Merge pull request #312 from stevvooe/add-layer-info-cache

registry: integrate layer info cache with registry and storage

Godeps/Godeps.json

@@ -64,6 +64,14 @@
 			"ImportPath": "github.com/docker/libtrust",
 			"Rev": "fa567046d9b14f6aa788882a950d69651d230b21"
 		},
+		{
+			"ImportPath": "github.com/garyburd/redigo/internal",
+			"Rev": "535138d7bcd717d6531c701ef5933d98b1866257"
+		},
+		{
+			"ImportPath": "github.com/garyburd/redigo/redis",
+			"Rev": "535138d7bcd717d6531c701ef5933d98b1866257"
+		},
 		{
 			"ImportPath": "github.com/gorilla/context",
 			"Rev": "14f550f51af52180c2eefed15e5fd18d63c0a64a"

Godeps/_workspace/src/github.com/garyburd/redigo/internal/commandinfo.go

@@ -0,0 +1,45 @@
+// Copyright 2014 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package internal
+
+import (
+	"strings"
+)
+
+const (
+	WatchState = 1 << iota
+	MultiState
+	SubscribeState
+	MonitorState
+)
+
+type CommandInfo struct {
+	Set, Clear int
+}
+
+var commandInfos = map[string]CommandInfo{
+	"WATCH":      {Set: WatchState},
+	"UNWATCH":    {Clear: WatchState},
+	"MULTI":      {Set: MultiState},
+	"EXEC":       {Clear: WatchState | MultiState},
+	"DISCARD":    {Clear: WatchState | MultiState},
+	"PSUBSCRIBE": {Set: SubscribeState},
+	"SUBSCRIBE":  {Set: SubscribeState},
+	"MONITOR":    {Set: MonitorState},
+}
+
+func LookupCommandInfo(commandName string) CommandInfo {
+	return commandInfos[strings.ToUpper(commandName)]
+}

Godeps/_workspace/src/github.com/garyburd/redigo/internal/redistest/testdb.go

@@ -0,0 +1,65 @@
+// Copyright 2014 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+// Package redistest contains utilities for writing Redigo tests.
+package redistest
+
+import (
+	"errors"
+	"time"
+
+	"github.com/garyburd/redigo/redis"
+)
+
+type testConn struct {
+	redis.Conn
+}
+
+func (t testConn) Close() error {
+	_, err := t.Conn.Do("SELECT", "9")
+	if err != nil {
+		return nil
+	}
+	_, err = t.Conn.Do("FLUSHDB")
+	if err != nil {
+		return err
+	}
+	return t.Conn.Close()
+}
+
+// Dial dials the local Redis server and selects database 9. To prevent
+// stomping on real data, DialTestDB fails if database 9 contains data. The
+// returned connection flushes database 9 on close.
+func Dial() (redis.Conn, error) {
+	c, err := redis.DialTimeout("tcp", ":6379", 0, 1*time.Second, 1*time.Second)
+	if err != nil {
+		return nil, err
+	}
+
+	_, err = c.Do("SELECT", "9")
+	if err != nil {
+		return nil, err
+	}
+
+	n, err := redis.Int(c.Do("DBSIZE"))
+	if err != nil {
+		return nil, err
+	}
+
+	if n != 0 {
+		return nil, errors.New("database #9 is not empty, test can not continue")
+	}
+
+	return testConn{c}, nil
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/conn.go

@@ -0,0 +1,455 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"bufio"
+	"bytes"
+	"errors"
+	"fmt"
+	"io"
+	"net"
+	"strconv"
+	"sync"
+	"time"
+)
+
+// conn is the low-level implementation of Conn
+type conn struct {
+
+	// Shared
+	mu      sync.Mutex
+	pending int
+	err     error
+	conn    net.Conn
+
+	// Read
+	readTimeout time.Duration
+	br          *bufio.Reader
+
+	// Write
+	writeTimeout time.Duration
+	bw           *bufio.Writer
+
+	// Scratch space for formatting argument length.
+	// '*' or '$', length, "\r\n"
+	lenScratch [32]byte
+
+	// Scratch space for formatting integers and floats.
+	numScratch [40]byte
+}
+
+// Dial connects to the Redis server at the given network and address.
+func Dial(network, address string) (Conn, error) {
+	dialer := xDialer{}
+	return dialer.Dial(network, address)
+}
+
+// DialTimeout acts like Dial but takes timeouts for establishing the
+// connection to the server, writing a command and reading a reply.
+func DialTimeout(network, address string, connectTimeout, readTimeout, writeTimeout time.Duration) (Conn, error) {
+	netDialer := net.Dialer{Timeout: connectTimeout}
+	dialer := xDialer{
+		NetDial:      netDialer.Dial,
+		ReadTimeout:  readTimeout,
+		WriteTimeout: writeTimeout,
+	}
+	return dialer.Dial(network, address)
+}
+
+// A Dialer specifies options for connecting to a Redis server.
+type xDialer struct {
+	// NetDial specifies the dial function for creating TCP connections. If
+	// NetDial is nil, then net.Dial is used.
+	NetDial func(network, addr string) (net.Conn, error)
+
+	// ReadTimeout specifies the timeout for reading a single command
+	// reply. If ReadTimeout is zero, then no timeout is used.
+	ReadTimeout time.Duration
+
+	// WriteTimeout specifies the timeout for writing a single command.  If
+	// WriteTimeout is zero, then no timeout is used.
+	WriteTimeout time.Duration
+}
+
+// Dial connects to the Redis server at address on the named network.
+func (d *xDialer) Dial(network, address string) (Conn, error) {
+	dial := d.NetDial
+	if dial == nil {
+		dial = net.Dial
+	}
+	netConn, err := dial(network, address)
+	if err != nil {
+		return nil, err
+	}
+	return &conn{
+		conn:         netConn,
+		bw:           bufio.NewWriter(netConn),
+		br:           bufio.NewReader(netConn),
+		readTimeout:  d.ReadTimeout,
+		writeTimeout: d.WriteTimeout,
+	}, nil
+}
+
+// NewConn returns a new Redigo connection for the given net connection.
+func NewConn(netConn net.Conn, readTimeout, writeTimeout time.Duration) Conn {
+	return &conn{
+		conn:         netConn,
+		bw:           bufio.NewWriter(netConn),
+		br:           bufio.NewReader(netConn),
+		readTimeout:  readTimeout,
+		writeTimeout: writeTimeout,
+	}
+}
+
+func (c *conn) Close() error {
+	c.mu.Lock()
+	err := c.err
+	if c.err == nil {
+		c.err = errors.New("redigo: closed")
+		err = c.conn.Close()
+	}
+	c.mu.Unlock()
+	return err
+}
+
+func (c *conn) fatal(err error) error {
+	c.mu.Lock()
+	if c.err == nil {
+		c.err = err
+		// Close connection to force errors on subsequent calls and to unblock
+		// other reader or writer.
+		c.conn.Close()
+	}
+	c.mu.Unlock()
+	return err
+}
+
+func (c *conn) Err() error {
+	c.mu.Lock()
+	err := c.err
+	c.mu.Unlock()
+	return err
+}
+
+func (c *conn) writeLen(prefix byte, n int) error {
+	c.lenScratch[len(c.lenScratch)-1] = '\n'
+	c.lenScratch[len(c.lenScratch)-2] = '\r'
+	i := len(c.lenScratch) - 3
+	for {
+		c.lenScratch[i] = byte('0' + n%10)
+		i -= 1
+		n = n / 10
+		if n == 0 {
+			break
+		}
+	}
+	c.lenScratch[i] = prefix
+	_, err := c.bw.Write(c.lenScratch[i:])
+	return err
+}
+
+func (c *conn) writeString(s string) error {
+	c.writeLen('$', len(s))
+	c.bw.WriteString(s)
+	_, err := c.bw.WriteString("\r\n")
+	return err
+}
+
+func (c *conn) writeBytes(p []byte) error {
+	c.writeLen('$', len(p))
+	c.bw.Write(p)
+	_, err := c.bw.WriteString("\r\n")
+	return err
+}
+
+func (c *conn) writeInt64(n int64) error {
+	return c.writeBytes(strconv.AppendInt(c.numScratch[:0], n, 10))
+}
+
+func (c *conn) writeFloat64(n float64) error {
+	return c.writeBytes(strconv.AppendFloat(c.numScratch[:0], n, 'g', -1, 64))
+}
+
+func (c *conn) writeCommand(cmd string, args []interface{}) (err error) {
+	c.writeLen('*', 1+len(args))
+	err = c.writeString(cmd)
+	for _, arg := range args {
+		if err != nil {
+			break
+		}
+		switch arg := arg.(type) {
+		case string:
+			err = c.writeString(arg)
+		case []byte:
+			err = c.writeBytes(arg)
+		case int:
+			err = c.writeInt64(int64(arg))
+		case int64:
+			err = c.writeInt64(arg)
+		case float64:
+			err = c.writeFloat64(arg)
+		case bool:
+			if arg {
+				err = c.writeString("1")
+			} else {
+				err = c.writeString("0")
+			}
+		case nil:
+			err = c.writeString("")
+		default:
+			var buf bytes.Buffer
+			fmt.Fprint(&buf, arg)
+			err = c.writeBytes(buf.Bytes())
+		}
+	}
+	return err
+}
+
+type protocolError string
+
+func (pe protocolError) Error() string {
+	return fmt.Sprintf("redigo: %s (possible server error or unsupported concurrent read by application)", string(pe))
+}
+
+func (c *conn) readLine() ([]byte, error) {
+	p, err := c.br.ReadSlice('\n')
+	if err == bufio.ErrBufferFull {
+		return nil, protocolError("long response line")
+	}
+	if err != nil {
+		return nil, err
+	}
+	i := len(p) - 2
+	if i < 0 || p[i] != '\r' {
+		return nil, protocolError("bad response line terminator")
+	}
+	return p[:i], nil
+}
+
+// parseLen parses bulk string and array lengths.
+func parseLen(p []byte) (int, error) {
+	if len(p) == 0 {
+		return -1, protocolError("malformed length")
+	}
+
+	if p[0] == '-' && len(p) == 2 && p[1] == '1' {
+		// handle $-1 and $-1 null replies.
+		return -1, nil
+	}
+
+	var n int
+	for _, b := range p {
+		n *= 10
+		if b < '0' || b > '9' {
+			return -1, protocolError("illegal bytes in length")
+		}
+		n += int(b - '0')
+	}
+
+	return n, nil
+}
+
+// parseInt parses an integer reply.
+func parseInt(p []byte) (interface{}, error) {
+	if len(p) == 0 {
+		return 0, protocolError("malformed integer")
+	}
+
+	var negate bool
+	if p[0] == '-' {
+		negate = true
+		p = p[1:]
+		if len(p) == 0 {
+			return 0, protocolError("malformed integer")
+		}
+	}
+
+	var n int64
+	for _, b := range p {
+		n *= 10
+		if b < '0' || b > '9' {
+			return 0, protocolError("illegal bytes in length")
+		}
+		n += int64(b - '0')
+	}
+
+	if negate {
+		n = -n
+	}
+	return n, nil
+}
+
+var (
+	okReply   interface{} = "OK"
+	pongReply interface{} = "PONG"
+)
+
+func (c *conn) readReply() (interface{}, error) {
+	line, err := c.readLine()
+	if err != nil {
+		return nil, err
+	}
+	if len(line) == 0 {
+		return nil, protocolError("short response line")
+	}
+	switch line[0] {
+	case '+':
+		switch {
+		case len(line) == 3 && line[1] == 'O' && line[2] == 'K':
+			// Avoid allocation for frequent "+OK" response.
+			return okReply, nil
+		case len(line) == 5 && line[1] == 'P' && line[2] == 'O' && line[3] == 'N' && line[4] == 'G':
+			// Avoid allocation in PING command benchmarks :)
+			return pongReply, nil
+		default:
+			return string(line[1:]), nil
+		}
+	case '-':
+		return Error(string(line[1:])), nil
+	case ':':
+		return parseInt(line[1:])
+	case '$':
+		n, err := parseLen(line[1:])
+		if n < 0 || err != nil {
+			return nil, err
+		}
+		p := make([]byte, n)
+		_, err = io.ReadFull(c.br, p)
+		if err != nil {
+			return nil, err
+		}
+		if line, err := c.readLine(); err != nil {
+			return nil, err
+		} else if len(line) != 0 {
+			return nil, protocolError("bad bulk string format")
+		}
+		return p, nil
+	case '*':
+		n, err := parseLen(line[1:])
+		if n < 0 || err != nil {
+			return nil, err
+		}
+		r := make([]interface{}, n)
+		for i := range r {
+			r[i], err = c.readReply()
+			if err != nil {
+				return nil, err
+			}
+		}
+		return r, nil
+	}
+	return nil, protocolError("unexpected response line")
+}
+
+func (c *conn) Send(cmd string, args ...interface{}) error {
+	c.mu.Lock()
+	c.pending += 1
+	c.mu.Unlock()
+	if c.writeTimeout != 0 {
+		c.conn.SetWriteDeadline(time.Now().Add(c.writeTimeout))
+	}
+	if err := c.writeCommand(cmd, args); err != nil {
+		return c.fatal(err)
+	}
+	return nil
+}
+
+func (c *conn) Flush() error {
+	if c.writeTimeout != 0 {
+		c.conn.SetWriteDeadline(time.Now().Add(c.writeTimeout))
+	}
+	if err := c.bw.Flush(); err != nil {
+		return c.fatal(err)
+	}
+	return nil
+}
+
+func (c *conn) Receive() (reply interface{}, err error) {
+	if c.readTimeout != 0 {
+		c.conn.SetReadDeadline(time.Now().Add(c.readTimeout))
+	}
+	if reply, err = c.readReply(); err != nil {
+		return nil, c.fatal(err)
+	}
+	// When using pub/sub, the number of receives can be greater than the
+	// number of sends. To enable normal use of the connection after
+	// unsubscribing from all channels, we do not decrement pending to a
+	// negative value.
+	//
+	// The pending field is decremented after the reply is read to handle the
+	// case where Receive is called before Send.
+	c.mu.Lock()
+	if c.pending > 0 {
+		c.pending -= 1
+	}
+	c.mu.Unlock()
+	if err, ok := reply.(Error); ok {
+		return nil, err
+	}
+	return
+}
+
+func (c *conn) Do(cmd string, args ...interface{}) (interface{}, error) {
+	c.mu.Lock()
+	pending := c.pending
+	c.pending = 0
+	c.mu.Unlock()
+
+	if cmd == "" && pending == 0 {
+		return nil, nil
+	}
+
+	if c.writeTimeout != 0 {
+		c.conn.SetWriteDeadline(time.Now().Add(c.writeTimeout))
+	}
+
+	if cmd != "" {
+		c.writeCommand(cmd, args)
+	}
+
+	if err := c.bw.Flush(); err != nil {
+		return nil, c.fatal(err)
+	}
+
+	if c.readTimeout != 0 {
+		c.conn.SetReadDeadline(time.Now().Add(c.readTimeout))
+	}
+
+	if cmd == "" {
+		reply := make([]interface{}, pending)
+		for i := range reply {
+			r, e := c.readReply()
+			if e != nil {
+				return nil, c.fatal(e)
+			}
+			reply[i] = r
+		}
+		return reply, nil
+	}
+
+	var err error
+	var reply interface{}
+	for i := 0; i <= pending; i++ {
+		var e error
+		if reply, e = c.readReply(); e != nil {
+			return nil, c.fatal(e)
+		}
+		if e, ok := reply.(Error); ok && err == nil {
+			err = e
+		}
+	}
+	return reply, err
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/conn_test.go

@@ -0,0 +1,542 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis_test
+
+import (
+	"bufio"
+	"bytes"
+	"math"
+	"net"
+	"reflect"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/garyburd/redigo/internal/redistest"
+	"github.com/garyburd/redigo/redis"
+)
+
+var writeTests = []struct {
+	args     []interface{}
+	expected string
+}{
+	{
+		[]interface{}{"SET", "key", "value"},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$5\r\nvalue\r\n",
+	},
+	{
+		[]interface{}{"SET", "key", "value"},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$5\r\nvalue\r\n",
+	},
+	{
+		[]interface{}{"SET", "key", byte(100)},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$3\r\n100\r\n",
+	},
+	{
+		[]interface{}{"SET", "key", 100},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$3\r\n100\r\n",
+	},
+	{
+		[]interface{}{"SET", "key", int64(math.MinInt64)},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$20\r\n-9223372036854775808\r\n",
+	},
+	{
+		[]interface{}{"SET", "key", float64(1349673917.939762)},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$21\r\n1.349673917939762e+09\r\n",
+	},
+	{
+		[]interface{}{"SET", "key", ""},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$0\r\n\r\n",
+	},
+	{
+		[]interface{}{"SET", "key", nil},
+		"*3\r\n$3\r\nSET\r\n$3\r\nkey\r\n$0\r\n\r\n",
+	},
+	{
+		[]interface{}{"ECHO", true, false},
+		"*3\r\n$4\r\nECHO\r\n$1\r\n1\r\n$1\r\n0\r\n",
+	},
+}
+
+func TestWrite(t *testing.T) {
+	for _, tt := range writeTests {
+		var buf bytes.Buffer
+		rw := bufio.ReadWriter{Writer: bufio.NewWriter(&buf)}
+		c := redis.NewConnBufio(rw)
+		err := c.Send(tt.args[0].(string), tt.args[1:]...)
+		if err != nil {
+			t.Errorf("Send(%v) returned error %v", tt.args, err)
+			continue
+		}
+		rw.Flush()
+		actual := buf.String()
+		if actual != tt.expected {
+			t.Errorf("Send(%v) = %q, want %q", tt.args, actual, tt.expected)
+		}
+	}
+}
+
+var errorSentinel = &struct{}{}
+
+var readTests = []struct {
+	reply    string
+	expected interface{}
+}{
+	{
+		"+OK\r\n",
+		"OK",
+	},
+	{
+		"+PONG\r\n",
+		"PONG",
+	},
+	{
+		"@OK\r\n",
+		errorSentinel,
+	},
+	{
+		"$6\r\nfoobar\r\n",
+		[]byte("foobar"),
+	},
+	{
+		"$-1\r\n",
+		nil,
+	},
+	{
+		":1\r\n",
+		int64(1),
+	},
+	{
+		":-2\r\n",
+		int64(-2),
+	},
+	{
+		"*0\r\n",
+		[]interface{}{},
+	},
+	{
+		"*-1\r\n",
+		nil,
+	},
+	{
+		"*4\r\n$3\r\nfoo\r\n$3\r\nbar\r\n$5\r\nHello\r\n$5\r\nWorld\r\n",
+		[]interface{}{[]byte("foo"), []byte("bar"), []byte("Hello"), []byte("World")},
+	},
+	{
+		"*3\r\n$3\r\nfoo\r\n$-1\r\n$3\r\nbar\r\n",
+		[]interface{}{[]byte("foo"), nil, []byte("bar")},
+	},
+
+	{
+		// "x" is not a valid length
+		"$x\r\nfoobar\r\n",
+		errorSentinel,
+	},
+	{
+		// -2 is not a valid length
+		"$-2\r\n",
+		errorSentinel,
+	},
+	{
+		// "x"  is not a valid integer
+		":x\r\n",
+		errorSentinel,
+	},
+	{
+		// missing \r\n following value
+		"$6\r\nfoobar",
+		errorSentinel,
+	},
+	{
+		// short value
+		"$6\r\nxx",
+		errorSentinel,
+	},
+	{
+		// long value
+		"$6\r\nfoobarx\r\n",
+		errorSentinel,
+	},
+}
+
+func TestRead(t *testing.T) {
+	for _, tt := range readTests {
+		rw := bufio.ReadWriter{
+			Reader: bufio.NewReader(strings.NewReader(tt.reply)),
+			Writer: bufio.NewWriter(nil), // writer need to support Flush
+		}
+		c := redis.NewConnBufio(rw)
+		actual, err := c.Receive()
+		if tt.expected == errorSentinel {
+			if err == nil {
+				t.Errorf("Receive(%q) did not return expected error", tt.reply)
+			}
+		} else {
+			if err != nil {
+				t.Errorf("Receive(%q) returned error %v", tt.reply, err)
+				continue
+			}
+			if !reflect.DeepEqual(actual, tt.expected) {
+				t.Errorf("Receive(%q) = %v, want %v", tt.reply, actual, tt.expected)
+			}
+		}
+	}
+}
+
+var testCommands = []struct {
+	args     []interface{}
+	expected interface{}
+}{
+	{
+		[]interface{}{"PING"},
+		"PONG",
+	},
+	{
+		[]interface{}{"SET", "foo", "bar"},
+		"OK",
+	},
+	{
+		[]interface{}{"GET", "foo"},
+		[]byte("bar"),
+	},
+	{
+		[]interface{}{"GET", "nokey"},
+		nil,
+	},
+	{
+		[]interface{}{"MGET", "nokey", "foo"},
+		[]interface{}{nil, []byte("bar")},
+	},
+	{
+		[]interface{}{"INCR", "mycounter"},
+		int64(1),
+	},
+	{
+		[]interface{}{"LPUSH", "mylist", "foo"},
+		int64(1),
+	},
+	{
+		[]interface{}{"LPUSH", "mylist", "bar"},
+		int64(2),
+	},
+	{
+		[]interface{}{"LRANGE", "mylist", 0, -1},
+		[]interface{}{[]byte("bar"), []byte("foo")},
+	},
+	{
+		[]interface{}{"MULTI"},
+		"OK",
+	},
+	{
+		[]interface{}{"LRANGE", "mylist", 0, -1},
+		"QUEUED",
+	},
+	{
+		[]interface{}{"PING"},
+		"QUEUED",
+	},
+	{
+		[]interface{}{"EXEC"},
+		[]interface{}{
+			[]interface{}{[]byte("bar"), []byte("foo")},
+			"PONG",
+		},
+	},
+}
+
+func TestDoCommands(t *testing.T) {
+	c, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer c.Close()
+
+	for _, cmd := range testCommands {
+		actual, err := c.Do(cmd.args[0].(string), cmd.args[1:]...)
+		if err != nil {
+			t.Errorf("Do(%v) returned error %v", cmd.args, err)
+			continue
+		}
+		if !reflect.DeepEqual(actual, cmd.expected) {
+			t.Errorf("Do(%v) = %v, want %v", cmd.args, actual, cmd.expected)
+		}
+	}
+}
+
+func TestPipelineCommands(t *testing.T) {
+	c, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer c.Close()
+
+	for _, cmd := range testCommands {
+		if err := c.Send(cmd.args[0].(string), cmd.args[1:]...); err != nil {
+			t.Fatalf("Send(%v) returned error %v", cmd.args, err)
+		}
+	}
+	if err := c.Flush(); err != nil {
+		t.Errorf("Flush() returned error %v", err)
+	}
+	for _, cmd := range testCommands {
+		actual, err := c.Receive()
+		if err != nil {
+			t.Fatalf("Receive(%v) returned error %v", cmd.args, err)
+		}
+		if !reflect.DeepEqual(actual, cmd.expected) {
+			t.Errorf("Receive(%v) = %v, want %v", cmd.args, actual, cmd.expected)
+		}
+	}
+}
+
+func TestBlankCommmand(t *testing.T) {
+	c, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer c.Close()
+
+	for _, cmd := range testCommands {
+		if err := c.Send(cmd.args[0].(string), cmd.args[1:]...); err != nil {
+			t.Fatalf("Send(%v) returned error %v", cmd.args, err)
+		}
+	}
+	reply, err := redis.Values(c.Do(""))
+	if err != nil {
+		t.Fatalf("Do() returned error %v", err)
+	}
+	if len(reply) != len(testCommands) {
+		t.Fatalf("len(reply)=%d, want %d", len(reply), len(testCommands))
+	}
+	for i, cmd := range testCommands {
+		actual := reply[i]
+		if !reflect.DeepEqual(actual, cmd.expected) {
+			t.Errorf("Receive(%v) = %v, want %v", cmd.args, actual, cmd.expected)
+		}
+	}
+}
+
+func TestRecvBeforeSend(t *testing.T) {
+	c, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer c.Close()
+	done := make(chan struct{})
+	go func() {
+		c.Receive()
+		close(done)
+	}()
+	time.Sleep(time.Millisecond)
+	c.Send("PING")
+	c.Flush()
+	<-done
+	_, err = c.Do("")
+	if err != nil {
+		t.Fatalf("error=%v", err)
+	}
+}
+
+func TestError(t *testing.T) {
+	c, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer c.Close()
+
+	c.Do("SET", "key", "val")
+	_, err = c.Do("HSET", "key", "fld", "val")
+	if err == nil {
+		t.Errorf("Expected err for HSET on string key.")
+	}
+	if c.Err() != nil {
+		t.Errorf("Conn has Err()=%v, expect nil", c.Err())
+	}
+	_, err = c.Do("SET", "key", "val")
+	if err != nil {
+		t.Errorf("Do(SET, key, val) returned error %v, expected nil.", err)
+	}
+}
+
+func TestReadDeadline(t *testing.T) {
+	l, err := net.Listen("tcp", "127.0.0.1:0")
+	if err != nil {
+		t.Fatalf("net.Listen returned %v", err)
+	}
+	defer l.Close()
+
+	go func() {
+		for {
+			c, err := l.Accept()
+			if err != nil {
+				return
+			}
+			go func() {
+				time.Sleep(time.Second)
+				c.Write([]byte("+OK\r\n"))
+				c.Close()
+			}()
+		}
+	}()
+
+	c1, err := redis.DialTimeout(l.Addr().Network(), l.Addr().String(), 0, time.Millisecond, 0)
+	if err != nil {
+		t.Fatalf("redis.Dial returned %v", err)
+	}
+	defer c1.Close()
+
+	_, err = c1.Do("PING")
+	if err == nil {
+		t.Fatalf("c1.Do() returned nil, expect error")
+	}
+	if c1.Err() == nil {
+		t.Fatalf("c1.Err() = nil, expect error")
+	}
+
+	c2, err := redis.DialTimeout(l.Addr().Network(), l.Addr().String(), 0, time.Millisecond, 0)
+	if err != nil {
+		t.Fatalf("redis.Dial returned %v", err)
+	}
+	defer c2.Close()
+
+	c2.Send("PING")
+	c2.Flush()
+	_, err = c2.Receive()
+	if err == nil {
+		t.Fatalf("c2.Receive() returned nil, expect error")
+	}
+	if c2.Err() == nil {
+		t.Fatalf("c2.Err() = nil, expect error")
+	}
+}
+
+// Connect to local instance of Redis running on the default port.
+func ExampleDial(x int) {
+	c, err := redis.Dial("tcp", ":6379")
+	if err != nil {
+		// handle error
+	}
+	defer c.Close()
+}
+
+// TextExecError tests handling of errors in a transaction. See
+// http://redis.io/topics/transactions for information on how Redis handles
+// errors in a transaction.
+func TestExecError(t *testing.T) {
+	c, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer c.Close()
+
+	// Execute commands that fail before EXEC is called.
+
+	c.Do("ZADD", "k0", 0, 0)
+	c.Send("MULTI")
+	c.Send("NOTACOMMAND", "k0", 0, 0)
+	c.Send("ZINCRBY", "k0", 0, 0)
+	v, err := c.Do("EXEC")
+	if err == nil {
+		t.Fatalf("EXEC returned values %v, expected error", v)
+	}
+
+	// Execute commands that fail after EXEC is called. The first command
+	// returns an error.
+
+	c.Do("ZADD", "k1", 0, 0)
+	c.Send("MULTI")
+	c.Send("HSET", "k1", 0, 0)
+	c.Send("ZINCRBY", "k1", 0, 0)
+	v, err = c.Do("EXEC")
+	if err != nil {
+		t.Fatalf("EXEC returned error %v", err)
+	}
+
+	vs, err := redis.Values(v, nil)
+	if err != nil {
+		t.Fatalf("Values(v) returned error %v", err)
+	}
+
+	if len(vs) != 2 {
+		t.Fatalf("len(vs) == %d, want 2", len(vs))
+	}
+
+	if _, ok := vs[0].(error); !ok {
+		t.Fatalf("first result is type %T, expected error", vs[0])
+	}
+
+	if _, ok := vs[1].([]byte); !ok {
+		t.Fatalf("second result is type %T, expected []byte", vs[2])
+	}
+
+	// Execute commands that fail after EXEC is called. The second command
+	// returns an error.
+
+	c.Do("ZADD", "k2", 0, 0)
+	c.Send("MULTI")
+	c.Send("ZINCRBY", "k2", 0, 0)
+	c.Send("HSET", "k2", 0, 0)
+	v, err = c.Do("EXEC")
+	if err != nil {
+		t.Fatalf("EXEC returned error %v", err)
+	}
+
+	vs, err = redis.Values(v, nil)
+	if err != nil {
+		t.Fatalf("Values(v) returned error %v", err)
+	}
+
+	if len(vs) != 2 {
+		t.Fatalf("len(vs) == %d, want 2", len(vs))
+	}
+
+	if _, ok := vs[0].([]byte); !ok {
+		t.Fatalf("first result is type %T, expected []byte", vs[0])
+	}
+
+	if _, ok := vs[1].(error); !ok {
+		t.Fatalf("second result is type %T, expected error", vs[2])
+	}
+}
+
+func BenchmarkDoEmpty(b *testing.B) {
+	b.StopTimer()
+	c, err := redistest.Dial()
+	if err != nil {
+		b.Fatal(err)
+	}
+	defer c.Close()
+	b.StartTimer()
+	for i := 0; i < b.N; i++ {
+		if _, err := c.Do(""); err != nil {
+			b.Fatal(err)
+		}
+	}
+}
+
+func BenchmarkDoPing(b *testing.B) {
+	b.StopTimer()
+	c, err := redistest.Dial()
+	if err != nil {
+		b.Fatal(err)
+	}
+	defer c.Close()
+	b.StartTimer()
+	for i := 0; i < b.N; i++ {
+		if _, err := c.Do("PING"); err != nil {
+			b.Fatal(err)
+		}
+	}
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/doc.go

@@ -0,0 +1,169 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+// Package redis is a client for the Redis database.
+//
+// The Redigo FAQ (https://github.com/garyburd/redigo/wiki/FAQ) contains more
+// documentation about this package.
+//
+// Connections
+//
+// The Conn interface is the primary interface for working with Redis.
+// Applications create connections by calling the Dial, DialWithTimeout or
+// NewConn functions. In the future, functions will be added for creating
+// sharded and other types of connections.
+//
+// The application must call the connection Close method when the application
+// is done with the connection.
+//
+// Executing Commands
+//
+// The Conn interface has a generic method for executing Redis commands:
+//
+//  Do(commandName string, args ...interface{}) (reply interface{}, err error)
+//
+// The Redis command reference (http://redis.io/commands) lists the available
+// commands. An example of using the Redis APPEND command is:
+//
+//  n, err := conn.Do("APPEND", "key", "value")
+//
+// The Do method converts command arguments to binary strings for transmission
+// to the server as follows:
+//
+//  Go Type                 Conversion
+//  []byte                  Sent as is
+//  string                  Sent as is
+//  int, int64              strconv.FormatInt(v)
+//  float64                 strconv.FormatFloat(v, 'g', -1, 64)
+//  bool                    true -> "1", false -> "0"
+//  nil                     ""
+//  all other types         fmt.Print(v)
+//
+// Redis command reply types are represented using the following Go types:
+//
+//  Redis type              Go type
+//  error                   redis.Error
+//  integer                 int64
+//  simple string           string
+//  bulk string             []byte or nil if value not present.
+//  array                   []interface{} or nil if value not present.
+//
+// Use type assertions or the reply helper functions to convert from
+// interface{} to the specific Go type for the command result.
+//
+// Pipelining
+//
+// Connections support pipelining using the Send, Flush and Receive methods.
+//
+//  Send(commandName string, args ...interface{}) error
+//  Flush() error
+//  Receive() (reply interface{}, err error)
+//
+// Send writes the command to the connection's output buffer. Flush flushes the
+// connection's output buffer to the server. Receive reads a single reply from
+// the server. The following example shows a simple pipeline.
+//
+//  c.Send("SET", "foo", "bar")
+//  c.Send("GET", "foo")
+//  c.Flush()
+//  c.Receive() // reply from SET
+//  v, err = c.Receive() // reply from GET
+//
+// The Do method combines the functionality of the Send, Flush and Receive
+// methods. The Do method starts by writing the command and flushing the output
+// buffer. Next, the Do method receives all pending replies including the reply
+// for the command just sent by Do. If any of the received replies is an error,
+// then Do returns the error. If there are no errors, then Do returns the last
+// reply. If the command argument to the Do method is "", then the Do method
+// will flush the output buffer and receive pending replies without sending a
+// command.
+//
+// Use the Send and Do methods to implement pipelined transactions.
+//
+//  c.Send("MULTI")
+//  c.Send("INCR", "foo")
+//  c.Send("INCR", "bar")
+//  r, err := c.Do("EXEC")
+//  fmt.Println(r) // prints [1, 1]
+//
+// Concurrency
+//
+// Connections do not support concurrent calls to the write methods (Send,
+// Flush) or concurrent calls to the read method (Receive). Connections do
+// allow a concurrent reader and writer.
+//
+// Because the Do method combines the functionality of Send, Flush and Receive,
+// the Do method cannot be called concurrently with the other methods.
+//
+// For full concurrent access to Redis, use the thread-safe Pool to get and
+// release connections from within a goroutine.
+//
+// Publish and Subscribe
+//
+// Use the Send, Flush and Receive methods to implement Pub/Sub subscribers.
+//
+//  c.Send("SUBSCRIBE", "example")
+//  c.Flush()
+//  for {
+//      reply, err := c.Receive()
+//      if err != nil {
+//          return err
+//      }
+//      // process pushed message
+//  }
+//
+// The PubSubConn type wraps a Conn with convenience methods for implementing
+// subscribers. The Subscribe, PSubscribe, Unsubscribe and PUnsubscribe methods
+// send and flush a subscription management command. The receive method
+// converts a pushed message to convenient types for use in a type switch.
+//
+//  psc := redis.PubSubConn{c}
+//  psc.Subscribe("example")
+//  for {
+//      switch v := psc.Receive().(type) {
+//      case redis.Message:
+//          fmt.Printf("%s: message: %s\n", v.Channel, v.Data)
+//      case redis.Subscription:
+//          fmt.Printf("%s: %s %d\n", v.Channel, v.Kind, v.Count)
+//      case error:
+//          return v
+//      }
+//  }
+//
+// Reply Helpers
+//
+// The Bool, Int, Bytes, String, Strings and Values functions convert a reply
+// to a value of a specific type. To allow convenient wrapping of calls to the
+// connection Do and Receive methods, the functions take a second argument of
+// type error.  If the error is non-nil, then the helper function returns the
+// error. If the error is nil, the function converts the reply to the specified
+// type:
+//
+//  exists, err := redis.Bool(c.Do("EXISTS", "foo"))
+//  if err != nil {
+//      // handle error return from c.Do or type conversion error.
+//  }
+//
+// The Scan function converts elements of a array reply to Go types:
+//
+//  var value1 int
+//  var value2 string
+//  reply, err := redis.Values(c.Do("MGET", "key1", "key2"))
+//  if err != nil {
+//      // handle error
+//  }
+//   if _, err := redis.Scan(reply, &value1, &value2); err != nil {
+//      // handle error
+//  }
+package redis

Godeps/_workspace/src/github.com/garyburd/redigo/redis/log.go

@@ -0,0 +1,117 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"bytes"
+	"fmt"
+	"log"
+)
+
+// NewLoggingConn returns a logging wrapper around a connection.
+func NewLoggingConn(conn Conn, logger *log.Logger, prefix string) Conn {
+	if prefix != "" {
+		prefix = prefix + "."
+	}
+	return &loggingConn{conn, logger, prefix}
+}
+
+type loggingConn struct {
+	Conn
+	logger *log.Logger
+	prefix string
+}
+
+func (c *loggingConn) Close() error {
+	err := c.Conn.Close()
+	var buf bytes.Buffer
+	fmt.Fprintf(&buf, "%sClose() -> (%v)", c.prefix, err)
+	c.logger.Output(2, buf.String())
+	return err
+}
+
+func (c *loggingConn) printValue(buf *bytes.Buffer, v interface{}) {
+	const chop = 32
+	switch v := v.(type) {
+	case []byte:
+		if len(v) > chop {
+			fmt.Fprintf(buf, "%q...", v[:chop])
+		} else {
+			fmt.Fprintf(buf, "%q", v)
+		}
+	case string:
+		if len(v) > chop {
+			fmt.Fprintf(buf, "%q...", v[:chop])
+		} else {
+			fmt.Fprintf(buf, "%q", v)
+		}
+	case []interface{}:
+		if len(v) == 0 {
+			buf.WriteString("[]")
+		} else {
+			sep := "["
+			fin := "]"
+			if len(v) > chop {
+				v = v[:chop]
+				fin = "...]"
+			}
+			for _, vv := range v {
+				buf.WriteString(sep)
+				c.printValue(buf, vv)
+				sep = ", "
+			}
+			buf.WriteString(fin)
+		}
+	default:
+		fmt.Fprint(buf, v)
+	}
+}
+
+func (c *loggingConn) print(method, commandName string, args []interface{}, reply interface{}, err error) {
+	var buf bytes.Buffer
+	fmt.Fprintf(&buf, "%s%s(", c.prefix, method)
+	if method != "Receive" {
+		buf.WriteString(commandName)
+		for _, arg := range args {
+			buf.WriteString(", ")
+			c.printValue(&buf, arg)
+		}
+	}
+	buf.WriteString(") -> (")
+	if method != "Send" {
+		c.printValue(&buf, reply)
+		buf.WriteString(", ")
+	}
+	fmt.Fprintf(&buf, "%v)", err)
+	c.logger.Output(3, buf.String())
+}
+
+func (c *loggingConn) Do(commandName string, args ...interface{}) (interface{}, error) {
+	reply, err := c.Conn.Do(commandName, args...)
+	c.print("Do", commandName, args, reply, err)
+	return reply, err
+}
+
+func (c *loggingConn) Send(commandName string, args ...interface{}) error {
+	err := c.Conn.Send(commandName, args...)
+	c.print("Send", commandName, args, nil, err)
+	return err
+}
+
+func (c *loggingConn) Receive() (interface{}, error) {
+	reply, err := c.Conn.Receive()
+	c.print("Receive", "", nil, reply, err)
+	return reply, err
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/pool.go

@@ -0,0 +1,389 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"bytes"
+	"container/list"
+	"crypto/rand"
+	"crypto/sha1"
+	"errors"
+	"io"
+	"strconv"
+	"sync"
+	"time"
+
+	"github.com/garyburd/redigo/internal"
+)
+
+var nowFunc = time.Now // for testing
+
+// ErrPoolExhausted is returned from a pool connection method (Do, Send,
+// Receive, Flush, Err) when the maximum number of database connections in the
+// pool has been reached.
+var ErrPoolExhausted = errors.New("redigo: connection pool exhausted")
+
+var (
+	errPoolClosed = errors.New("redigo: connection pool closed")
+	errConnClosed = errors.New("redigo: connection closed")
+)
+
+// Pool maintains a pool of connections. The application calls the Get method
+// to get a connection from the pool and the connection's Close method to
+// return the connection's resources to the pool.
+//
+// The following example shows how to use a pool in a web application. The
+// application creates a pool at application startup and makes it available to
+// request handlers using a global variable.
+//
+//  func newPool(server, password string) *redis.Pool {
+//      return &redis.Pool{
+//          MaxIdle: 3,
+//          IdleTimeout: 240 * time.Second,
+//          Dial: func () (redis.Conn, error) {
+//              c, err := redis.Dial("tcp", server)
+//              if err != nil {
+//                  return nil, err
+//              }
+//              if _, err := c.Do("AUTH", password); err != nil {
+//                  c.Close()
+//                  return nil, err
+//              }
+//              return c, err
+//          },
+//          TestOnBorrow: func(c redis.Conn, t time.Time) error {
+//              _, err := c.Do("PING")
+//              return err
+//          },
+//      }
+//  }
+//
+//  var (
+//      pool *redis.Pool
+//      redisServer = flag.String("redisServer", ":6379", "")
+//      redisPassword = flag.String("redisPassword", "", "")
+//  )
+//
+//  func main() {
+//      flag.Parse()
+//      pool = newPool(*redisServer, *redisPassword)
+//      ...
+//  }
+//
+// A request handler gets a connection from the pool and closes the connection
+// when the handler is done:
+//
+//  func serveHome(w http.ResponseWriter, r *http.Request) {
+//      conn := pool.Get()
+//      defer conn.Close()
+//      ....
+//  }
+//
+type Pool struct {
+
+	// Dial is an application supplied function for creating and configuring a
+	// connection
+	Dial func() (Conn, error)
+
+	// TestOnBorrow is an optional application supplied function for checking
+	// the health of an idle connection before the connection is used again by
+	// the application. Argument t is the time that the connection was returned
+	// to the pool. If the function returns an error, then the connection is
+	// closed.
+	TestOnBorrow func(c Conn, t time.Time) error
+
+	// Maximum number of idle connections in the pool.
+	MaxIdle int
+
+	// Maximum number of connections allocated by the pool at a given time.
+	// When zero, there is no limit on the number of connections in the pool.
+	MaxActive int
+
+	// Close connections after remaining idle for this duration. If the value
+	// is zero, then idle connections are not closed. Applications should set
+	// the timeout to a value less than the server's timeout.
+	IdleTimeout time.Duration
+
+	// If Wait is true and the pool is at the MaxIdle limit, then Get() waits
+	// for a connection to be returned to the pool before returning.
+	Wait bool
+
+	// mu protects fields defined below.
+	mu     sync.Mutex
+	cond   *sync.Cond
+	closed bool
+	active int
+
+	// Stack of idleConn with most recently used at the front.
+	idle list.List
+}
+
+type idleConn struct {
+	c Conn
+	t time.Time
+}
+
+// NewPool creates a new pool. This function is deprecated. Applications should
+// initialize the Pool fields directly as shown in example.
+func NewPool(newFn func() (Conn, error), maxIdle int) *Pool {
+	return &Pool{Dial: newFn, MaxIdle: maxIdle}
+}
+
+// Get gets a connection. The application must close the returned connection.
+// This method always returns a valid connection so that applications can defer
+// error handling to the first use of the connection. If there is an error
+// getting an underlying connection, then the connection Err, Do, Send, Flush
+// and Receive methods return that error.
+func (p *Pool) Get() Conn {
+	c, err := p.get()
+	if err != nil {
+		return errorConnection{err}
+	}
+	return &pooledConnection{p: p, c: c}
+}
+
+// ActiveCount returns the number of active connections in the pool.
+func (p *Pool) ActiveCount() int {
+	p.mu.Lock()
+	active := p.active
+	p.mu.Unlock()
+	return active
+}
+
+// Close releases the resources used by the pool.
+func (p *Pool) Close() error {
+	p.mu.Lock()
+	idle := p.idle
+	p.idle.Init()
+	p.closed = true
+	p.active -= idle.Len()
+	if p.cond != nil {
+		p.cond.Broadcast()
+	}
+	p.mu.Unlock()
+	for e := idle.Front(); e != nil; e = e.Next() {
+		e.Value.(idleConn).c.Close()
+	}
+	return nil
+}
+
+// release decrements the active count and signals waiters. The caller must
+// hold p.mu during the call.
+func (p *Pool) release() {
+	p.active -= 1
+	if p.cond != nil {
+		p.cond.Signal()
+	}
+}
+
+// get prunes stale connections and returns a connection from the idle list or
+// creates a new connection.
+func (p *Pool) get() (Conn, error) {
+	p.mu.Lock()
+
+	// Prune stale connections.
+
+	if timeout := p.IdleTimeout; timeout > 0 {
+		for i, n := 0, p.idle.Len(); i < n; i++ {
+			e := p.idle.Back()
+			if e == nil {
+				break
+			}
+			ic := e.Value.(idleConn)
+			if ic.t.Add(timeout).After(nowFunc()) {
+				break
+			}
+			p.idle.Remove(e)
+			p.release()
+			p.mu.Unlock()
+			ic.c.Close()
+			p.mu.Lock()
+		}
+	}
+
+	for {
+
+		// Get idle connection.
+
+		for i, n := 0, p.idle.Len(); i < n; i++ {
+			e := p.idle.Front()
+			if e == nil {
+				break
+			}
+			ic := e.Value.(idleConn)
+			p.idle.Remove(e)
+			test := p.TestOnBorrow
+			p.mu.Unlock()
+			if test == nil || test(ic.c, ic.t) == nil {
+				return ic.c, nil
+			}
+			ic.c.Close()
+			p.mu.Lock()
+			p.release()
+		}
+
+		// Check for pool closed before dialing a new connection.
+
+		if p.closed {
+			p.mu.Unlock()
+			return nil, errors.New("redigo: get on closed pool")
+		}
+
+		// Dial new connection if under limit.
+
+		if p.MaxActive == 0 || p.active < p.MaxActive {
+			dial := p.Dial
+			p.active += 1
+			p.mu.Unlock()
+			c, err := dial()
+			if err != nil {
+				p.mu.Lock()
+				p.release()
+				p.mu.Unlock()
+				c = nil
+			}
+			return c, err
+		}
+
+		if !p.Wait {
+			p.mu.Unlock()
+			return nil, ErrPoolExhausted
+		}
+
+		if p.cond == nil {
+			p.cond = sync.NewCond(&p.mu)
+		}
+		p.cond.Wait()
+	}
+}
+
+func (p *Pool) put(c Conn, forceClose bool) error {
+	err := c.Err()
+	p.mu.Lock()
+	if !p.closed && err == nil && !forceClose {
+		p.idle.PushFront(idleConn{t: nowFunc(), c: c})
+		if p.idle.Len() > p.MaxIdle {
+			c = p.idle.Remove(p.idle.Back()).(idleConn).c
+		} else {
+			c = nil
+		}
+	}
+
+	if c == nil {
+		if p.cond != nil {
+			p.cond.Signal()
+		}
+		p.mu.Unlock()
+		return nil
+	}
+
+	p.release()
+	p.mu.Unlock()
+	return c.Close()
+}
+
+type pooledConnection struct {
+	p     *Pool
+	c     Conn
+	state int
+}
+
+var (
+	sentinel     []byte
+	sentinelOnce sync.Once
+)
+
+func initSentinel() {
+	p := make([]byte, 64)
+	if _, err := rand.Read(p); err == nil {
+		sentinel = p
+	} else {
+		h := sha1.New()
+		io.WriteString(h, "Oops, rand failed. Use time instead.")
+		io.WriteString(h, strconv.FormatInt(time.Now().UnixNano(), 10))
+		sentinel = h.Sum(nil)
+	}
+}
+
+func (pc *pooledConnection) Close() error {
+	c := pc.c
+	if _, ok := c.(errorConnection); ok {
+		return nil
+	}
+	pc.c = errorConnection{errConnClosed}
+
+	if pc.state&internal.MultiState != 0 {
+		c.Send("DISCARD")
+		pc.state &^= (internal.MultiState | internal.WatchState)
+	} else if pc.state&internal.WatchState != 0 {
+		c.Send("UNWATCH")
+		pc.state &^= internal.WatchState
+	}
+	if pc.state&internal.SubscribeState != 0 {
+		c.Send("UNSUBSCRIBE")
+		c.Send("PUNSUBSCRIBE")
+		// To detect the end of the message stream, ask the server to echo
+		// a sentinel value and read until we see that value.
+		sentinelOnce.Do(initSentinel)
+		c.Send("ECHO", sentinel)
+		c.Flush()
+		for {
+			p, err := c.Receive()
+			if err != nil {
+				break
+			}
+			if p, ok := p.([]byte); ok && bytes.Equal(p, sentinel) {
+				pc.state &^= internal.SubscribeState
+				break
+			}
+		}
+	}
+	c.Do("")
+	pc.p.put(c, pc.state != 0)
+	return nil
+}
+
+func (pc *pooledConnection) Err() error {
+	return pc.c.Err()
+}
+
+func (pc *pooledConnection) Do(commandName string, args ...interface{}) (reply interface{}, err error) {
+	ci := internal.LookupCommandInfo(commandName)
+	pc.state = (pc.state | ci.Set) &^ ci.Clear
+	return pc.c.Do(commandName, args...)
+}
+
+func (pc *pooledConnection) Send(commandName string, args ...interface{}) error {
+	ci := internal.LookupCommandInfo(commandName)
+	pc.state = (pc.state | ci.Set) &^ ci.Clear
+	return pc.c.Send(commandName, args...)
+}
+
+func (pc *pooledConnection) Flush() error {
+	return pc.c.Flush()
+}
+
+func (pc *pooledConnection) Receive() (reply interface{}, err error) {
+	return pc.c.Receive()
+}
+
+type errorConnection struct{ err error }
+
+func (ec errorConnection) Do(string, ...interface{}) (interface{}, error) { return nil, ec.err }
+func (ec errorConnection) Send(string, ...interface{}) error              { return ec.err }
+func (ec errorConnection) Err() error                                     { return ec.err }
+func (ec errorConnection) Close() error                                   { return ec.err }
+func (ec errorConnection) Flush() error                                   { return ec.err }
+func (ec errorConnection) Receive() (interface{}, error)                  { return nil, ec.err }

Godeps/_workspace/src/github.com/garyburd/redigo/redis/pool_test.go

@@ -0,0 +1,674 @@
+// Copyright 2011 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis_test
+
+import (
+	"errors"
+	"io"
+	"reflect"
+	"sync"
+	"testing"
+	"time"
+
+	"github.com/garyburd/redigo/internal/redistest"
+	"github.com/garyburd/redigo/redis"
+)
+
+type poolTestConn struct {
+	d   *poolDialer
+	err error
+	redis.Conn
+}
+
+func (c *poolTestConn) Close() error { c.d.open -= 1; return nil }
+func (c *poolTestConn) Err() error   { return c.err }
+
+func (c *poolTestConn) Do(commandName string, args ...interface{}) (reply interface{}, err error) {
+	if commandName == "ERR" {
+		c.err = args[0].(error)
+		commandName = "PING"
+	}
+	if commandName != "" {
+		c.d.commands = append(c.d.commands, commandName)
+	}
+	return c.Conn.Do(commandName, args...)
+}
+
+func (c *poolTestConn) Send(commandName string, args ...interface{}) error {
+	c.d.commands = append(c.d.commands, commandName)
+	return c.Conn.Send(commandName, args...)
+}
+
+type poolDialer struct {
+	t        *testing.T
+	dialed   int
+	open     int
+	commands []string
+	dialErr  error
+}
+
+func (d *poolDialer) dial() (redis.Conn, error) {
+	d.dialed += 1
+	if d.dialErr != nil {
+		return nil, d.dialErr
+	}
+	c, err := redistest.Dial()
+	if err != nil {
+		return nil, err
+	}
+	d.open += 1
+	return &poolTestConn{d: d, Conn: c}, nil
+}
+
+func (d *poolDialer) check(message string, p *redis.Pool, dialed, open int) {
+	if d.dialed != dialed {
+		d.t.Errorf("%s: dialed=%d, want %d", message, d.dialed, dialed)
+	}
+	if d.open != open {
+		d.t.Errorf("%s: open=%d, want %d", message, d.open, open)
+	}
+	if active := p.ActiveCount(); active != open {
+		d.t.Errorf("%s: active=%d, want %d", message, active, open)
+	}
+}
+
+func TestPoolReuse(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle: 2,
+		Dial:    d.dial,
+	}
+
+	for i := 0; i < 10; i++ {
+		c1 := p.Get()
+		c1.Do("PING")
+		c2 := p.Get()
+		c2.Do("PING")
+		c1.Close()
+		c2.Close()
+	}
+
+	d.check("before close", p, 2, 2)
+	p.Close()
+	d.check("after close", p, 2, 0)
+}
+
+func TestPoolMaxIdle(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle: 2,
+		Dial:    d.dial,
+	}
+	for i := 0; i < 10; i++ {
+		c1 := p.Get()
+		c1.Do("PING")
+		c2 := p.Get()
+		c2.Do("PING")
+		c3 := p.Get()
+		c3.Do("PING")
+		c1.Close()
+		c2.Close()
+		c3.Close()
+	}
+	d.check("before close", p, 12, 2)
+	p.Close()
+	d.check("after close", p, 12, 0)
+}
+
+func TestPoolError(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle: 2,
+		Dial:    d.dial,
+	}
+
+	c := p.Get()
+	c.Do("ERR", io.EOF)
+	if c.Err() == nil {
+		t.Errorf("expected c.Err() != nil")
+	}
+	c.Close()
+
+	c = p.Get()
+	c.Do("ERR", io.EOF)
+	c.Close()
+
+	d.check(".", p, 2, 0)
+}
+
+func TestPoolClose(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle: 2,
+		Dial:    d.dial,
+	}
+
+	c1 := p.Get()
+	c1.Do("PING")
+	c2 := p.Get()
+	c2.Do("PING")
+	c3 := p.Get()
+	c3.Do("PING")
+
+	c1.Close()
+	if _, err := c1.Do("PING"); err == nil {
+		t.Errorf("expected error after connection closed")
+	}
+
+	c2.Close()
+	c2.Close()
+
+	p.Close()
+
+	d.check("after pool close", p, 3, 1)
+
+	if _, err := c1.Do("PING"); err == nil {
+		t.Errorf("expected error after connection and pool closed")
+	}
+
+	c3.Close()
+
+	d.check("after conn close", p, 3, 0)
+
+	c1 = p.Get()
+	if _, err := c1.Do("PING"); err == nil {
+		t.Errorf("expected error after pool closed")
+	}
+}
+
+func TestPoolTimeout(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:     2,
+		IdleTimeout: 300 * time.Second,
+		Dial:        d.dial,
+	}
+
+	now := time.Now()
+	redis.SetNowFunc(func() time.Time { return now })
+	defer redis.SetNowFunc(time.Now)
+
+	c := p.Get()
+	c.Do("PING")
+	c.Close()
+
+	d.check("1", p, 1, 1)
+
+	now = now.Add(p.IdleTimeout)
+
+	c = p.Get()
+	c.Do("PING")
+	c.Close()
+
+	d.check("2", p, 2, 1)
+
+	p.Close()
+}
+
+func TestPoolConcurrenSendReceive(t *testing.T) {
+	p := &redis.Pool{
+		Dial: redistest.Dial,
+	}
+	c := p.Get()
+	done := make(chan error, 1)
+	go func() {
+		_, err := c.Receive()
+		done <- err
+	}()
+	c.Send("PING")
+	c.Flush()
+	err := <-done
+	if err != nil {
+		t.Fatalf("Receive() returned error %v", err)
+	}
+	_, err = c.Do("")
+	if err != nil {
+		t.Fatalf("Do() returned error %v", err)
+	}
+	c.Close()
+	p.Close()
+}
+
+func TestPoolBorrowCheck(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:      2,
+		Dial:         d.dial,
+		TestOnBorrow: func(redis.Conn, time.Time) error { return redis.Error("BLAH") },
+	}
+
+	for i := 0; i < 10; i++ {
+		c := p.Get()
+		c.Do("PING")
+		c.Close()
+	}
+	d.check("1", p, 10, 1)
+	p.Close()
+}
+
+func TestPoolMaxActive(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   2,
+		MaxActive: 2,
+		Dial:      d.dial,
+	}
+	c1 := p.Get()
+	c1.Do("PING")
+	c2 := p.Get()
+	c2.Do("PING")
+
+	d.check("1", p, 2, 2)
+
+	c3 := p.Get()
+	if _, err := c3.Do("PING"); err != redis.ErrPoolExhausted {
+		t.Errorf("expected pool exhausted")
+	}
+
+	c3.Close()
+	d.check("2", p, 2, 2)
+	c2.Close()
+	d.check("3", p, 2, 2)
+
+	c3 = p.Get()
+	if _, err := c3.Do("PING"); err != nil {
+		t.Errorf("expected good channel, err=%v", err)
+	}
+	c3.Close()
+
+	d.check("4", p, 2, 2)
+	p.Close()
+}
+
+func TestPoolMonitorCleanup(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   2,
+		MaxActive: 2,
+		Dial:      d.dial,
+	}
+	c := p.Get()
+	c.Send("MONITOR")
+	c.Close()
+
+	d.check("", p, 1, 0)
+	p.Close()
+}
+
+func TestPoolPubSubCleanup(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   2,
+		MaxActive: 2,
+		Dial:      d.dial,
+	}
+
+	c := p.Get()
+	c.Send("SUBSCRIBE", "x")
+	c.Close()
+
+	want := []string{"SUBSCRIBE", "UNSUBSCRIBE", "PUNSUBSCRIBE", "ECHO"}
+	if !reflect.DeepEqual(d.commands, want) {
+		t.Errorf("got commands %v, want %v", d.commands, want)
+	}
+	d.commands = nil
+
+	c = p.Get()
+	c.Send("PSUBSCRIBE", "x*")
+	c.Close()
+
+	want = []string{"PSUBSCRIBE", "UNSUBSCRIBE", "PUNSUBSCRIBE", "ECHO"}
+	if !reflect.DeepEqual(d.commands, want) {
+		t.Errorf("got commands %v, want %v", d.commands, want)
+	}
+	d.commands = nil
+
+	p.Close()
+}
+
+func TestPoolTransactionCleanup(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   2,
+		MaxActive: 2,
+		Dial:      d.dial,
+	}
+
+	c := p.Get()
+	c.Do("WATCH", "key")
+	c.Do("PING")
+	c.Close()
+
+	want := []string{"WATCH", "PING", "UNWATCH"}
+	if !reflect.DeepEqual(d.commands, want) {
+		t.Errorf("got commands %v, want %v", d.commands, want)
+	}
+	d.commands = nil
+
+	c = p.Get()
+	c.Do("WATCH", "key")
+	c.Do("UNWATCH")
+	c.Do("PING")
+	c.Close()
+
+	want = []string{"WATCH", "UNWATCH", "PING"}
+	if !reflect.DeepEqual(d.commands, want) {
+		t.Errorf("got commands %v, want %v", d.commands, want)
+	}
+	d.commands = nil
+
+	c = p.Get()
+	c.Do("WATCH", "key")
+	c.Do("MULTI")
+	c.Do("PING")
+	c.Close()
+
+	want = []string{"WATCH", "MULTI", "PING", "DISCARD"}
+	if !reflect.DeepEqual(d.commands, want) {
+		t.Errorf("got commands %v, want %v", d.commands, want)
+	}
+	d.commands = nil
+
+	c = p.Get()
+	c.Do("WATCH", "key")
+	c.Do("MULTI")
+	c.Do("DISCARD")
+	c.Do("PING")
+	c.Close()
+
+	want = []string{"WATCH", "MULTI", "DISCARD", "PING"}
+	if !reflect.DeepEqual(d.commands, want) {
+		t.Errorf("got commands %v, want %v", d.commands, want)
+	}
+	d.commands = nil
+
+	c = p.Get()
+	c.Do("WATCH", "key")
+	c.Do("MULTI")
+	c.Do("EXEC")
+	c.Do("PING")
+	c.Close()
+
+	want = []string{"WATCH", "MULTI", "EXEC", "PING"}
+	if !reflect.DeepEqual(d.commands, want) {
+		t.Errorf("got commands %v, want %v", d.commands, want)
+	}
+	d.commands = nil
+
+	p.Close()
+}
+
+func startGoroutines(p *redis.Pool, cmd string, args ...interface{}) chan error {
+	errs := make(chan error, 10)
+	for i := 0; i < cap(errs); i++ {
+		go func() {
+			c := p.Get()
+			_, err := c.Do(cmd, args...)
+			errs <- err
+			c.Close()
+		}()
+	}
+
+	// Wait for goroutines to block.
+	time.Sleep(time.Second / 4)
+
+	return errs
+}
+
+func TestWaitPool(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   1,
+		MaxActive: 1,
+		Dial:      d.dial,
+		Wait:      true,
+	}
+	defer p.Close()
+	c := p.Get()
+	errs := startGoroutines(p, "PING")
+	d.check("before close", p, 1, 1)
+	c.Close()
+	timeout := time.After(2 * time.Second)
+	for i := 0; i < cap(errs); i++ {
+		select {
+		case err := <-errs:
+			if err != nil {
+				t.Fatal(err)
+			}
+		case <-timeout:
+			t.Fatalf("timeout waiting for blocked goroutine %d", i)
+		}
+	}
+	d.check("done", p, 1, 1)
+}
+
+func TestWaitPoolClose(t *testing.T) {
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   1,
+		MaxActive: 1,
+		Dial:      d.dial,
+		Wait:      true,
+	}
+	c := p.Get()
+	if _, err := c.Do("PING"); err != nil {
+		t.Fatal(err)
+	}
+	errs := startGoroutines(p, "PING")
+	d.check("before close", p, 1, 1)
+	p.Close()
+	timeout := time.After(2 * time.Second)
+	for i := 0; i < cap(errs); i++ {
+		select {
+		case err := <-errs:
+			switch err {
+			case nil:
+				t.Fatal("blocked goroutine did not get error")
+			case redis.ErrPoolExhausted:
+				t.Fatal("blocked goroutine got pool exhausted error")
+			}
+		case <-timeout:
+			t.Fatal("timeout waiting for blocked goroutine")
+		}
+	}
+	c.Close()
+	d.check("done", p, 1, 0)
+}
+
+func TestWaitPoolCommandError(t *testing.T) {
+	testErr := errors.New("test")
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   1,
+		MaxActive: 1,
+		Dial:      d.dial,
+		Wait:      true,
+	}
+	defer p.Close()
+	c := p.Get()
+	errs := startGoroutines(p, "ERR", testErr)
+	d.check("before close", p, 1, 1)
+	c.Close()
+	timeout := time.After(2 * time.Second)
+	for i := 0; i < cap(errs); i++ {
+		select {
+		case err := <-errs:
+			if err != nil {
+				t.Fatal(err)
+			}
+		case <-timeout:
+			t.Fatalf("timeout waiting for blocked goroutine %d", i)
+		}
+	}
+	d.check("done", p, cap(errs), 0)
+}
+
+func TestWaitPoolDialError(t *testing.T) {
+	testErr := errors.New("test")
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   1,
+		MaxActive: 1,
+		Dial:      d.dial,
+		Wait:      true,
+	}
+	defer p.Close()
+	c := p.Get()
+	errs := startGoroutines(p, "ERR", testErr)
+	d.check("before close", p, 1, 1)
+
+	d.dialErr = errors.New("dial")
+	c.Close()
+
+	nilCount := 0
+	errCount := 0
+	timeout := time.After(2 * time.Second)
+	for i := 0; i < cap(errs); i++ {
+		select {
+		case err := <-errs:
+			switch err {
+			case nil:
+				nilCount++
+			case d.dialErr:
+				errCount++
+			default:
+				t.Fatalf("expected dial error or nil, got %v", err)
+			}
+		case <-timeout:
+			t.Fatalf("timeout waiting for blocked goroutine %d", i)
+		}
+	}
+	if nilCount != 1 {
+		t.Errorf("expected one nil error, got %d", nilCount)
+	}
+	if errCount != cap(errs)-1 {
+		t.Errorf("expected %d dial erors, got %d", cap(errs)-1, errCount)
+	}
+	d.check("done", p, cap(errs), 0)
+}
+
+// Borrowing requires us to iterate over the idle connections, unlock the pool,
+// and perform a blocking operation to check the connection still works. If
+// TestOnBorrow fails, we must reacquire the lock and continue iteration. This
+// test ensures that iteration will work correctly if multiple threads are
+// iterating simultaneously.
+func TestLocking_TestOnBorrowFails_PoolDoesntCrash(t *testing.T) {
+	count := 100
+
+	// First we'll Create a pool where the pilfering of idle connections fails.
+	d := poolDialer{t: t}
+	p := &redis.Pool{
+		MaxIdle:   count,
+		MaxActive: count,
+		Dial:      d.dial,
+		TestOnBorrow: func(c redis.Conn, t time.Time) error {
+			return errors.New("No way back into the real world.")
+		},
+	}
+	defer p.Close()
+
+	// Fill the pool with idle connections.
+	b1 := sync.WaitGroup{}
+	b1.Add(count)
+	b2 := sync.WaitGroup{}
+	b2.Add(count)
+	for i := 0; i < count; i++ {
+		go func() {
+			c := p.Get()
+			if c.Err() != nil {
+				t.Errorf("pool get failed: %v", c.Err())
+			}
+			b1.Done()
+			b1.Wait()
+			c.Close()
+			b2.Done()
+		}()
+	}
+	b2.Wait()
+	if d.dialed != count {
+		t.Errorf("Expected %d dials, got %d", count, d.dialed)
+	}
+
+	// Spawn a bunch of goroutines to thrash the pool.
+	b2.Add(count)
+	for i := 0; i < count; i++ {
+		go func() {
+			c := p.Get()
+			if c.Err() != nil {
+				t.Errorf("pool get failed: %v", c.Err())
+			}
+			c.Close()
+			b2.Done()
+		}()
+	}
+	b2.Wait()
+	if d.dialed != count*2 {
+		t.Errorf("Expected %d dials, got %d", count*2, d.dialed)
+	}
+}
+
+func BenchmarkPoolGet(b *testing.B) {
+	b.StopTimer()
+	p := redis.Pool{Dial: redistest.Dial, MaxIdle: 2}
+	c := p.Get()
+	if err := c.Err(); err != nil {
+		b.Fatal(err)
+	}
+	c.Close()
+	defer p.Close()
+	b.StartTimer()
+	for i := 0; i < b.N; i++ {
+		c = p.Get()
+		c.Close()
+	}
+}
+
+func BenchmarkPoolGetErr(b *testing.B) {
+	b.StopTimer()
+	p := redis.Pool{Dial: redistest.Dial, MaxIdle: 2}
+	c := p.Get()
+	if err := c.Err(); err != nil {
+		b.Fatal(err)
+	}
+	c.Close()
+	defer p.Close()
+	b.StartTimer()
+	for i := 0; i < b.N; i++ {
+		c = p.Get()
+		if err := c.Err(); err != nil {
+			b.Fatal(err)
+		}
+		c.Close()
+	}
+}
+
+func BenchmarkPoolGetPing(b *testing.B) {
+	b.StopTimer()
+	p := redis.Pool{Dial: redistest.Dial, MaxIdle: 2}
+	c := p.Get()
+	if err := c.Err(); err != nil {
+		b.Fatal(err)
+	}
+	c.Close()
+	defer p.Close()
+	b.StartTimer()
+	for i := 0; i < b.N; i++ {
+		c = p.Get()
+		if _, err := c.Do("PING"); err != nil {
+			b.Fatal(err)
+		}
+		c.Close()
+	}
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/pubsub.go

@@ -0,0 +1,129 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"errors"
+)
+
+// Subscription represents a subscribe or unsubscribe notification.
+type Subscription struct {
+
+	// Kind is "subscribe", "unsubscribe", "psubscribe" or "punsubscribe"
+	Kind string
+
+	// The channel that was changed.
+	Channel string
+
+	// The current number of subscriptions for connection.
+	Count int
+}
+
+// Message represents a message notification.
+type Message struct {
+
+	// The originating channel.
+	Channel string
+
+	// The message data.
+	Data []byte
+}
+
+// PMessage represents a pmessage notification.
+type PMessage struct {
+
+	// The matched pattern.
+	Pattern string
+
+	// The originating channel.
+	Channel string
+
+	// The message data.
+	Data []byte
+}
+
+// PubSubConn wraps a Conn with convenience methods for subscribers.
+type PubSubConn struct {
+	Conn Conn
+}
+
+// Close closes the connection.
+func (c PubSubConn) Close() error {
+	return c.Conn.Close()
+}
+
+// Subscribe subscribes the connection to the specified channels.
+func (c PubSubConn) Subscribe(channel ...interface{}) error {
+	c.Conn.Send("SUBSCRIBE", channel...)
+	return c.Conn.Flush()
+}
+
+// PSubscribe subscribes the connection to the given patterns.
+func (c PubSubConn) PSubscribe(channel ...interface{}) error {
+	c.Conn.Send("PSUBSCRIBE", channel...)
+	return c.Conn.Flush()
+}
+
+// Unsubscribe unsubscribes the connection from the given channels, or from all
+// of them if none is given.
+func (c PubSubConn) Unsubscribe(channel ...interface{}) error {
+	c.Conn.Send("UNSUBSCRIBE", channel...)
+	return c.Conn.Flush()
+}
+
+// PUnsubscribe unsubscribes the connection from the given patterns, or from all
+// of them if none is given.
+func (c PubSubConn) PUnsubscribe(channel ...interface{}) error {
+	c.Conn.Send("PUNSUBSCRIBE", channel...)
+	return c.Conn.Flush()
+}
+
+// Receive returns a pushed message as a Subscription, Message, PMessage or
+// error. The return value is intended to be used directly in a type switch as
+// illustrated in the PubSubConn example.
+func (c PubSubConn) Receive() interface{} {
+	reply, err := Values(c.Conn.Receive())
+	if err != nil {
+		return err
+	}
+
+	var kind string
+	reply, err = Scan(reply, &kind)
+	if err != nil {
+		return err
+	}
+
+	switch kind {
+	case "message":
+		var m Message
+		if _, err := Scan(reply, &m.Channel, &m.Data); err != nil {
+			return err
+		}
+		return m
+	case "pmessage":
+		var pm PMessage
+		if _, err := Scan(reply, &pm.Pattern, &pm.Channel, &pm.Data); err != nil {
+			return err
+		}
+		return pm
+	case "subscribe", "psubscribe", "unsubscribe", "punsubscribe":
+		s := Subscription{Kind: kind}
+		if _, err := Scan(reply, &s.Channel, &s.Count); err != nil {
+			return err
+		}
+		return s
+	}
+	return errors.New("redigo: unknown pubsub notification")
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/pubsub_test.go

@@ -0,0 +1,143 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis_test
+
+import (
+	"fmt"
+	"net"
+	"reflect"
+	"sync"
+	"testing"
+	"time"
+
+	"github.com/garyburd/redigo/internal/redistest"
+	"github.com/garyburd/redigo/redis"
+)
+
+func publish(channel, value interface{}) {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+	c.Do("PUBLISH", channel, value)
+}
+
+// Applications can receive pushed messages from one goroutine and manage subscriptions from another goroutine.
+func ExamplePubSubConn() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+	var wg sync.WaitGroup
+	wg.Add(2)
+
+	psc := redis.PubSubConn{Conn: c}
+
+	// This goroutine receives and prints pushed notifications from the server.
+	// The goroutine exits when the connection is unsubscribed from all
+	// channels or there is an error.
+	go func() {
+		defer wg.Done()
+		for {
+			switch n := psc.Receive().(type) {
+			case redis.Message:
+				fmt.Printf("Message: %s %s\n", n.Channel, n.Data)
+			case redis.PMessage:
+				fmt.Printf("PMessage: %s %s %s\n", n.Pattern, n.Channel, n.Data)
+			case redis.Subscription:
+				fmt.Printf("Subscription: %s %s %d\n", n.Kind, n.Channel, n.Count)
+				if n.Count == 0 {
+					return
+				}
+			case error:
+				fmt.Printf("error: %v\n", n)
+				return
+			}
+		}
+	}()
+
+	// This goroutine manages subscriptions for the connection.
+	go func() {
+		defer wg.Done()
+
+		psc.Subscribe("example")
+		psc.PSubscribe("p*")
+
+		// The following function calls publish a message using another
+		// connection to the Redis server.
+		publish("example", "hello")
+		publish("example", "world")
+		publish("pexample", "foo")
+		publish("pexample", "bar")
+
+		// Unsubscribe from all connections. This will cause the receiving
+		// goroutine to exit.
+		psc.Unsubscribe()
+		psc.PUnsubscribe()
+	}()
+
+	wg.Wait()
+
+	// Output:
+	// Subscription: subscribe example 1
+	// Subscription: psubscribe p* 2
+	// Message: example hello
+	// Message: example world
+	// PMessage: p* pexample foo
+	// PMessage: p* pexample bar
+	// Subscription: unsubscribe example 1
+	// Subscription: punsubscribe p* 0
+}
+
+func expectPushed(t *testing.T, c redis.PubSubConn, message string, expected interface{}) {
+	actual := c.Receive()
+	if !reflect.DeepEqual(actual, expected) {
+		t.Errorf("%s = %v, want %v", message, actual, expected)
+	}
+}
+
+func TestPushed(t *testing.T) {
+	pc, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer pc.Close()
+
+	nc, err := net.Dial("tcp", ":6379")
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer nc.Close()
+	nc.SetReadDeadline(time.Now().Add(4 * time.Second))
+
+	c := redis.PubSubConn{Conn: redis.NewConn(nc, 0, 0)}
+
+	c.Subscribe("c1")
+	expectPushed(t, c, "Subscribe(c1)", redis.Subscription{Kind: "subscribe", Channel: "c1", Count: 1})
+	c.Subscribe("c2")
+	expectPushed(t, c, "Subscribe(c2)", redis.Subscription{Kind: "subscribe", Channel: "c2", Count: 2})
+	c.PSubscribe("p1")
+	expectPushed(t, c, "PSubscribe(p1)", redis.Subscription{Kind: "psubscribe", Channel: "p1", Count: 3})
+	c.PSubscribe("p2")
+	expectPushed(t, c, "PSubscribe(p2)", redis.Subscription{Kind: "psubscribe", Channel: "p2", Count: 4})
+	c.PUnsubscribe()
+	expectPushed(t, c, "Punsubscribe(p1)", redis.Subscription{Kind: "punsubscribe", Channel: "p1", Count: 3})
+	expectPushed(t, c, "Punsubscribe()", redis.Subscription{Kind: "punsubscribe", Channel: "p2", Count: 2})
+
+	pc.Do("PUBLISH", "c1", "hello")
+	expectPushed(t, c, "PUBLISH c1 hello", redis.Message{Channel: "c1", Data: []byte("hello")})
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/redis.go

@@ -0,0 +1,44 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+// Error represents an error returned in a command reply.
+type Error string
+
+func (err Error) Error() string { return string(err) }
+
+// Conn represents a connection to a Redis server.
+type Conn interface {
+	// Close closes the connection.
+	Close() error
+
+	// Err returns a non-nil value if the connection is broken. The returned
+	// value is either the first non-nil value returned from the underlying
+	// network connection or a protocol parsing error. Applications should
+	// close broken connections.
+	Err() error
+
+	// Do sends a command to the server and returns the received reply.
+	Do(commandName string, args ...interface{}) (reply interface{}, err error)
+
+	// Send writes the command to the client's output buffer.
+	Send(commandName string, args ...interface{}) error
+
+	// Flush flushes the output buffer to the Redis server.
+	Flush() error
+
+	// Receive receives a single reply from the Redis server
+	Receive() (reply interface{}, err error)
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/reply.go

@@ -0,0 +1,312 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"errors"
+	"fmt"
+	"strconv"
+)
+
+// ErrNil indicates that a reply value is nil.
+var ErrNil = errors.New("redigo: nil returned")
+
+// Int is a helper that converts a command reply to an integer. If err is not
+// equal to nil, then Int returns 0, err. Otherwise, Int converts the
+// reply to an int as follows:
+//
+//  Reply type    Result
+//  integer       int(reply), nil
+//  bulk string   parsed reply, nil
+//  nil           0, ErrNil
+//  other         0, error
+func Int(reply interface{}, err error) (int, error) {
+	if err != nil {
+		return 0, err
+	}
+	switch reply := reply.(type) {
+	case int64:
+		x := int(reply)
+		if int64(x) != reply {
+			return 0, strconv.ErrRange
+		}
+		return x, nil
+	case []byte:
+		n, err := strconv.ParseInt(string(reply), 10, 0)
+		return int(n), err
+	case nil:
+		return 0, ErrNil
+	case Error:
+		return 0, reply
+	}
+	return 0, fmt.Errorf("redigo: unexpected type for Int, got type %T", reply)
+}
+
+// Int64 is a helper that converts a command reply to 64 bit integer. If err is
+// not equal to nil, then Int returns 0, err. Otherwise, Int64 converts the
+// reply to an int64 as follows:
+//
+//  Reply type    Result
+//  integer       reply, nil
+//  bulk string   parsed reply, nil
+//  nil           0, ErrNil
+//  other         0, error
+func Int64(reply interface{}, err error) (int64, error) {
+	if err != nil {
+		return 0, err
+	}
+	switch reply := reply.(type) {
+	case int64:
+		return reply, nil
+	case []byte:
+		n, err := strconv.ParseInt(string(reply), 10, 64)
+		return n, err
+	case nil:
+		return 0, ErrNil
+	case Error:
+		return 0, reply
+	}
+	return 0, fmt.Errorf("redigo: unexpected type for Int64, got type %T", reply)
+}
+
+var errNegativeInt = errors.New("redigo: unexpected value for Uint64")
+
+// Uint64 is a helper that converts a command reply to 64 bit integer. If err is
+// not equal to nil, then Int returns 0, err. Otherwise, Int64 converts the
+// reply to an int64 as follows:
+//
+//  Reply type    Result
+//  integer       reply, nil
+//  bulk string   parsed reply, nil
+//  nil           0, ErrNil
+//  other         0, error
+func Uint64(reply interface{}, err error) (uint64, error) {
+	if err != nil {
+		return 0, err
+	}
+	switch reply := reply.(type) {
+	case int64:
+		if reply < 0 {
+			return 0, errNegativeInt
+		}
+		return uint64(reply), nil
+	case []byte:
+		n, err := strconv.ParseUint(string(reply), 10, 64)
+		return n, err
+	case nil:
+		return 0, ErrNil
+	case Error:
+		return 0, reply
+	}
+	return 0, fmt.Errorf("redigo: unexpected type for Uint64, got type %T", reply)
+}
+
+// Float64 is a helper that converts a command reply to 64 bit float. If err is
+// not equal to nil, then Float64 returns 0, err. Otherwise, Float64 converts
+// the reply to an int as follows:
+//
+//  Reply type    Result
+//  bulk string   parsed reply, nil
+//  nil           0, ErrNil
+//  other         0, error
+func Float64(reply interface{}, err error) (float64, error) {
+	if err != nil {
+		return 0, err
+	}
+	switch reply := reply.(type) {
+	case []byte:
+		n, err := strconv.ParseFloat(string(reply), 64)
+		return n, err
+	case nil:
+		return 0, ErrNil
+	case Error:
+		return 0, reply
+	}
+	return 0, fmt.Errorf("redigo: unexpected type for Float64, got type %T", reply)
+}
+
+// String is a helper that converts a command reply to a string. If err is not
+// equal to nil, then String returns "", err. Otherwise String converts the
+// reply to a string as follows:
+//
+//  Reply type      Result
+//  bulk string     string(reply), nil
+//  simple string   reply, nil
+//  nil             "",  ErrNil
+//  other           "",  error
+func String(reply interface{}, err error) (string, error) {
+	if err != nil {
+		return "", err
+	}
+	switch reply := reply.(type) {
+	case []byte:
+		return string(reply), nil
+	case string:
+		return reply, nil
+	case nil:
+		return "", ErrNil
+	case Error:
+		return "", reply
+	}
+	return "", fmt.Errorf("redigo: unexpected type for String, got type %T", reply)
+}
+
+// Bytes is a helper that converts a command reply to a slice of bytes. If err
+// is not equal to nil, then Bytes returns nil, err. Otherwise Bytes converts
+// the reply to a slice of bytes as follows:
+//
+//  Reply type      Result
+//  bulk string     reply, nil
+//  simple string   []byte(reply), nil
+//  nil             nil, ErrNil
+//  other           nil, error
+func Bytes(reply interface{}, err error) ([]byte, error) {
+	if err != nil {
+		return nil, err
+	}
+	switch reply := reply.(type) {
+	case []byte:
+		return reply, nil
+	case string:
+		return []byte(reply), nil
+	case nil:
+		return nil, ErrNil
+	case Error:
+		return nil, reply
+	}
+	return nil, fmt.Errorf("redigo: unexpected type for Bytes, got type %T", reply)
+}
+
+// Bool is a helper that converts a command reply to a boolean. If err is not
+// equal to nil, then Bool returns false, err. Otherwise Bool converts the
+// reply to boolean as follows:
+//
+//  Reply type      Result
+//  integer         value != 0, nil
+//  bulk string     strconv.ParseBool(reply)
+//  nil             false, ErrNil
+//  other           false, error
+func Bool(reply interface{}, err error) (bool, error) {
+	if err != nil {
+		return false, err
+	}
+	switch reply := reply.(type) {
+	case int64:
+		return reply != 0, nil
+	case []byte:
+		return strconv.ParseBool(string(reply))
+	case nil:
+		return false, ErrNil
+	case Error:
+		return false, reply
+	}
+	return false, fmt.Errorf("redigo: unexpected type for Bool, got type %T", reply)
+}
+
+// MultiBulk is deprecated. Use Values.
+func MultiBulk(reply interface{}, err error) ([]interface{}, error) { return Values(reply, err) }
+
+// Values is a helper that converts an array command reply to a []interface{}.
+// If err is not equal to nil, then Values returns nil, err. Otherwise, Values
+// converts the reply as follows:
+//
+//  Reply type      Result
+//  array           reply, nil
+//  nil             nil, ErrNil
+//  other           nil, error
+func Values(reply interface{}, err error) ([]interface{}, error) {
+	if err != nil {
+		return nil, err
+	}
+	switch reply := reply.(type) {
+	case []interface{}:
+		return reply, nil
+	case nil:
+		return nil, ErrNil
+	case Error:
+		return nil, reply
+	}
+	return nil, fmt.Errorf("redigo: unexpected type for Values, got type %T", reply)
+}
+
+// Strings is a helper that converts an array command reply to a []string. If
+// err is not equal to nil, then Strings returns nil, err. Nil array items are
+// converted to "" in the output slice. Strings returns an error if an array
+// item is not a bulk string or nil.
+func Strings(reply interface{}, err error) ([]string, error) {
+	if err != nil {
+		return nil, err
+	}
+	switch reply := reply.(type) {
+	case []interface{}:
+		result := make([]string, len(reply))
+		for i := range reply {
+			if reply[i] == nil {
+				continue
+			}
+			p, ok := reply[i].([]byte)
+			if !ok {
+				return nil, fmt.Errorf("redigo: unexpected element type for Strings, got type %T", reply[i])
+			}
+			result[i] = string(p)
+		}
+		return result, nil
+	case nil:
+		return nil, ErrNil
+	case Error:
+		return nil, reply
+	}
+	return nil, fmt.Errorf("redigo: unexpected type for Strings, got type %T", reply)
+}
+
+// Ints is a helper that converts an array command reply to a []int. If
+// err is not equal to nil, then Ints returns nil, err.
+func Ints(reply interface{}, err error) ([]int, error) {
+	var ints []int
+	if reply == nil {
+		return ints, ErrNil
+	}
+	values, err := Values(reply, err)
+	if err != nil {
+		return ints, err
+	}
+	if err := ScanSlice(values, &ints); err != nil {
+		return ints, err
+	}
+	return ints, nil
+}
+
+// StringMap is a helper that converts an array of strings (alternating key, value)
+// into a map[string]string. The HGETALL and CONFIG GET commands return replies in this format.
+// Requires an even number of values in result.
+func StringMap(result interface{}, err error) (map[string]string, error) {
+	values, err := Values(result, err)
+	if err != nil {
+		return nil, err
+	}
+	if len(values)%2 != 0 {
+		return nil, errors.New("redigo: StringMap expects even number of values result")
+	}
+	m := make(map[string]string, len(values)/2)
+	for i := 0; i < len(values); i += 2 {
+		key, okKey := values[i].([]byte)
+		value, okValue := values[i+1].([]byte)
+		if !okKey || !okValue {
+			return nil, errors.New("redigo: ScanMap key not a bulk string value")
+		}
+		m[string(key)] = string(value)
+	}
+	return m, nil
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/reply_test.go

@@ -0,0 +1,166 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis_test
+
+import (
+	"fmt"
+	"reflect"
+	"testing"
+
+	"github.com/garyburd/redigo/internal/redistest"
+	"github.com/garyburd/redigo/redis"
+)
+
+type valueError struct {
+	v   interface{}
+	err error
+}
+
+func ve(v interface{}, err error) valueError {
+	return valueError{v, err}
+}
+
+var replyTests = []struct {
+	name     interface{}
+	actual   valueError
+	expected valueError
+}{
+	{
+		"ints([v1, v2])",
+		ve(redis.Ints([]interface{}{[]byte("4"), []byte("5")}, nil)),
+		ve([]int{4, 5}, nil),
+	},
+	{
+		"ints(nil)",
+		ve(redis.Ints(nil, nil)),
+		ve([]int(nil), redis.ErrNil),
+	},
+	{
+		"strings([v1, v2])",
+		ve(redis.Strings([]interface{}{[]byte("v1"), []byte("v2")}, nil)),
+		ve([]string{"v1", "v2"}, nil),
+	},
+	{
+		"strings(nil)",
+		ve(redis.Strings(nil, nil)),
+		ve([]string(nil), redis.ErrNil),
+	},
+	{
+		"values([v1, v2])",
+		ve(redis.Values([]interface{}{[]byte("v1"), []byte("v2")}, nil)),
+		ve([]interface{}{[]byte("v1"), []byte("v2")}, nil),
+	},
+	{
+		"values(nil)",
+		ve(redis.Values(nil, nil)),
+		ve([]interface{}(nil), redis.ErrNil),
+	},
+	{
+		"float64(1.0)",
+		ve(redis.Float64([]byte("1.0"), nil)),
+		ve(float64(1.0), nil),
+	},
+	{
+		"float64(nil)",
+		ve(redis.Float64(nil, nil)),
+		ve(float64(0.0), redis.ErrNil),
+	},
+	{
+		"uint64(1)",
+		ve(redis.Uint64(int64(1), nil)),
+		ve(uint64(1), nil),
+	},
+	{
+		"uint64(-1)",
+		ve(redis.Uint64(int64(-1), nil)),
+		ve(uint64(0), redis.ErrNegativeInt),
+	},
+}
+
+func TestReply(t *testing.T) {
+	for _, rt := range replyTests {
+		if rt.actual.err != rt.expected.err {
+			t.Errorf("%s returned err %v, want %v", rt.name, rt.actual.err, rt.expected.err)
+			continue
+		}
+		if !reflect.DeepEqual(rt.actual.v, rt.expected.v) {
+			t.Errorf("%s=%+v, want %+v", rt.name, rt.actual.v, rt.expected.v)
+		}
+	}
+}
+
+// dial wraps DialTestDB() with a more suitable function name for examples.
+func dial() (redis.Conn, error) {
+	return redistest.Dial()
+}
+
+func ExampleBool() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+
+	c.Do("SET", "foo", 1)
+	exists, _ := redis.Bool(c.Do("EXISTS", "foo"))
+	fmt.Printf("%#v\n", exists)
+	// Output:
+	// true
+}
+
+func ExampleInt() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+
+	c.Do("SET", "k1", 1)
+	n, _ := redis.Int(c.Do("GET", "k1"))
+	fmt.Printf("%#v\n", n)
+	n, _ = redis.Int(c.Do("INCR", "k1"))
+	fmt.Printf("%#v\n", n)
+	// Output:
+	// 1
+	// 2
+}
+
+func ExampleInts() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+
+	c.Do("SADD", "set_with_integers", 4, 5, 6)
+	ints, _ := redis.Ints(c.Do("SMEMBERS", "set_with_integers"))
+	fmt.Printf("%#v\n", ints)
+	// Output:
+	// []int{4, 5, 6}
+}
+
+func ExampleString() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+
+	c.Do("SET", "hello", "world")
+	s, err := redis.String(c.Do("GET", "hello"))
+	fmt.Printf("%#v\n", s)
+	// Output:
+	// "world"
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/scan.go

@@ -0,0 +1,513 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"errors"
+	"fmt"
+	"reflect"
+	"strconv"
+	"strings"
+	"sync"
+)
+
+func ensureLen(d reflect.Value, n int) {
+	if n > d.Cap() {
+		d.Set(reflect.MakeSlice(d.Type(), n, n))
+	} else {
+		d.SetLen(n)
+	}
+}
+
+func cannotConvert(d reflect.Value, s interface{}) error {
+	return fmt.Errorf("redigo: Scan cannot convert from %s to %s",
+		reflect.TypeOf(s), d.Type())
+}
+
+func convertAssignBytes(d reflect.Value, s []byte) (err error) {
+	switch d.Type().Kind() {
+	case reflect.Float32, reflect.Float64:
+		var x float64
+		x, err = strconv.ParseFloat(string(s), d.Type().Bits())
+		d.SetFloat(x)
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+		var x int64
+		x, err = strconv.ParseInt(string(s), 10, d.Type().Bits())
+		d.SetInt(x)
+	case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64:
+		var x uint64
+		x, err = strconv.ParseUint(string(s), 10, d.Type().Bits())
+		d.SetUint(x)
+	case reflect.Bool:
+		var x bool
+		x, err = strconv.ParseBool(string(s))
+		d.SetBool(x)
+	case reflect.String:
+		d.SetString(string(s))
+	case reflect.Slice:
+		if d.Type().Elem().Kind() != reflect.Uint8 {
+			err = cannotConvert(d, s)
+		} else {
+			d.SetBytes(s)
+		}
+	default:
+		err = cannotConvert(d, s)
+	}
+	return
+}
+
+func convertAssignInt(d reflect.Value, s int64) (err error) {
+	switch d.Type().Kind() {
+	case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
+		d.SetInt(s)
+		if d.Int() != s {
+			err = strconv.ErrRange
+			d.SetInt(0)
+		}
+	case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64:
+		if s < 0 {
+			err = strconv.ErrRange
+		} else {
+			x := uint64(s)
+			d.SetUint(x)
+			if d.Uint() != x {
+				err = strconv.ErrRange
+				d.SetUint(0)
+			}
+		}
+	case reflect.Bool:
+		d.SetBool(s != 0)
+	default:
+		err = cannotConvert(d, s)
+	}
+	return
+}
+
+func convertAssignValue(d reflect.Value, s interface{}) (err error) {
+	switch s := s.(type) {
+	case []byte:
+		err = convertAssignBytes(d, s)
+	case int64:
+		err = convertAssignInt(d, s)
+	default:
+		err = cannotConvert(d, s)
+	}
+	return err
+}
+
+func convertAssignValues(d reflect.Value, s []interface{}) error {
+	if d.Type().Kind() != reflect.Slice {
+		return cannotConvert(d, s)
+	}
+	ensureLen(d, len(s))
+	for i := 0; i < len(s); i++ {
+		if err := convertAssignValue(d.Index(i), s[i]); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func convertAssign(d interface{}, s interface{}) (err error) {
+	// Handle the most common destination types using type switches and
+	// fall back to reflection for all other types.
+	switch s := s.(type) {
+	case nil:
+		// ingore
+	case []byte:
+		switch d := d.(type) {
+		case *string:
+			*d = string(s)
+		case *int:
+			*d, err = strconv.Atoi(string(s))
+		case *bool:
+			*d, err = strconv.ParseBool(string(s))
+		case *[]byte:
+			*d = s
+		case *interface{}:
+			*d = s
+		case nil:
+			// skip value
+		default:
+			if d := reflect.ValueOf(d); d.Type().Kind() != reflect.Ptr {
+				err = cannotConvert(d, s)
+			} else {
+				err = convertAssignBytes(d.Elem(), s)
+			}
+		}
+	case int64:
+		switch d := d.(type) {
+		case *int:
+			x := int(s)
+			if int64(x) != s {
+				err = strconv.ErrRange
+				x = 0
+			}
+			*d = x
+		case *bool:
+			*d = s != 0
+		case *interface{}:
+			*d = s
+		case nil:
+			// skip value
+		default:
+			if d := reflect.ValueOf(d); d.Type().Kind() != reflect.Ptr {
+				err = cannotConvert(d, s)
+			} else {
+				err = convertAssignInt(d.Elem(), s)
+			}
+		}
+	case []interface{}:
+		switch d := d.(type) {
+		case *[]interface{}:
+			*d = s
+		case *interface{}:
+			*d = s
+		case nil:
+			// skip value
+		default:
+			if d := reflect.ValueOf(d); d.Type().Kind() != reflect.Ptr {
+				err = cannotConvert(d, s)
+			} else {
+				err = convertAssignValues(d.Elem(), s)
+			}
+		}
+	case Error:
+		err = s
+	default:
+		err = cannotConvert(reflect.ValueOf(d), s)
+	}
+	return
+}
+
+// Scan copies from src to the values pointed at by dest.
+//
+// The values pointed at by dest must be an integer, float, boolean, string,
+// []byte, interface{} or slices of these types. Scan uses the standard strconv
+// package to convert bulk strings to numeric and boolean types.
+//
+// If a dest value is nil, then the corresponding src value is skipped.
+//
+// If a src element is nil, then the corresponding dest value is not modified.
+//
+// To enable easy use of Scan in a loop, Scan returns the slice of src
+// following the copied values.
+func Scan(src []interface{}, dest ...interface{}) ([]interface{}, error) {
+	if len(src) < len(dest) {
+		return nil, errors.New("redigo: Scan array short")
+	}
+	var err error
+	for i, d := range dest {
+		err = convertAssign(d, src[i])
+		if err != nil {
+			break
+		}
+	}
+	return src[len(dest):], err
+}
+
+type fieldSpec struct {
+	name  string
+	index []int
+	//omitEmpty bool
+}
+
+type structSpec struct {
+	m map[string]*fieldSpec
+	l []*fieldSpec
+}
+
+func (ss *structSpec) fieldSpec(name []byte) *fieldSpec {
+	return ss.m[string(name)]
+}
+
+func compileStructSpec(t reflect.Type, depth map[string]int, index []int, ss *structSpec) {
+	for i := 0; i < t.NumField(); i++ {
+		f := t.Field(i)
+		switch {
+		case f.PkgPath != "":
+			// Ignore unexported fields.
+		case f.Anonymous:
+			// TODO: Handle pointers. Requires change to decoder and
+			// protection against infinite recursion.
+			if f.Type.Kind() == reflect.Struct {
+				compileStructSpec(f.Type, depth, append(index, i), ss)
+			}
+		default:
+			fs := &fieldSpec{name: f.Name}
+			tag := f.Tag.Get("redis")
+			p := strings.Split(tag, ",")
+			if len(p) > 0 {
+				if p[0] == "-" {
+					continue
+				}
+				if len(p[0]) > 0 {
+					fs.name = p[0]
+				}
+				for _, s := range p[1:] {
+					switch s {
+					//case "omitempty":
+					//  fs.omitempty = true
+					default:
+						panic(errors.New("redigo: unknown field flag " + s + " for type " + t.Name()))
+					}
+				}
+			}
+			d, found := depth[fs.name]
+			if !found {
+				d = 1 << 30
+			}
+			switch {
+			case len(index) == d:
+				// At same depth, remove from result.
+				delete(ss.m, fs.name)
+				j := 0
+				for i := 0; i < len(ss.l); i++ {
+					if fs.name != ss.l[i].name {
+						ss.l[j] = ss.l[i]
+						j += 1
+					}
+				}
+				ss.l = ss.l[:j]
+			case len(index) < d:
+				fs.index = make([]int, len(index)+1)
+				copy(fs.index, index)
+				fs.index[len(index)] = i
+				depth[fs.name] = len(index)
+				ss.m[fs.name] = fs
+				ss.l = append(ss.l, fs)
+			}
+		}
+	}
+}
+
+var (
+	structSpecMutex  sync.RWMutex
+	structSpecCache  = make(map[reflect.Type]*structSpec)
+	defaultFieldSpec = &fieldSpec{}
+)
+
+func structSpecForType(t reflect.Type) *structSpec {
+
+	structSpecMutex.RLock()
+	ss, found := structSpecCache[t]
+	structSpecMutex.RUnlock()
+	if found {
+		return ss
+	}
+
+	structSpecMutex.Lock()
+	defer structSpecMutex.Unlock()
+	ss, found = structSpecCache[t]
+	if found {
+		return ss
+	}
+
+	ss = &structSpec{m: make(map[string]*fieldSpec)}
+	compileStructSpec(t, make(map[string]int), nil, ss)
+	structSpecCache[t] = ss
+	return ss
+}
+
+var errScanStructValue = errors.New("redigo: ScanStruct value must be non-nil pointer to a struct")
+
+// ScanStruct scans alternating names and values from src to a struct. The
+// HGETALL and CONFIG GET commands return replies in this format.
+//
+// ScanStruct uses exported field names to match values in the response. Use
+// 'redis' field tag to override the name:
+//
+//      Field int `redis:"myName"`
+//
+// Fields with the tag redis:"-" are ignored.
+//
+// Integer, float, boolean, string and []byte fields are supported. Scan uses the
+// standard strconv package to convert bulk string values to numeric and
+// boolean types.
+//
+// If a src element is nil, then the corresponding field is not modified.
+func ScanStruct(src []interface{}, dest interface{}) error {
+	d := reflect.ValueOf(dest)
+	if d.Kind() != reflect.Ptr || d.IsNil() {
+		return errScanStructValue
+	}
+	d = d.Elem()
+	if d.Kind() != reflect.Struct {
+		return errScanStructValue
+	}
+	ss := structSpecForType(d.Type())
+
+	if len(src)%2 != 0 {
+		return errors.New("redigo: ScanStruct expects even number of values in values")
+	}
+
+	for i := 0; i < len(src); i += 2 {
+		s := src[i+1]
+		if s == nil {
+			continue
+		}
+		name, ok := src[i].([]byte)
+		if !ok {
+			return errors.New("redigo: ScanStruct key not a bulk string value")
+		}
+		fs := ss.fieldSpec(name)
+		if fs == nil {
+			continue
+		}
+		if err := convertAssignValue(d.FieldByIndex(fs.index), s); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+var (
+	errScanSliceValue = errors.New("redigo: ScanSlice dest must be non-nil pointer to a struct")
+)
+
+// ScanSlice scans src to the slice pointed to by dest. The elements the dest
+// slice must be integer, float, boolean, string, struct or pointer to struct
+// values.
+//
+// Struct fields must be integer, float, boolean or string values. All struct
+// fields are used unless a subset is specified using fieldNames.
+func ScanSlice(src []interface{}, dest interface{}, fieldNames ...string) error {
+	d := reflect.ValueOf(dest)
+	if d.Kind() != reflect.Ptr || d.IsNil() {
+		return errScanSliceValue
+	}
+	d = d.Elem()
+	if d.Kind() != reflect.Slice {
+		return errScanSliceValue
+	}
+
+	isPtr := false
+	t := d.Type().Elem()
+	if t.Kind() == reflect.Ptr && t.Elem().Kind() == reflect.Struct {
+		isPtr = true
+		t = t.Elem()
+	}
+
+	if t.Kind() != reflect.Struct {
+		ensureLen(d, len(src))
+		for i, s := range src {
+			if s == nil {
+				continue
+			}
+			if err := convertAssignValue(d.Index(i), s); err != nil {
+				return err
+			}
+		}
+		return nil
+	}
+
+	ss := structSpecForType(t)
+	fss := ss.l
+	if len(fieldNames) > 0 {
+		fss = make([]*fieldSpec, len(fieldNames))
+		for i, name := range fieldNames {
+			fss[i] = ss.m[name]
+			if fss[i] == nil {
+				return errors.New("redigo: ScanSlice bad field name " + name)
+			}
+		}
+	}
+
+	if len(fss) == 0 {
+		return errors.New("redigo: ScanSlice no struct fields")
+	}
+
+	n := len(src) / len(fss)
+	if n*len(fss) != len(src) {
+		return errors.New("redigo: ScanSlice length not a multiple of struct field count")
+	}
+
+	ensureLen(d, n)
+	for i := 0; i < n; i++ {
+		d := d.Index(i)
+		if isPtr {
+			if d.IsNil() {
+				d.Set(reflect.New(t))
+			}
+			d = d.Elem()
+		}
+		for j, fs := range fss {
+			s := src[i*len(fss)+j]
+			if s == nil {
+				continue
+			}
+			if err := convertAssignValue(d.FieldByIndex(fs.index), s); err != nil {
+				return err
+			}
+		}
+	}
+	return nil
+}
+
+// Args is a helper for constructing command arguments from structured values.
+type Args []interface{}
+
+// Add returns the result of appending value to args.
+func (args Args) Add(value ...interface{}) Args {
+	return append(args, value...)
+}
+
+// AddFlat returns the result of appending the flattened value of v to args.
+//
+// Maps are flattened by appending the alternating keys and map values to args.
+//
+// Slices are flattened by appending the slice elements to args.
+//
+// Structs are flattened by appending the alternating names and values of
+// exported fields to args. If v is a nil struct pointer, then nothing is
+// appended. The 'redis' field tag overrides struct field names. See ScanStruct
+// for more information on the use of the 'redis' field tag.
+//
+// Other types are appended to args as is.
+func (args Args) AddFlat(v interface{}) Args {
+	rv := reflect.ValueOf(v)
+	switch rv.Kind() {
+	case reflect.Struct:
+		args = flattenStruct(args, rv)
+	case reflect.Slice:
+		for i := 0; i < rv.Len(); i++ {
+			args = append(args, rv.Index(i).Interface())
+		}
+	case reflect.Map:
+		for _, k := range rv.MapKeys() {
+			args = append(args, k.Interface(), rv.MapIndex(k).Interface())
+		}
+	case reflect.Ptr:
+		if rv.Type().Elem().Kind() == reflect.Struct {
+			if !rv.IsNil() {
+				args = flattenStruct(args, rv.Elem())
+			}
+		} else {
+			args = append(args, v)
+		}
+	default:
+		args = append(args, v)
+	}
+	return args
+}
+
+func flattenStruct(args Args, v reflect.Value) Args {
+	ss := structSpecForType(v.Type())
+	for _, fs := range ss.l {
+		fv := v.FieldByIndex(fs.index)
+		args = append(args, fs.name, fv.Interface())
+	}
+	return args
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/scan_test.go

@@ -0,0 +1,412 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis_test
+
+import (
+	"fmt"
+	"github.com/garyburd/redigo/redis"
+	"math"
+	"reflect"
+	"testing"
+)
+
+var scanConversionTests = []struct {
+	src  interface{}
+	dest interface{}
+}{
+	{[]byte("-inf"), math.Inf(-1)},
+	{[]byte("+inf"), math.Inf(1)},
+	{[]byte("0"), float64(0)},
+	{[]byte("3.14159"), float64(3.14159)},
+	{[]byte("3.14"), float32(3.14)},
+	{[]byte("-100"), int(-100)},
+	{[]byte("101"), int(101)},
+	{int64(102), int(102)},
+	{[]byte("103"), uint(103)},
+	{int64(104), uint(104)},
+	{[]byte("105"), int8(105)},
+	{int64(106), int8(106)},
+	{[]byte("107"), uint8(107)},
+	{int64(108), uint8(108)},
+	{[]byte("0"), false},
+	{int64(0), false},
+	{[]byte("f"), false},
+	{[]byte("1"), true},
+	{int64(1), true},
+	{[]byte("t"), true},
+	{[]byte("hello"), "hello"},
+	{[]byte("world"), []byte("world")},
+	{[]interface{}{[]byte("foo")}, []interface{}{[]byte("foo")}},
+	{[]interface{}{[]byte("foo")}, []string{"foo"}},
+	{[]interface{}{[]byte("hello"), []byte("world")}, []string{"hello", "world"}},
+	{[]interface{}{[]byte("bar")}, [][]byte{[]byte("bar")}},
+	{[]interface{}{[]byte("1")}, []int{1}},
+	{[]interface{}{[]byte("1"), []byte("2")}, []int{1, 2}},
+	{[]interface{}{[]byte("1"), []byte("2")}, []float64{1, 2}},
+	{[]interface{}{[]byte("1")}, []byte{1}},
+	{[]interface{}{[]byte("1")}, []bool{true}},
+}
+
+func TestScanConversion(t *testing.T) {
+	for _, tt := range scanConversionTests {
+		values := []interface{}{tt.src}
+		dest := reflect.New(reflect.TypeOf(tt.dest))
+		values, err := redis.Scan(values, dest.Interface())
+		if err != nil {
+			t.Errorf("Scan(%v) returned error %v", tt, err)
+			continue
+		}
+		if !reflect.DeepEqual(tt.dest, dest.Elem().Interface()) {
+			t.Errorf("Scan(%v) returned %v, want %v", tt, dest.Elem().Interface(), tt.dest)
+		}
+	}
+}
+
+var scanConversionErrorTests = []struct {
+	src  interface{}
+	dest interface{}
+}{
+	{[]byte("1234"), byte(0)},
+	{int64(1234), byte(0)},
+	{[]byte("-1"), byte(0)},
+	{int64(-1), byte(0)},
+	{[]byte("junk"), false},
+	{redis.Error("blah"), false},
+}
+
+func TestScanConversionError(t *testing.T) {
+	for _, tt := range scanConversionErrorTests {
+		values := []interface{}{tt.src}
+		dest := reflect.New(reflect.TypeOf(tt.dest))
+		values, err := redis.Scan(values, dest.Interface())
+		if err == nil {
+			t.Errorf("Scan(%v) did not return error", tt)
+		}
+	}
+}
+
+func ExampleScan() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+
+	c.Send("HMSET", "album:1", "title", "Red", "rating", 5)
+	c.Send("HMSET", "album:2", "title", "Earthbound", "rating", 1)
+	c.Send("HMSET", "album:3", "title", "Beat")
+	c.Send("LPUSH", "albums", "1")
+	c.Send("LPUSH", "albums", "2")
+	c.Send("LPUSH", "albums", "3")
+	values, err := redis.Values(c.Do("SORT", "albums",
+		"BY", "album:*->rating",
+		"GET", "album:*->title",
+		"GET", "album:*->rating"))
+	if err != nil {
+		panic(err)
+	}
+
+	for len(values) > 0 {
+		var title string
+		rating := -1 // initialize to illegal value to detect nil.
+		values, err = redis.Scan(values, &title, &rating)
+		if err != nil {
+			panic(err)
+		}
+		if rating == -1 {
+			fmt.Println(title, "not-rated")
+		} else {
+			fmt.Println(title, rating)
+		}
+	}
+	// Output:
+	// Beat not-rated
+	// Earthbound 1
+	// Red 5
+}
+
+type s0 struct {
+	X  int
+	Y  int `redis:"y"`
+	Bt bool
+}
+
+type s1 struct {
+	X  int    `redis:"-"`
+	I  int    `redis:"i"`
+	U  uint   `redis:"u"`
+	S  string `redis:"s"`
+	P  []byte `redis:"p"`
+	B  bool   `redis:"b"`
+	Bt bool
+	Bf bool
+	s0
+}
+
+var scanStructTests = []struct {
+	title string
+	reply []string
+	value interface{}
+}{
+	{"basic",
+		[]string{"i", "-1234", "u", "5678", "s", "hello", "p", "world", "b", "t", "Bt", "1", "Bf", "0", "X", "123", "y", "456"},
+		&s1{I: -1234, U: 5678, S: "hello", P: []byte("world"), B: true, Bt: true, Bf: false, s0: s0{X: 123, Y: 456}},
+	},
+}
+
+func TestScanStruct(t *testing.T) {
+	for _, tt := range scanStructTests {
+
+		var reply []interface{}
+		for _, v := range tt.reply {
+			reply = append(reply, []byte(v))
+		}
+
+		value := reflect.New(reflect.ValueOf(tt.value).Type().Elem())
+
+		if err := redis.ScanStruct(reply, value.Interface()); err != nil {
+			t.Fatalf("ScanStruct(%s) returned error %v", tt.title, err)
+		}
+
+		if !reflect.DeepEqual(value.Interface(), tt.value) {
+			t.Fatalf("ScanStruct(%s) returned %v, want %v", tt.title, value.Interface(), tt.value)
+		}
+	}
+}
+
+func TestBadScanStructArgs(t *testing.T) {
+	x := []interface{}{"A", "b"}
+	test := func(v interface{}) {
+		if err := redis.ScanStruct(x, v); err == nil {
+			t.Errorf("Expect error for ScanStruct(%T, %T)", x, v)
+		}
+	}
+
+	test(nil)
+
+	var v0 *struct{}
+	test(v0)
+
+	var v1 int
+	test(&v1)
+
+	x = x[:1]
+	v2 := struct{ A string }{}
+	test(&v2)
+}
+
+var scanSliceTests = []struct {
+	src        []interface{}
+	fieldNames []string
+	ok         bool
+	dest       interface{}
+}{
+	{
+		[]interface{}{[]byte("1"), nil, []byte("-1")},
+		nil,
+		true,
+		[]int{1, 0, -1},
+	},
+	{
+		[]interface{}{[]byte("1"), nil, []byte("2")},
+		nil,
+		true,
+		[]uint{1, 0, 2},
+	},
+	{
+		[]interface{}{[]byte("-1")},
+		nil,
+		false,
+		[]uint{1},
+	},
+	{
+		[]interface{}{[]byte("hello"), nil, []byte("world")},
+		nil,
+		true,
+		[][]byte{[]byte("hello"), nil, []byte("world")},
+	},
+	{
+		[]interface{}{[]byte("hello"), nil, []byte("world")},
+		nil,
+		true,
+		[]string{"hello", "", "world"},
+	},
+	{
+		[]interface{}{[]byte("a1"), []byte("b1"), []byte("a2"), []byte("b2")},
+		nil,
+		true,
+		[]struct{ A, B string }{{"a1", "b1"}, {"a2", "b2"}},
+	},
+	{
+		[]interface{}{[]byte("a1"), []byte("b1")},
+		nil,
+		false,
+		[]struct{ A, B, C string }{{"a1", "b1", ""}},
+	},
+	{
+		[]interface{}{[]byte("a1"), []byte("b1"), []byte("a2"), []byte("b2")},
+		nil,
+		true,
+		[]*struct{ A, B string }{{"a1", "b1"}, {"a2", "b2"}},
+	},
+	{
+		[]interface{}{[]byte("a1"), []byte("b1"), []byte("a2"), []byte("b2")},
+		[]string{"A", "B"},
+		true,
+		[]struct{ A, C, B string }{{"a1", "", "b1"}, {"a2", "", "b2"}},
+	},
+	{
+		[]interface{}{[]byte("a1"), []byte("b1"), []byte("a2"), []byte("b2")},
+		nil,
+		false,
+		[]struct{}{},
+	},
+}
+
+func TestScanSlice(t *testing.T) {
+	for _, tt := range scanSliceTests {
+
+		typ := reflect.ValueOf(tt.dest).Type()
+		dest := reflect.New(typ)
+
+		err := redis.ScanSlice(tt.src, dest.Interface(), tt.fieldNames...)
+		if tt.ok != (err == nil) {
+			t.Errorf("ScanSlice(%v, []%s, %v) returned error %v", tt.src, typ, tt.fieldNames, err)
+			continue
+		}
+		if tt.ok && !reflect.DeepEqual(dest.Elem().Interface(), tt.dest) {
+			t.Errorf("ScanSlice(src, []%s) returned %#v, want %#v", typ, dest.Elem().Interface(), tt.dest)
+		}
+	}
+}
+
+func ExampleScanSlice() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+
+	c.Send("HMSET", "album:1", "title", "Red", "rating", 5)
+	c.Send("HMSET", "album:2", "title", "Earthbound", "rating", 1)
+	c.Send("HMSET", "album:3", "title", "Beat", "rating", 4)
+	c.Send("LPUSH", "albums", "1")
+	c.Send("LPUSH", "albums", "2")
+	c.Send("LPUSH", "albums", "3")
+	values, err := redis.Values(c.Do("SORT", "albums",
+		"BY", "album:*->rating",
+		"GET", "album:*->title",
+		"GET", "album:*->rating"))
+	if err != nil {
+		panic(err)
+	}
+
+	var albums []struct {
+		Title  string
+		Rating int
+	}
+	if err := redis.ScanSlice(values, &albums); err != nil {
+		panic(err)
+	}
+	fmt.Printf("%v\n", albums)
+	// Output:
+	// [{Earthbound 1} {Beat 4} {Red 5}]
+}
+
+var argsTests = []struct {
+	title    string
+	actual   redis.Args
+	expected redis.Args
+}{
+	{"struct ptr",
+		redis.Args{}.AddFlat(&struct {
+			I  int    `redis:"i"`
+			U  uint   `redis:"u"`
+			S  string `redis:"s"`
+			P  []byte `redis:"p"`
+			Bt bool
+			Bf bool
+		}{
+			-1234, 5678, "hello", []byte("world"), true, false,
+		}),
+		redis.Args{"i", int(-1234), "u", uint(5678), "s", "hello", "p", []byte("world"), "Bt", true, "Bf", false},
+	},
+	{"struct",
+		redis.Args{}.AddFlat(struct{ I int }{123}),
+		redis.Args{"I", 123},
+	},
+	{"slice",
+		redis.Args{}.Add(1).AddFlat([]string{"a", "b", "c"}).Add(2),
+		redis.Args{1, "a", "b", "c", 2},
+	},
+}
+
+func TestArgs(t *testing.T) {
+	for _, tt := range argsTests {
+		if !reflect.DeepEqual(tt.actual, tt.expected) {
+			t.Fatalf("%s is %v, want %v", tt.title, tt.actual, tt.expected)
+		}
+	}
+}
+
+func ExampleArgs() {
+	c, err := dial()
+	if err != nil {
+		panic(err)
+	}
+	defer c.Close()
+
+	var p1, p2 struct {
+		Title  string `redis:"title"`
+		Author string `redis:"author"`
+		Body   string `redis:"body"`
+	}
+
+	p1.Title = "Example"
+	p1.Author = "Gary"
+	p1.Body = "Hello"
+
+	if _, err := c.Do("HMSET", redis.Args{}.Add("id1").AddFlat(&p1)...); err != nil {
+		panic(err)
+	}
+
+	m := map[string]string{
+		"title":  "Example2",
+		"author": "Steve",
+		"body":   "Map",
+	}
+
+	if _, err := c.Do("HMSET", redis.Args{}.Add("id2").AddFlat(m)...); err != nil {
+		panic(err)
+	}
+
+	for _, id := range []string{"id1", "id2"} {
+
+		v, err := redis.Values(c.Do("HGETALL", id))
+		if err != nil {
+			panic(err)
+		}
+
+		if err := redis.ScanStruct(v, &p2); err != nil {
+			panic(err)
+		}
+
+		fmt.Printf("%+v\n", p2)
+	}
+
+	// Output:
+	// {Title:Example Author:Gary Body:Hello}
+	// {Title:Example2 Author:Steve Body:Map}
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/script.go

@@ -0,0 +1,86 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"crypto/sha1"
+	"encoding/hex"
+	"io"
+	"strings"
+)
+
+// Script encapsulates the source, hash and key count for a Lua script. See
+// http://redis.io/commands/eval for information on scripts in Redis.
+type Script struct {
+	keyCount int
+	src      string
+	hash     string
+}
+
+// NewScript returns a new script object. If keyCount is greater than or equal
+// to zero, then the count is automatically inserted in the EVAL command
+// argument list. If keyCount is less than zero, then the application supplies
+// the count as the first value in the keysAndArgs argument to the Do, Send and
+// SendHash methods.
+func NewScript(keyCount int, src string) *Script {
+	h := sha1.New()
+	io.WriteString(h, src)
+	return &Script{keyCount, src, hex.EncodeToString(h.Sum(nil))}
+}
+
+func (s *Script) args(spec string, keysAndArgs []interface{}) []interface{} {
+	var args []interface{}
+	if s.keyCount < 0 {
+		args = make([]interface{}, 1+len(keysAndArgs))
+		args[0] = spec
+		copy(args[1:], keysAndArgs)
+	} else {
+		args = make([]interface{}, 2+len(keysAndArgs))
+		args[0] = spec
+		args[1] = s.keyCount
+		copy(args[2:], keysAndArgs)
+	}
+	return args
+}
+
+// Do evaluates the script. Under the covers, Do optimistically evaluates the
+// script using the EVALSHA command. If the command fails because the script is
+// not loaded, then Do evaluates the script using the EVAL command (thus
+// causing the script to load).
+func (s *Script) Do(c Conn, keysAndArgs ...interface{}) (interface{}, error) {
+	v, err := c.Do("EVALSHA", s.args(s.hash, keysAndArgs)...)
+	if e, ok := err.(Error); ok && strings.HasPrefix(string(e), "NOSCRIPT ") {
+		v, err = c.Do("EVAL", s.args(s.src, keysAndArgs)...)
+	}
+	return v, err
+}
+
+// SendHash evaluates the script without waiting for the reply. The script is
+// evaluated with the EVALSHA command. The application must ensure that the
+// script is loaded by a previous call to Send, Do or Load methods.
+func (s *Script) SendHash(c Conn, keysAndArgs ...interface{}) error {
+	return c.Send("EVALSHA", s.args(s.hash, keysAndArgs)...)
+}
+
+// Send evaluates the script without waiting for the reply.
+func (s *Script) Send(c Conn, keysAndArgs ...interface{}) error {
+	return c.Send("EVAL", s.args(s.src, keysAndArgs)...)
+}
+
+// Load loads the script without evaluating it.
+func (s *Script) Load(c Conn) error {
+	_, err := c.Do("SCRIPT", "LOAD", s.src)
+	return err
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/script_test.go

@@ -0,0 +1,93 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis_test
+
+import (
+	"fmt"
+	"reflect"
+	"testing"
+	"time"
+
+	"github.com/garyburd/redigo/internal/redistest"
+	"github.com/garyburd/redigo/redis"
+)
+
+func ExampleScript(c redis.Conn, reply interface{}, err error) {
+	// Initialize a package-level variable with a script.
+	var getScript = redis.NewScript(1, `return redis.call('get', KEYS[1])`)
+
+	// In a function, use the script Do method to evaluate the script. The Do
+	// method optimistically uses the EVALSHA command. If the script is not
+	// loaded, then the Do method falls back to the EVAL command.
+	reply, err = getScript.Do(c, "foo")
+}
+
+func TestScript(t *testing.T) {
+	c, err := redistest.Dial()
+	if err != nil {
+		t.Fatalf("error connection to database, %v", err)
+	}
+	defer c.Close()
+
+	// To test fall back in Do, we make script unique by adding comment with current time.
+	script := fmt.Sprintf("--%d\nreturn {KEYS[1],KEYS[2],ARGV[1],ARGV[2]}", time.Now().UnixNano())
+	s := redis.NewScript(2, script)
+	reply := []interface{}{[]byte("key1"), []byte("key2"), []byte("arg1"), []byte("arg2")}
+
+	v, err := s.Do(c, "key1", "key2", "arg1", "arg2")
+	if err != nil {
+		t.Errorf("s.Do(c, ...) returned %v", err)
+	}
+
+	if !reflect.DeepEqual(v, reply) {
+		t.Errorf("s.Do(c, ..); = %v, want %v", v, reply)
+	}
+
+	err = s.Load(c)
+	if err != nil {
+		t.Errorf("s.Load(c) returned %v", err)
+	}
+
+	err = s.SendHash(c, "key1", "key2", "arg1", "arg2")
+	if err != nil {
+		t.Errorf("s.SendHash(c, ...) returned %v", err)
+	}
+
+	err = c.Flush()
+	if err != nil {
+		t.Errorf("c.Flush() returned %v", err)
+	}
+
+	v, err = c.Receive()
+	if !reflect.DeepEqual(v, reply) {
+		t.Errorf("s.SendHash(c, ..); c.Receive() = %v, want %v", v, reply)
+	}
+
+	err = s.Send(c, "key1", "key2", "arg1", "arg2")
+	if err != nil {
+		t.Errorf("s.Send(c, ...) returned %v", err)
+	}
+
+	err = c.Flush()
+	if err != nil {
+		t.Errorf("c.Flush() returned %v", err)
+	}
+
+	v, err = c.Receive()
+	if !reflect.DeepEqual(v, reply) {
+		t.Errorf("s.Send(c, ..); c.Receive() = %v, want %v", v, reply)
+	}
+
+}

Godeps/_workspace/src/github.com/garyburd/redigo/redis/test_test.go

@@ -0,0 +1,38 @@
+// Copyright 2012 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis
+
+import (
+	"bufio"
+	"net"
+	"time"
+)
+
+func SetNowFunc(f func() time.Time) {
+	nowFunc = f
+}
+
+type nopCloser struct{ net.Conn }
+
+func (nopCloser) Close() error { return nil }
+
+// NewConnBufio is a hook for tests.
+func NewConnBufio(rw bufio.ReadWriter) Conn {
+	return &conn{br: rw.Reader, bw: rw.Writer, conn: nopCloser{}}
+}
+
+var (
+	ErrNegativeInt = errNegativeInt
+)

Godeps/_workspace/src/github.com/garyburd/redigo/redis/zpop_example_test.go

@@ -0,0 +1,113 @@
+// Copyright 2013 Gary Burd
+//
+// Licensed under the Apache License, Version 2.0 (the "License"): you may
+// not use this file except in compliance with the License. You may obtain
+// a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+// WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+// License for the specific language governing permissions and limitations
+// under the License.
+
+package redis_test
+
+import (
+	"fmt"
+	"github.com/garyburd/redigo/redis"
+)
+
+// zpop pops a value from the ZSET key using WATCH/MULTI/EXEC commands.
+func zpop(c redis.Conn, key string) (result string, err error) {
+
+	defer func() {
+		// Return connection to normal state on error.
+		if err != nil {
+			c.Do("DISCARD")
+		}
+	}()
+
+	// Loop until transaction is successful.
+	for {
+		if _, err := c.Do("WATCH", key); err != nil {
+			return "", err
+		}
+
+		members, err := redis.Strings(c.Do("ZRANGE", key, 0, 0))
+		if err != nil {
+			return "", err
+		}
+		if len(members) != 1 {
+			return "", redis.ErrNil
+		}
+
+		c.Send("MULTI")
+		c.Send("ZREM", key, members[0])
+		queued, err := c.Do("EXEC")
+		if err != nil {
+			return "", err
+		}
+
+		if queued != nil {
+			result = members[0]
+			break
+		}
+	}
+
+	return result, nil
+}
+
+// zpopScript pops a value from a ZSET.
+var zpopScript = redis.NewScript(1, `
+    local r = redis.call('ZRANGE', KEYS[1], 0, 0)
+    if r ~= nil then
+        r = r[1]
+        redis.call('ZREM', KEYS[1], r)
+    end
+    return r
+`)
+
+// This example implements ZPOP as described at
+// http://redis.io/topics/transactions using WATCH/MULTI/EXEC and scripting.
+func Example_zpop() {
+	c, err := dial()
+	if err != nil {
+		fmt.Println(err)
+		return
+	}
+	defer c.Close()
+
+	// Add test data using a pipeline.
+
+	for i, member := range []string{"red", "blue", "green"} {
+		c.Send("ZADD", "zset", i, member)
+	}
+	if _, err := c.Do(""); err != nil {
+		fmt.Println(err)
+		return
+	}
+
+	// Pop using WATCH/MULTI/EXEC
+
+	v, err := zpop(c, "zset")
+	if err != nil {
+		fmt.Println(err)
+		return
+	}
+	fmt.Println(v)
+
+	// Pop using a script.
+
+	v, err = redis.String(zpopScript.Do(c, "zset"))
+	if err != nil {
+		fmt.Println(err)
+		return
+	}
+	fmt.Println(v)
+
+	// Output:
+	// red
+	// blue
+}

cmd/registry/config.yml

@@ -5,6 +5,8 @@ log:
     service: registry
     environment: development
 storage:
+    cache:
+        layerinfo: inmemory
     filesystem:
         rootdirectory: /tmp/registry-dev
 http:
@@ -12,6 +14,15 @@ http:
     secret: asecretforlocaldevelopment
     debug:
         addr: localhost:5001
+redis:
+  addr: localhost:6379
+  pool:
+    maxidle: 16
+    maxactive: 64
+    idletimeout: 300s
+  dialtimeout: 10ms
+  readtimeout: 10ms
+  writetimeout: 10ms
 notifications:
     endpoints:
         - name: local-8082

configuration/configuration.go

@@ -91,6 +91,36 @@ type Configuration struct {
 	// Notifications specifies configuration about various endpoint to which
 	// registry events are dispatched.
 	Notifications Notifications `yaml:"notifications,omitempty"`
+
+	// Redis configures the redis pool available to the registry webapp.
+	Redis struct {
+		// Addr specifies the the redis instance available to the application.
+		Addr string `yaml:"addr,omitempty"`
+
+		// Password string to use when making a connection.
+		Password string `yaml:"password,omitempty"`
+
+		// DB specifies the database to connect to on the redis instance.
+		DB int `yaml:"db,omitempty"`
+
+		DialTimeout  time.Duration `yaml:"dialtimeout,omitempty"`  // timeout for connect
+		ReadTimeout  time.Duration `yaml:"readtimeout,omitempty"`  // timeout for reads of data
+		WriteTimeout time.Duration `yaml:"writetimeout,omitempty"` // timeout for writes of data
+
+		// Pool configures the behavior of the redis connection pool.
+		Pool struct {
+			// MaxIdle sets the maximum number of idle connections.
+			MaxIdle int `yaml:"maxidle,omitempty"`
+
+			// MaxActive sets the maximum number of connections that should be
+			// opened before blocking a connection request.
+			MaxActive int `yaml:"maxactive,omitempty"`
+
+			// IdleTimeout sets the amount time to wait before closing
+			// inactive connections.
+			IdleTimeout time.Duration `yaml:"idletimeout,omitempty"`
+		} `yaml:"pool,omitempty"`
+	} `yaml:"redis,omitempty"`
 }
 
 // v0_1Configuration is a Version 0.1 Configuration struct
@@ -157,8 +187,13 @@ type Storage map[string]Parameters
 func (storage Storage) Type() string {
 	// Return only key in this map
 	for k := range storage {
+		switch k {
+		case "cache":
+			// allow configuration of caching
+		default:
 			return k
 		}
+	}
 	return ""
 }
 
@@ -181,10 +216,18 @@ func (storage *Storage) UnmarshalYAML(unmarshal func(interface{}) error) error {
 		if len(storageMap) > 1 {
 			types := make([]string, 0, len(storageMap))
 			for k := range storageMap {
+				switch k {
+				case "cache":
+					// allow configuration of caching
+				default:
 					types = append(types, k)
 				}
+			}
+
+			if len(types) > 1 {
 				return fmt.Errorf("Must provide exactly one storage type. Provided: %v", types)
 			}
+		}
 		*storage = storageMap
 		return nil
 	}

context/logger.go

@@ -44,6 +44,20 @@ func WithLogger(ctx Context, logger Logger) Context {
 	return WithValue(ctx, "logger", logger)
 }
 
+// GetLoggerWithField returns a logger instance with the specified field key
+// and value without affecting the context. Extra specified keys will be
+// resolved from the context.
+func GetLoggerWithField(ctx context.Context, key, value interface{}, keys ...interface{}) Logger {
+	return getLogrusLogger(ctx, keys...).WithField(fmt.Sprint(key), value)
+}
+
+// GetLoggerWithFields returns a logger instance with the specified fields
+// without affecting the context. Extra specified keys will be resolved from
+// the context.
+func GetLoggerWithFields(ctx context.Context, fields map[string]interface{}, keys ...interface{}) Logger {
+	return getLogrusLogger(ctx, keys...).WithFields(logrus.Fields(fields))
+}
+
 // GetLogger returns the logger from the current context, if present. If one
 // or more keys are provided, they will be resolved on the context and
 // included in the logger. While context.Value takes an interface, any key

doc/configuration.md

@@ -28,6 +28,8 @@ storage:
 		v4auth: true
 		chunksize: 5242880
 		rootdirectory: /s3/object/name/prefix
+	cache:
+		layerinfo: inmemory
 auth:
 	silly:
 		realm: silly-realm
@@ -82,6 +84,17 @@ notifications:
 		  timeout: 500
 		  threshold: 5
 		  backoff: 1000
+redis:
+	addr: localhost:6379
+	password: asecret
+	db: 0
+	dialtimeout: 10ms
+	readtimeout: 10ms
+	writetimeout: 10ms
+	pool:
+		maxidle: 16
+		maxactive: 64
+		idletimeout: 300s
 ```
 
 N.B. In some instances a configuration option may be marked **optional** but contain child options marked as **required**. This indicates that a parent may be omitted with all its children, however, if the parent is included, the children marked **required** must be included.
@@ -151,10 +164,20 @@ storage:
 		v4auth: true
 		chunksize: 5242880
 		rootdirectory: /s3/object/name/prefix
+	cache:
+		layerinfo: inmemory
 ```
 
 The storage option is **required** and defines which storage backend is in use. At the moment only one backend may be configured, an error is returned when the registry is started with more than one storage backend configured.
 
+A `cache` subsection can be used to enable caching of data accessed in the
+storage backend. Currently, the only available cache provides fast access to
+layer metadata. This if configured using the `layerinfo` field. The following
+cache implementations are available:
+
+- redis: using the redis pool to cache layer meta data.
+- inmemory: use an in memory map to cache layer meta data.
+
 The following backends may be configured, **all options for a given storage backend are required**:
 
 ### filesystem
@@ -347,3 +370,48 @@ Endpoints is a list of named services (URLs) that can accept event notifications
 - threshold: **Required** - TODO: fill in description
 - backoff: **Required** - TODO: fill in description
 
+## redis
+
+```yaml
+redis:
+	addr: localhost:6379
+	password: asecret
+	db: 0
+	dialtimeout: 10ms
+	readtimeout: 10ms
+	writetimeout: 10ms
+	pool:
+		maxidle: 16
+		maxactive: 64
+		idletimeout: 300s
+```
+
+Declare parameters for constructing the redis connections. Registry instances
+may use the redis instance for several applications. The current purpose is
+caching information about immutable blobs. Most of the options below control
+how the registry connects to redis. The behavior of the pool can be controlled
+with the [pool](#pool) subsection.
+
+- addr: **Required** - Address (host and port) of redis instance.
+- password: **Optional** - A password used to authenticate to the redis instance.
+- db: **Optional** - Selects the db for each connection.
+- dialtimeout: **Optional** - Timeout for connecting to a redis instance.
+- readtimeout: **Optional** - Timeout for reading from redis connections.
+- writetimeout: **Optional** - Timeout for writing to redis connections.
+
+### pool
+
+```yaml
+pool:
+	maxidle: 16
+	maxactive: 64
+	idletimeout: 300s
+```
+
+Configure the behavior of the redis connection pool.
+
+- maxidle: **Optional** - sets the maximum number of idle connections.
+- maxactive: **Optional** - sets the maximum number of connections that should
+  be opened before blocking a connection request.
+- idletimeout: **Optional** - sets the amount time to wait before closing
+  inactive connections.

notifications/listener_test.go

@@ -9,6 +9,7 @@ import (
 	"github.com/docker/distribution/digest"
 	"github.com/docker/distribution/manifest"
 	"github.com/docker/distribution/registry/storage"
+	"github.com/docker/distribution/registry/storage/cache"
 	"github.com/docker/distribution/registry/storage/driver/inmemory"
 	"github.com/docker/distribution/testutil"
 	"github.com/docker/libtrust"
@@ -16,7 +17,7 @@ import (
 )
 
 func TestListener(t *testing.T) {
-	registry := storage.NewRegistryWithDriver(inmemory.New())
+	registry := storage.NewRegistryWithDriver(inmemory.New(), cache.NewInMemoryLayerInfoCache())
 	tl := &testListener{
 		ops: make(map[string]int),
 	}

registry.go

@@ -33,6 +33,10 @@ type Repository interface {
 	Signatures() SignatureService
 }
 
+// TODO(stevvooe): Must add close methods to all these. May want to change the
+// way instances are created to better reflect internal dependency
+// relationships.
+
 // ManifestService provides operations on image manifests.
 type ManifestService interface {
 	// Exists returns true if the manifest exists.

registry/doc.go

@@ -1,3 +1,3 @@
 // Package registry is a placeholder package for registry interface
-// destinations and utilities.
+// definitions and utilities.
 package registry

registry/handlers/app.go

@@ -1,10 +1,12 @@
 package handlers
 
 import (
+	"expvar"
 	"fmt"
 	"net"
 	"net/http"
 	"os"
+	"time"
 
 	"code.google.com/p/go-uuid/uuid"
 	"github.com/docker/distribution"
@@ -16,9 +18,11 @@ import (
 	registrymiddleware "github.com/docker/distribution/registry/middleware/registry"
 	repositorymiddleware "github.com/docker/distribution/registry/middleware/repository"
 	"github.com/docker/distribution/registry/storage"
+	"github.com/docker/distribution/registry/storage/cache"
 	storagedriver "github.com/docker/distribution/registry/storage/driver"
 	"github.com/docker/distribution/registry/storage/driver/factory"
 	storagemiddleware "github.com/docker/distribution/registry/storage/driver/middleware"
+	"github.com/garyburd/redigo/redis"
 	"github.com/gorilla/mux"
 	"golang.org/x/net/context"
 )
@@ -44,6 +48,8 @@ type App struct {
 		sink   notifications.Sink
 		source notifications.SourceRecord
 	}
+
+	redis *redis.Pool
 }
 
 // Value intercepts calls context.Context.Value, returning the current app id,
@@ -95,8 +101,32 @@ func NewApp(ctx context.Context, configuration configuration.Configuration) *App
 	}
 
 	app.configureEvents(&configuration)
+	app.configureRedis(&configuration)
+
+	// configure storage caches
+	if cc, ok := configuration.Storage["cache"]; ok {
+		switch cc["layerinfo"] {
+		case "redis":
+			if app.redis == nil {
+				panic("redis configuration required to use for layerinfo cache")
+			}
+			app.registry = storage.NewRegistryWithDriver(app.driver, cache.NewRedisLayerInfoCache(app.redis))
+			ctxu.GetLogger(app).Infof("using redis layerinfo cache")
+		case "inmemory":
+			app.registry = storage.NewRegistryWithDriver(app.driver, cache.NewInMemoryLayerInfoCache())
+			ctxu.GetLogger(app).Infof("using inmemory layerinfo cache")
+		default:
+			if cc["layerinfo"] != "" {
+				ctxu.GetLogger(app).Warnf("unkown cache type %q, caching disabled", configuration.Storage["cache"])
+			}
+		}
+	}
+
+	if app.registry == nil {
+		// configure the registry if no cache section is available.
+		app.registry = storage.NewRegistryWithDriver(app.driver, nil)
+	}
 
-	app.registry = storage.NewRegistryWithDriver(app.driver)
 	app.registry, err = applyRegistryMiddleware(app.registry, configuration.Middleware["registry"])
 	if err != nil {
 		panic(err)
@@ -174,6 +204,88 @@ func (app *App) configureEvents(configuration *configuration.Configuration) {
 	}
 }
 
+func (app *App) configureRedis(configuration *configuration.Configuration) {
+	if configuration.Redis.Addr == "" {
+		ctxu.GetLogger(app).Infof("redis not configured")
+		return
+	}
+
+	pool := &redis.Pool{
+		Dial: func() (redis.Conn, error) {
+			// TODO(stevvooe): Yet another use case for contextual timing.
+			ctx := context.WithValue(app, "redis.connect.startedat", time.Now())
+
+			done := func(err error) {
+				logger := ctxu.GetLoggerWithField(ctx, "redis.connect.duration",
+					ctxu.Since(ctx, "redis.connect.startedat"))
+				if err != nil {
+					logger.Errorf("redis: error connecting: %v", err)
+				} else {
+					logger.Infof("redis: connect %v", configuration.Redis.Addr)
+				}
+			}
+
+			conn, err := redis.DialTimeout("tcp",
+				configuration.Redis.Addr,
+				configuration.Redis.DialTimeout,
+				configuration.Redis.ReadTimeout,
+				configuration.Redis.WriteTimeout)
+			if err != nil {
+				ctxu.GetLogger(app).Errorf("error connecting to redis instance %s: %v",
+					configuration.Redis.Addr, err)
+				done(err)
+				return nil, err
+			}
+
+			// authorize the connection
+			if configuration.Redis.Password != "" {
+				if _, err = conn.Do("AUTH", configuration.Redis.Password); err != nil {
+					defer conn.Close()
+					done(err)
+					return nil, err
+				}
+			}
+
+			// select the database to use
+			if configuration.Redis.DB != 0 {
+				if _, err = conn.Do("SELECT", configuration.Redis.DB); err != nil {
+					defer conn.Close()
+					done(err)
+					return nil, err
+				}
+			}
+
+			done(nil)
+			return conn, nil
+		},
+		MaxIdle:     configuration.Redis.Pool.MaxIdle,
+		MaxActive:   configuration.Redis.Pool.MaxActive,
+		IdleTimeout: configuration.Redis.Pool.IdleTimeout,
+		TestOnBorrow: func(c redis.Conn, t time.Time) error {
+			// TODO(stevvooe): We can probably do something more interesting
+			// here with the health package.
+			_, err := c.Do("PING")
+			return err
+		},
+		Wait: false, // if a connection is not avialable, proceed without cache.
+	}
+
+	app.redis = pool
+
+	// setup expvar
+	registry := expvar.Get("registry")
+	if registry == nil {
+		registry = expvar.NewMap("registry")
+	}
+
+	registry.(*expvar.Map).Set("redis", expvar.Func(func() interface{} {
+		return map[string]interface{}{
+			"Config": configuration.Redis,
+			"Active": app.redis.ActiveCount(),
+		}
+	}))
+}
+
 func (app *App) ServeHTTP(w http.ResponseWriter, r *http.Request) {
 	defer r.Body.Close() // ensure that request body is always closed.
 

registry/handlers/app_test.go

@@ -13,6 +13,7 @@ import (
 	"github.com/docker/distribution/registry/auth"
 	_ "github.com/docker/distribution/registry/auth/silly"
 	"github.com/docker/distribution/registry/storage"
+	"github.com/docker/distribution/registry/storage/cache"
 	"github.com/docker/distribution/registry/storage/driver/inmemory"
 	"golang.org/x/net/context"
 )
@@ -28,7 +29,7 @@ func TestAppDispatcher(t *testing.T) {
 		Context:  context.Background(),
 		router:   v2.Router(),
 		driver:   driver,
-		registry: storage.NewRegistryWithDriver(driver),
+		registry: storage.NewRegistryWithDriver(driver, cache.NewInMemoryLayerInfoCache()),
 	}
 	server := httptest.NewServer(app)
 	router := v2.Router()

registry/storage/blobstore.go

@@ -18,7 +18,8 @@ import (
 // abstraction, providing utility methods that support creating and traversing
 // backend links.
 type blobStore struct {
-	*registry
+	driver storagedriver.StorageDriver
+	pm     *pathMapper
 	ctx    context.Context
 }
 

registry/storage/cache/cache.go

@@ -0,0 +1,98 @@
+// Package cache provides facilities to speed up access to the storage
+// backend. Typically cache implementations deal with internal implementation
+// details at the backend level, rather than generalized caches for
+// distribution related interfaces. In other words, unless the cache is
+// specific to the storage package, it belongs in another package.
+package cache
+
+import (
+	"fmt"
+
+	"github.com/docker/distribution/digest"
+	"golang.org/x/net/context"
+)
+
+// ErrNotFound is returned when a meta item is not found.
+var ErrNotFound = fmt.Errorf("not found")
+
+// LayerMeta describes the backend location and length of layer data.
+type LayerMeta struct {
+	Path   string
+	Length int64
+}
+
+// LayerInfoCache is a driver-aware cache of layer metadata. Basically, it
+// provides a fast cache for checks against repository metadata, avoiding
+// round trips to backend storage. Note that this is different from a pure
+// layer cache, which would also provide access to backing data, as well. Such
+// a cache should be implemented as a middleware, rather than integrated with
+// the storage backend.
+//
+// Note that most implementations rely on the caller to do strict checks on on
+// repo and dgst arguments, since these are mostly used behind existing
+// implementations.
+type LayerInfoCache interface {
+	// Contains returns true if the repository with name contains the layer.
+	Contains(ctx context.Context, repo string, dgst digest.Digest) (bool, error)
+
+	// Add includes the layer in the given repository cache.
+	Add(ctx context.Context, repo string, dgst digest.Digest) error
+
+	// Meta provides the location of the layer on the backend and its size. Membership of a
+	// repository should be tested before using the result, if required.
+	Meta(ctx context.Context, dgst digest.Digest) (LayerMeta, error)
+
+	// SetMeta sets the meta data for the given layer.
+	SetMeta(ctx context.Context, dgst digest.Digest, meta LayerMeta) error
+}
+
+// base implements common checks between cache implementations. Note that
+// these are not full checks of input, since that should be done by the
+// caller.
+type base struct {
+	LayerInfoCache
+}
+
+func (b *base) Contains(ctx context.Context, repo string, dgst digest.Digest) (bool, error) {
+	if repo == "" {
+		return false, fmt.Errorf("cache: cannot check for empty repository name")
+	}
+
+	if dgst == "" {
+		return false, fmt.Errorf("cache: cannot check for empty digests")
+	}
+
+	return b.LayerInfoCache.Contains(ctx, repo, dgst)
+}
+
+func (b *base) Add(ctx context.Context, repo string, dgst digest.Digest) error {
+	if repo == "" {
+		return fmt.Errorf("cache: cannot add empty repository name")
+	}
+
+	if dgst == "" {
+		return fmt.Errorf("cache: cannot add empty digest")
+	}
+
+	return b.LayerInfoCache.Add(ctx, repo, dgst)
+}
+
+func (b *base) Meta(ctx context.Context, dgst digest.Digest) (LayerMeta, error) {
+	if dgst == "" {
+		return LayerMeta{}, fmt.Errorf("cache: cannot get meta for empty digest")
+	}
+
+	return b.LayerInfoCache.Meta(ctx, dgst)
+}
+
+func (b *base) SetMeta(ctx context.Context, dgst digest.Digest, meta LayerMeta) error {
+	if dgst == "" {
+		return fmt.Errorf("cache: cannot set meta for empty digest")
+	}
+
+	if meta.Path == "" {
+		return fmt.Errorf("cache: cannot set empty path for meta")
+	}
+
+	return b.LayerInfoCache.SetMeta(ctx, dgst, meta)
+}

registry/storage/cache/cache_test.go

@@ -0,0 +1,86 @@
+package cache
+
+import (
+	"testing"
+
+	"golang.org/x/net/context"
+)
+
+// checkLayerInfoCache takes a cache implementation through a common set of
+// operations. If adding new tests, please add them here so new
+// implementations get the benefit.
+func checkLayerInfoCache(t *testing.T, lic LayerInfoCache) {
+	ctx := context.Background()
+
+	exists, err := lic.Contains(ctx, "", "fake:abc")
+	if err == nil {
+		t.Fatalf("expected error checking for cache item with empty repo")
+	}
+
+	exists, err = lic.Contains(ctx, "foo/bar", "")
+	if err == nil {
+		t.Fatalf("expected error checking for cache item with empty digest")
+	}
+
+	exists, err = lic.Contains(ctx, "foo/bar", "fake:abc")
+	if err != nil {
+		t.Fatalf("unexpected error checking for cache item: %v", err)
+	}
+
+	if exists {
+		t.Fatalf("item should not exist")
+	}
+
+	if err := lic.Add(ctx, "", "fake:abc"); err == nil {
+		t.Fatalf("expected error adding cache item with empty name")
+	}
+
+	if err := lic.Add(ctx, "foo/bar", ""); err == nil {
+		t.Fatalf("expected error adding cache item with empty digest")
+	}
+
+	if err := lic.Add(ctx, "foo/bar", "fake:abc"); err != nil {
+		t.Fatalf("unexpected error adding item: %v", err)
+	}
+
+	exists, err = lic.Contains(ctx, "foo/bar", "fake:abc")
+	if err != nil {
+		t.Fatalf("unexpected error checking for cache item: %v", err)
+	}
+
+	if !exists {
+		t.Fatalf("item should exist")
+	}
+
+	_, err = lic.Meta(ctx, "")
+	if err == nil || err == ErrNotFound {
+		t.Fatalf("expected error getting meta for cache item with empty digest")
+	}
+
+	_, err = lic.Meta(ctx, "fake:abc")
+	if err != ErrNotFound {
+		t.Fatalf("expected unknown layer error getting meta for cache item with empty digest")
+	}
+
+	if err = lic.SetMeta(ctx, "", LayerMeta{}); err == nil {
+		t.Fatalf("expected error setting meta for cache item with empty digest")
+	}
+
+	if err = lic.SetMeta(ctx, "foo/bar", LayerMeta{}); err == nil {
+		t.Fatalf("expected error setting meta for cache item with empty meta")
+	}
+
+	expected := LayerMeta{Path: "/foo/bar", Length: 20}
+	if err := lic.SetMeta(ctx, "foo/bar", expected); err != nil {
+		t.Fatalf("unexpected error setting meta: %v", err)
+	}
+
+	meta, err := lic.Meta(ctx, "foo/bar")
+	if err != nil {
+		t.Fatalf("unexpected error getting meta: %v", err)
+	}
+
+	if meta != expected {
+		t.Fatalf("retrieved meta data did not match: %v", err)
+	}
+}

registry/storage/cache/memory.go

@@ -0,0 +1,63 @@
+package cache
+
+import (
+	"github.com/docker/distribution/digest"
+	"golang.org/x/net/context"
+)
+
+// inmemoryLayerInfoCache is a map-based implementation of LayerInfoCache.
+type inmemoryLayerInfoCache struct {
+	membership map[string]map[digest.Digest]struct{}
+	meta       map[digest.Digest]LayerMeta
+}
+
+// NewInMemoryLayerInfoCache provides an implementation of LayerInfoCache that
+// stores results in memory.
+func NewInMemoryLayerInfoCache() LayerInfoCache {
+	return &base{&inmemoryLayerInfoCache{
+		membership: make(map[string]map[digest.Digest]struct{}),
+		meta:       make(map[digest.Digest]LayerMeta),
+	}}
+}
+
+func (ilic *inmemoryLayerInfoCache) Contains(ctx context.Context, repo string, dgst digest.Digest) (bool, error) {
+	members, ok := ilic.membership[repo]
+	if !ok {
+		return false, nil
+	}
+
+	_, ok = members[dgst]
+	return ok, nil
+}
+
+// Add adds the layer to the redis repository blob set.
+func (ilic *inmemoryLayerInfoCache) Add(ctx context.Context, repo string, dgst digest.Digest) error {
+	members, ok := ilic.membership[repo]
+	if !ok {
+		members = make(map[digest.Digest]struct{})
+		ilic.membership[repo] = members
+	}
+
+	members[dgst] = struct{}{}
+
+	return nil
+}
+
+// Meta retrieves the layer meta data from the redis hash, returning
+// ErrUnknownLayer if not found.
+func (ilic *inmemoryLayerInfoCache) Meta(ctx context.Context, dgst digest.Digest) (LayerMeta, error) {
+	meta, ok := ilic.meta[dgst]
+	if !ok {
+		return LayerMeta{}, ErrNotFound
+	}
+
+	return meta, nil
+}
+
+// SetMeta sets the meta data for the given digest using a redis hash. A hash
+// is used here since we may store unrelated fields about a layer in the
+// future.
+func (ilic *inmemoryLayerInfoCache) SetMeta(ctx context.Context, dgst digest.Digest, meta LayerMeta) error {
+	ilic.meta[dgst] = meta
+	return nil
+}

registry/storage/cache/memory_test.go

@@ -0,0 +1,9 @@
+package cache
+
+import "testing"
+
+// TestInMemoryLayerInfoCache checks the in memory implementation is working
+// correctly.
+func TestInMemoryLayerInfoCache(t *testing.T) {
+	checkLayerInfoCache(t, NewInMemoryLayerInfoCache())
+}

registry/storage/cache/redis.go

@@ -0,0 +1,98 @@
+package cache
+
+import (
+	ctxu "github.com/docker/distribution/context"
+	"github.com/docker/distribution/digest"
+	"github.com/garyburd/redigo/redis"
+	"golang.org/x/net/context"
+)
+
+// redisLayerInfoCache provides an implementation of storage.LayerInfoCache
+// based on redis. Layer info is stored in two parts. The first provide fast
+// access to repository membership through a redis set for each repo. The
+// second is a redis hash keyed by the digest of the layer, providing path and
+// length information. Note that there is no implied relationship between
+// these two caches. The layer may exist in one, both or none and the code
+// must be written this way.
+type redisLayerInfoCache struct {
+	pool *redis.Pool
+
+	// TODO(stevvooe): We use a pool because we don't have great control over
+	// the cache lifecycle to manage connections. A new connection if fetched
+	// for each operation. Once we have better lifecycle management of the
+	// request objects, we can change this to a connection.
+}
+
+// NewRedisLayerInfoCache returns a new redis-based LayerInfoCache using the
+// provided redis connection pool.
+func NewRedisLayerInfoCache(pool *redis.Pool) LayerInfoCache {
+	return &base{&redisLayerInfoCache{
+		pool: pool,
+	}}
+}
+
+// Contains does a membership check on the repository blob set in redis. This
+// is used as an access check before looking up global path information. If
+// false is returned, the caller should still check the backend to if it
+// exists elsewhere.
+func (rlic *redisLayerInfoCache) Contains(ctx context.Context, repo string, dgst digest.Digest) (bool, error) {
+	conn := rlic.pool.Get()
+	defer conn.Close()
+
+	ctxu.GetLogger(ctx).Debugf("(*redisLayerInfoCache).Contains(%q, %q)", repo, dgst)
+	return redis.Bool(conn.Do("SISMEMBER", rlic.repositoryBlobSetKey(repo), dgst))
+}
+
+// Add adds the layer to the redis repository blob set.
+func (rlic *redisLayerInfoCache) Add(ctx context.Context, repo string, dgst digest.Digest) error {
+	conn := rlic.pool.Get()
+	defer conn.Close()
+
+	ctxu.GetLogger(ctx).Debugf("(*redisLayerInfoCache).Add(%q, %q)", repo, dgst)
+	_, err := conn.Do("SADD", rlic.repositoryBlobSetKey(repo), dgst)
+	return err
+}
+
+// Meta retrieves the layer meta data from the redis hash, returning
+// ErrUnknownLayer if not found.
+func (rlic *redisLayerInfoCache) Meta(ctx context.Context, dgst digest.Digest) (LayerMeta, error) {
+	conn := rlic.pool.Get()
+	defer conn.Close()
+
+	reply, err := redis.Values(conn.Do("HMGET", rlic.blobMetaHashKey(dgst), "path", "length"))
+	if err != nil {
+		return LayerMeta{}, err
+	}
+
+	if len(reply) < 2 || reply[0] == nil || reply[1] == nil {
+		return LayerMeta{}, ErrNotFound
+	}
+
+	var meta LayerMeta
+	if _, err := redis.Scan(reply, &meta.Path, &meta.Length); err != nil {
+		return LayerMeta{}, err
+	}
+
+	return meta, nil
+}
+
+// SetMeta sets the meta data for the given digest using a redis hash. A hash
+// is used here since we may store unrelated fields about a layer in the
+// future.
+func (rlic *redisLayerInfoCache) SetMeta(ctx context.Context, dgst digest.Digest, meta LayerMeta) error {
+	conn := rlic.pool.Get()
+	defer conn.Close()
+
+	_, err := conn.Do("HMSET", rlic.blobMetaHashKey(dgst), "path", meta.Path, "length", meta.Length)
+	return err
+}
+
+// repositoryBlobSetKey returns the key for the blob set in the cache.
+func (rlic *redisLayerInfoCache) repositoryBlobSetKey(repo string) string {
+	return "repository::" + repo + "::blobs"
+}
+
+// blobMetaHashKey returns the cache key for immutable blob meta data.
+func (rlic *redisLayerInfoCache) blobMetaHashKey(dgst digest.Digest) string {
+	return "blobs::" + dgst.String()
+}

registry/storage/cache/redis_test.go

@@ -0,0 +1,50 @@
+package cache
+
+import (
+	"flag"
+	"os"
+	"testing"
+	"time"
+
+	"github.com/garyburd/redigo/redis"
+)
+
+var redisAddr string
+
+func init() {
+	flag.StringVar(&redisAddr, "test.registry.storage.cache.redis.addr", "", "configure the address of a test instance of redis")
+}
+
+// TestRedisLayerInfoCache exercises a live redis instance using the cache
+// implementation.
+func TestRedisLayerInfoCache(t *testing.T) {
+	if redisAddr == "" {
+		// fallback to an environement variable
+		redisAddr = os.Getenv("TEST_REGISTRY_STORAGE_CACHE_REDIS_ADDR")
+	}
+
+	if redisAddr == "" {
+		// skip if still not set
+		t.Skip("please set -registry.storage.cache.redis to test layer info cache against redis")
+	}
+
+	pool := &redis.Pool{
+		Dial: func() (redis.Conn, error) {
+			return redis.Dial("tcp", redisAddr)
+		},
+		MaxIdle:   1,
+		MaxActive: 2,
+		TestOnBorrow: func(c redis.Conn, t time.Time) error {
+			_, err := c.Do("PING")
+			return err
+		},
+		Wait: false, // if a connection is not avialable, proceed without cache.
+	}
+
+	// Clear the database
+	if _, err := pool.Get().Do("FLUSHDB"); err != nil {
+		t.Fatalf("unexpected error flushing redis db: %v", err)
+	}
+
+	checkLayerInfoCache(t, NewRedisLayerInfoCache(pool))
+}

registry/storage/filereader.go

@@ -27,8 +27,8 @@ type fileReader struct {
 
 	// identifying fields
 	path    string
-	size    int64 // size is the total layer size, must be set.
-	modtime time.Time
+	size    int64     // size is the total size, must be set.
+	modtime time.Time // TODO(stevvooe): This is not needed anymore.
 
 	// mutable fields
 	rc     io.ReadCloser // remote read closer

registry/storage/layer_test.go

@@ -11,6 +11,7 @@ import (
 
 	"github.com/docker/distribution"
 	"github.com/docker/distribution/digest"
+	"github.com/docker/distribution/registry/storage/cache"
 	storagedriver "github.com/docker/distribution/registry/storage/driver"
 	"github.com/docker/distribution/registry/storage/driver/inmemory"
 	"github.com/docker/distribution/testutil"
@@ -35,7 +36,7 @@ func TestSimpleLayerUpload(t *testing.T) {
 	ctx := context.Background()
 	imageName := "foo/bar"
 	driver := inmemory.New()
-	registry := NewRegistryWithDriver(driver)
+	registry := NewRegistryWithDriver(driver, cache.NewInMemoryLayerInfoCache())
 	repository, err := registry.Repository(ctx, imageName)
 	if err != nil {
 		t.Fatalf("unexpected error getting repo: %v", err)
@@ -143,7 +144,7 @@ func TestSimpleLayerRead(t *testing.T) {
 	ctx := context.Background()
 	imageName := "foo/bar"
 	driver := inmemory.New()
-	registry := NewRegistryWithDriver(driver)
+	registry := NewRegistryWithDriver(driver, cache.NewInMemoryLayerInfoCache())
 	repository, err := registry.Repository(ctx, imageName)
 	if err != nil {
 		t.Fatalf("unexpected error getting repo: %v", err)
@@ -180,7 +181,7 @@ func TestSimpleLayerRead(t *testing.T) {
 		t.Fatalf("unexpected error fetching non-existent layer: %v", err)
 	}
 
-	randomLayerDigest, err := writeTestLayer(driver, ls.(*layerStore).repository.pm, imageName, dgst, randomLayerReader)
+	randomLayerDigest, err := writeTestLayer(driver, defaultPathMapper, imageName, dgst, randomLayerReader)
 	if err != nil {
 		t.Fatalf("unexpected error writing test layer: %v", err)
 	}
@@ -252,7 +253,7 @@ func TestLayerUploadZeroLength(t *testing.T) {
 	ctx := context.Background()
 	imageName := "foo/bar"
 	driver := inmemory.New()
-	registry := NewRegistryWithDriver(driver)
+	registry := NewRegistryWithDriver(driver, cache.NewInMemoryLayerInfoCache())
 	repository, err := registry.Repository(ctx, imageName)
 	if err != nil {
 		t.Fatalf("unexpected error getting repo: %v", err)

registry/storage/layercache.go

@@ -0,0 +1,202 @@
+package storage
+
+import (
+	"expvar"
+	"sync/atomic"
+	"time"
+
+	"github.com/docker/distribution"
+	ctxu "github.com/docker/distribution/context"
+	"github.com/docker/distribution/digest"
+	"github.com/docker/distribution/registry/storage/cache"
+	"github.com/docker/distribution/registry/storage/driver"
+	"golang.org/x/net/context"
+)
+
+// cachedLayerService implements the layer service with path-aware caching,
+// using a LayerInfoCache interface.
+type cachedLayerService struct {
+	distribution.LayerService // upstream layer service
+	repository                distribution.Repository
+	ctx                       context.Context
+	driver                    driver.StorageDriver
+	*blobStore                // global blob store
+	cache                     cache.LayerInfoCache
+}
+
+// Exists checks for existence of the digest in the cache, immediately
+// returning if it exists for the repository. If not, the upstream is checked.
+// When a positive result is found, it is written into the cache.
+func (lc *cachedLayerService) Exists(dgst digest.Digest) (bool, error) {
+	ctxu.GetLogger(lc.ctx).Debugf("(*cachedLayerService).Exists(%q)", dgst)
+	now := time.Now()
+	defer func() {
+		// TODO(stevvooe): Replace this with a decent context-based metrics solution
+		ctxu.GetLoggerWithField(lc.ctx, "blob.exists.duration", time.Since(now)).
+			Infof("(*cachedLayerService).Exists(%q)", dgst)
+	}()
+
+	atomic.AddUint64(&layerInfoCacheMetrics.Exists.Requests, 1)
+	available, err := lc.cache.Contains(lc.ctx, lc.repository.Name(), dgst)
+	if err != nil {
+		ctxu.GetLogger(lc.ctx).Errorf("error checking availability of %v@%v: %v", lc.repository.Name(), dgst, err)
+		goto fallback
+	}
+
+	if available {
+		atomic.AddUint64(&layerInfoCacheMetrics.Exists.Hits, 1)
+		return true, nil
+	}
+
+fallback:
+	atomic.AddUint64(&layerInfoCacheMetrics.Exists.Misses, 1)
+	exists, err := lc.LayerService.Exists(dgst)
+	if err != nil {
+		return exists, err
+	}
+
+	if exists {
+		// we can only cache this if the existence is positive.
+		if err := lc.cache.Add(lc.ctx, lc.repository.Name(), dgst); err != nil {
+			ctxu.GetLogger(lc.ctx).Errorf("error adding %v@%v to cache: %v", lc.repository.Name(), dgst, err)
+		}
+	}
+
+	return exists, err
+}
+
+// Fetch checks for the availability of the layer in the repository via the
+// cache. If present, the metadata is resolved and the layer is returned. If
+// any operation fails, the layer is read directly from the upstream. The
+// results are cached, if possible.
+func (lc *cachedLayerService) Fetch(dgst digest.Digest) (distribution.Layer, error) {
+	ctxu.GetLogger(lc.ctx).Debugf("(*layerInfoCache).Fetch(%q)", dgst)
+	now := time.Now()
+	defer func() {
+		ctxu.GetLoggerWithField(lc.ctx, "blob.fetch.duration", time.Since(now)).
+			Infof("(*layerInfoCache).Fetch(%q)", dgst)
+	}()
+
+	atomic.AddUint64(&layerInfoCacheMetrics.Fetch.Requests, 1)
+	available, err := lc.cache.Contains(lc.ctx, lc.repository.Name(), dgst)
+	if err != nil {
+		ctxu.GetLogger(lc.ctx).Errorf("error checking availability of %v@%v: %v", lc.repository.Name(), dgst, err)
+		goto fallback
+	}
+
+	if available {
+		// fast path: get the layer info and return
+		meta, err := lc.cache.Meta(lc.ctx, dgst)
+		if err != nil {
+			ctxu.GetLogger(lc.ctx).Errorf("error fetching %v@%v from cache: %v", lc.repository.Name(), dgst, err)
+			goto fallback
+		}
+
+		atomic.AddUint64(&layerInfoCacheMetrics.Fetch.Hits, 1)
+		return newLayerReader(lc.driver, dgst, meta.Path, meta.Length)
+	}
+
+	// NOTE(stevvooe): Unfortunately, the cache here only makes checks for
+	// existing layers faster. We'd have to provide more careful
+	// synchronization with the backend to make the missing case as fast.
+
+fallback:
+	atomic.AddUint64(&layerInfoCacheMetrics.Fetch.Misses, 1)
+	layer, err := lc.LayerService.Fetch(dgst)
+	if err != nil {
+		return nil, err
+	}
+
+	// add the layer to the repository
+	if err := lc.cache.Add(lc.ctx, lc.repository.Name(), dgst); err != nil {
+		ctxu.GetLogger(lc.ctx).
+			Errorf("error caching repository relationship for %v@%v: %v", lc.repository.Name(), dgst, err)
+	}
+
+	// lookup layer path and add it to the cache, if it succeds. Note that we
+	// still return the layer even if we have trouble caching it.
+	if path, err := lc.resolveLayerPath(layer); err != nil {
+		ctxu.GetLogger(lc.ctx).
+			Errorf("error resolving path while caching %v@%v: %v", lc.repository.Name(), dgst, err)
+	} else {
+		// add the layer to the cache once we've resolved the path.
+		if err := lc.cache.SetMeta(lc.ctx, dgst, cache.LayerMeta{Path: path, Length: layer.Length()}); err != nil {
+			ctxu.GetLogger(lc.ctx).Errorf("error adding meta for %v@%v to cache: %v", lc.repository.Name(), dgst, err)
+		}
+	}
+
+	return layer, err
+}
+
+// extractLayerInfo pulls the layerInfo from the layer, attempting to get the
+// path information from either the concrete object or by resolving the
+// primary blob store path.
+func (lc *cachedLayerService) resolveLayerPath(layer distribution.Layer) (path string, err error) {
+	// try and resolve the type and driver, so we don't have to traverse links
+	switch v := layer.(type) {
+	case *layerReader:
+		// only set path if we have same driver instance.
+		if v.driver == lc.driver {
+			return v.path, nil
+		}
+	}
+
+	ctxu.GetLogger(lc.ctx).Warnf("resolving layer path during cache lookup (%v@%v)", lc.repository.Name(), layer.Digest())
+	// we have to do an expensive stat to resolve the layer location but no
+	// need to check the link, since we already have layer instance for this
+	// repository.
+	bp, err := lc.blobStore.path(layer.Digest())
+	if err != nil {
+		return "", err
+	}
+
+	return bp, nil
+}
+
+// layerInfoCacheMetrics keeps track of cache metrics for layer info cache
+// requests. Note this is kept globally and made available via expvar. For
+// more detailed metrics, its recommend to instrument a particular cache
+// implementation.
+var layerInfoCacheMetrics struct {
+	// Exists tracks calls to the Exists caches.
+	Exists struct {
+		Requests uint64
+		Hits     uint64
+		Misses   uint64
+	}
+
+	// Fetch tracks calls to the fetch caches.
+	Fetch struct {
+		Requests uint64
+		Hits     uint64
+		Misses   uint64
+	}
+}
+
+func init() {
+	registry := expvar.Get("registry")
+	if registry == nil {
+		registry = expvar.NewMap("registry")
+	}
+
+	cache := registry.(*expvar.Map).Get("cache")
+	if cache == nil {
+		cache = &expvar.Map{}
+		cache.(*expvar.Map).Init()
+		registry.(*expvar.Map).Set("cache", cache)
+	}
+
+	storage := cache.(*expvar.Map).Get("storage")
+	if storage == nil {
+		storage = &expvar.Map{}
+		storage.(*expvar.Map).Init()
+		cache.(*expvar.Map).Set("storage", storage)
+	}
+
+	storage.(*expvar.Map).Set("layerinfo", expvar.Func(func() interface{} {
+		// no need for synchronous access: the increments are atomic and
+		// during reading, we don't care if the data is up to date. The
+		// numbers will always *eventually* be reported correctly.
+		return layerInfoCacheMetrics
+	}))
+}

registry/storage/layerreader.go

@@ -17,6 +17,21 @@ type layerReader struct {
 	digest digest.Digest
 }
 
+// newLayerReader returns a new layerReader with the digest, path and length,
+// eliding round trips to the storage backend.
+func newLayerReader(driver driver.StorageDriver, dgst digest.Digest, path string, length int64) (*layerReader, error) {
+	fr := &fileReader{
+		driver: driver,
+		path:   path,
+		size:   length,
+	}
+
+	return &layerReader{
+		fileReader: *fr,
+		digest:     dgst,
+	}, nil
+}
+
 var _ distribution.Layer = &layerReader{}
 
 func (lr *layerReader) Digest() digest.Digest {

registry/storage/manifeststore_test.go

@@ -6,6 +6,8 @@ import (
 	"reflect"
 	"testing"
 
+	"github.com/docker/distribution/registry/storage/cache"
+
 	"github.com/docker/distribution"
 	"github.com/docker/distribution/digest"
 	"github.com/docker/distribution/manifest"
@@ -28,7 +30,7 @@ type manifestStoreTestEnv struct {
 func newManifestStoreTestEnv(t *testing.T, name, tag string) *manifestStoreTestEnv {
 	ctx := context.Background()
 	driver := inmemory.New()
-	registry := NewRegistryWithDriver(driver)
+	registry := NewRegistryWithDriver(driver, cache.NewInMemoryLayerInfoCache())
 
 	repo, err := registry.Repository(ctx, name)
 	if err != nil {

registry/storage/registry.go

@@ -3,6 +3,7 @@ package storage
 import (
 	"github.com/docker/distribution"
 	"github.com/docker/distribution/registry/api/v2"
+	"github.com/docker/distribution/registry/storage/cache"
 	storagedriver "github.com/docker/distribution/registry/storage/driver"
 	"golang.org/x/net/context"
 )
@@ -13,25 +14,26 @@ type registry struct {
 	driver         storagedriver.StorageDriver
 	pm             *pathMapper
 	blobStore      *blobStore
+	layerInfoCache cache.LayerInfoCache
 }
 
 // NewRegistryWithDriver creates a new registry instance from the provided
 // driver. The resulting registry may be shared by multiple goroutines but is
 // cheap to allocate.
-func NewRegistryWithDriver(driver storagedriver.StorageDriver) distribution.Registry {
-	bs := &blobStore{}
+func NewRegistryWithDriver(driver storagedriver.StorageDriver, layerInfoCache cache.LayerInfoCache) distribution.Registry {
+	bs := &blobStore{
+		driver: driver,
+		pm:     defaultPathMapper,
+	}
 
-	reg := &registry{
+	return &registry{
 		driver:    driver,
 		blobStore: bs,
 
 		// TODO(sday): This should be configurable.
 		pm:             defaultPathMapper,
+		layerInfoCache: layerInfoCache,
 	}
-
-	reg.blobStore.registry = reg
-
-	return reg
 }
 
 // Repository returns an instance of the repository tied to the registry.
@@ -83,9 +85,29 @@ func (repo *repository) Manifests() distribution.ManifestService {
 // may be context sensitive in the future. The instance should be used similar
 // to a request local.
 func (repo *repository) Layers() distribution.LayerService {
-	return &layerStore{
+	ls := &layerStore{
 		repository: repo,
 	}
+
+	if repo.registry.layerInfoCache != nil {
+		// TODO(stevvooe): This is not the best place to setup a cache. We would
+		// really like to decouple the cache from the backend but also have the
+		// manifeset service use the layer service cache. For now, we can simply
+		// integrate the cache directly. The main issue is that we have layer
+		// access and layer data coupled in a single object. Work is already under
+		// way to decouple this.
+
+		return &cachedLayerService{
+			LayerService: ls,
+			repository:   repo,
+			ctx:          repo.ctx,
+			driver:       repo.driver,
+			blobStore:    repo.blobStore,
+			cache:        repo.registry.layerInfoCache,
+		}
+	}
+
+	return ls
 }
 
 func (repo *repository) Signatures() distribution.SignatureService {