil || err.Error() != "meta 0 copy: error injected for tests" {
 		t.Fatalf("unexpected error: %v", err)
 	}
 }