First public commit

55 files changed, 7616 insertions, 0 deletions

diff --git a/vendor/github.com/go-kit/kit/LICENSE b/vendor/github.com/go-kit/kit/LICENSE
new file mode 100644
index 0000000..9d83342
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/LICENSE
@@ -0,0 +1,22 @@
+The MIT License (MIT)
+
+Copyright (c) 2015 Peter Bourgon
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
diff --git a/vendor/github.com/go-kit/kit/log/README.md b/vendor/github.com/go-kit/kit/log/README.md
new file mode 100644
index 0000000..7222f80
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/README.md
@@ -0,0 +1,147 @@
+# package log
+
+`package log` provides a minimal interface for structured logging in services.
+It may be wrapped to encode conventions, enforce type-safety, provide leveled
+logging, and so on. It can be used for both typical application log events,
+and log-structured data streams.
+
+## Structured logging
+
+Structured logging is, basically, conceding to the reality that logs are
+_data_, and warrant some level of schematic rigor. Using a stricter,
+key/value-oriented message format for our logs, containing contextual and
+semantic information, makes it much easier to get insight into the
+operational activity of the systems we build. Consequently, `package log` is
+of the strong belief that "[the benefits of structured logging outweigh the
+minimal effort involved](https://www.thoughtworks.com/radar/techniques/structured-logging)".
+
+Migrating from unstructured to structured logging is probably a lot easier
+than you'd expect.
+
+```go
+// Unstructured
+log.Printf("HTTP server listening on %s", addr)
+
+// Structured
+logger.Log("transport", "HTTP", "addr", addr, "msg", "listening")
+```
+
+## Usage
+
+### Typical application logging
+
+```go
+w := log.NewSyncWriter(os.Stderr)
+logger := log.NewLogfmtLogger(w)
+logger.Log("question", "what is the meaning of life?", "answer", 42)
+
+// Output:
+// question="what is the meaning of life?" answer=42
+```
+
+### Contextual Loggers
+
+```go
+func main() {
+	var logger log.Logger
+	logger = log.NewLogfmtLogger(log.NewSyncWriter(os.Stderr))
+	logger = log.With(logger, "instance_id", 123)
+
+	logger.Log("msg", "starting")
+	NewWorker(log.With(logger, "component", "worker")).Run()
+	NewSlacker(log.With(logger, "component", "slacker")).Run()
+}
+
+// Output:
+// instance_id=123 msg=starting
+// instance_id=123 component=worker msg=running
+// instance_id=123 component=slacker msg=running
+```
+
+### Interact with stdlib logger
+
+Redirect stdlib logger to Go kit logger.
+
+```go
+import (
+	"os"
+	stdlog "log"
+	kitlog "github.com/go-kit/kit/log"
+)
+
+func main() {
+	logger := kitlog.NewJSONLogger(kitlog.NewSyncWriter(os.Stdout))
+	stdlog.SetOutput(kitlog.NewStdlibAdapter(logger))
+	stdlog.Print("I sure like pie")
+}
+
+// Output:
+// {"msg":"I sure like pie","ts":"2016/01/01 12:34:56"}
+```
+
+Or, if, for legacy reasons, you need to pipe all of your logging through the
+stdlib log package, you can redirect Go kit logger to the stdlib logger.
+
+```go
+logger := kitlog.NewLogfmtLogger(kitlog.StdlibWriter{})
+logger.Log("legacy", true, "msg", "at least it's something")
+
+// Output:
+// 2016/01/01 12:34:56 legacy=true msg="at least it's something"
+```
+
+### Timestamps and callers
+
+```go
+var logger log.Logger
+logger = log.NewLogfmtLogger(log.NewSyncWriter(os.Stderr))
+logger = log.With(logger, "ts", log.DefaultTimestampUTC, "caller", log.DefaultCaller)
+
+logger.Log("msg", "hello")
+
+// Output:
+// ts=2016-01-01T12:34:56Z caller=main.go:15 msg=hello
+```
+
+## Supported output formats
+
+- [Logfmt](https://brandur.org/logfmt) ([see also](https://blog.codeship.com/logfmt-a-log-format-thats-easy-to-read-and-write))
+- JSON
+
+## Enhancements
+
+`package log` is centered on the one-method Logger interface.
+
+```go
+type Logger interface {
+	Log(keyvals ...interface{}) error
+}
+```
+
+This interface, and its supporting code like is the product of much iteration
+and evaluation. For more details on the evolution of the Logger interface,
+see [The Hunt for a Logger Interface](http://go-talks.appspot.com/github.com/ChrisHines/talks/structured-logging/structured-logging.slide#1),
+a talk by [Chris Hines](https://github.com/ChrisHines).
+Also, please see
+[#63](https://github.com/go-kit/kit/issues/63),
+[#76](https://github.com/go-kit/kit/pull/76),
+[#131](https://github.com/go-kit/kit/issues/131),
+[#157](https://github.com/go-kit/kit/pull/157),
+[#164](https://github.com/go-kit/kit/issues/164), and
+[#252](https://github.com/go-kit/kit/pull/252)
+to review historical conversations about package log and the Logger interface.
+
+Value-add packages and suggestions,
+like improvements to [the leveled logger](https://godoc.org/github.com/go-kit/kit/log/level),
+are of course welcome. Good proposals should
+
+- Be composable with [contextual loggers](https://godoc.org/github.com/go-kit/kit/log#With),
+- Not break the behavior of [log.Caller](https://godoc.org/github.com/go-kit/kit/log#Caller) in any wrapped contextual loggers, and
+- Be friendly to packages that accept only an unadorned log.Logger.
+
+## Benchmarks & comparisons
+
+There are a few Go logging benchmarks and comparisons that include Go kit's package log.
+
+- [imkira/go-loggers-bench](https://github.com/imkira/go-loggers-bench) includes kit/log
+- [uber-common/zap](https://github.com/uber-common/zap), a zero-alloc logging library, includes a comparison with kit/log
diff --git a/vendor/github.com/go-kit/kit/log/doc.go b/vendor/github.com/go-kit/kit/log/doc.go
new file mode 100644
index 0000000..918c0af
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/doc.go
@@ -0,0 +1,116 @@
+// Package log provides a structured logger.
+//
+// Structured logging produces logs easily consumed later by humans or
+// machines. Humans might be interested in debugging errors, or tracing
+// specific requests. Machines might be interested in counting interesting
+// events, or aggregating information for off-line processing. In both cases,
+// it is important that the log messages are structured and actionable.
+// Package log is designed to encourage both of these best practices.
+//
+// Basic Usage
+//
+// The fundamental interface is Logger. Loggers create log events from
+// key/value data. The Logger interface has a single method, Log, which
+// accepts a sequence of alternating key/value pairs, which this package names
+// keyvals.
+//
+//    type Logger interface {
+//        Log(keyvals ...interface{}) error
+//    }
+//
+// Here is an example of a function using a Logger to create log events.
+//
+//    func RunTask(task Task, logger log.Logger) string {
+//        logger.Log("taskID", task.ID, "event", "starting task")
+//        ...
+//        logger.Log("taskID", task.ID, "event", "task complete")
+//    }
+//
+// The keys in the above example are "taskID" and "event". The values are
+// task.ID, "starting task", and "task complete". Every key is followed
+// immediately by its value.
+//
+// Keys are usually plain strings. Values may be any type that has a sensible
+// encoding in the chosen log format. With structured logging it is a good
+// idea to log simple values without formatting them. This practice allows
+// the chosen logger to encode values in the most appropriate way.
+//
+// Contextual Loggers
+//
+// A contextual logger stores keyvals that it includes in all log events.
+// Building appropriate contextual loggers reduces repetition and aids
+// consistency in the resulting log output. With and WithPrefix add context to
+// a logger. We can use With to improve the RunTask example.
+//
+//    func RunTask(task Task, logger log.Logger) string {
+//        logger = log.With(logger, "taskID", task.ID)
+//        logger.Log("event", "starting task")
+//        ...
+//        taskHelper(task.Cmd, logger)
+//        ...
+//        logger.Log("event", "task complete")
+//    }
+//
+// The improved version emits the same log events as the original for the
+// first and last calls to Log. Passing the contextual logger to taskHelper
+// enables each log event created by taskHelper to include the task.ID even
+// though taskHelper does not have access to that value. Using contextual
+// loggers this way simplifies producing log output that enables tracing the
+// life cycle of individual tasks. (See the Contextual example for the full
+// code of the above snippet.)
+//
+// Dynamic Contextual Values
+//
+// A Valuer function stored in a contextual logger generates a new value each
+// time an event is logged. The Valuer example demonstrates how this feature
+// works.
+//
+// Valuers provide the basis for consistently logging timestamps and source
+// code location. The log package defines several valuers for that purpose.
+// See Timestamp, DefaultTimestamp, DefaultTimestampUTC, Caller, and
+// DefaultCaller. A common logger initialization sequence that ensures all log
+// entries contain a timestamp and source location looks like this:
+//
+//    logger := log.NewLogfmtLogger(log.NewSyncWriter(os.Stdout))
+//    logger = log.With(logger, "ts", log.DefaultTimestampUTC, "caller", log.DefaultCaller)
+//
+// Concurrent Safety
+//
+// Applications with multiple goroutines want each log event written to the
+// same logger to remain separate from other log events. Package log provides
+// two simple solutions for concurrent safe logging.
+//
+// NewSyncWriter wraps an io.Writer and serializes each call to its Write
+// method. Using a SyncWriter has the benefit that the smallest practical
+// portion of the logging logic is performed within a mutex, but it requires
+// the formatting Logger to make only one call to Write per log event.
+//
+// NewSyncLogger wraps any Logger and serializes each call to its Log method.
+// Using a SyncLogger has the benefit that it guarantees each log event is
+// handled atomically within the wrapped logger, but it typically serializes
+// both the formatting and output logic. Use a SyncLogger if the formatting
+// logger may perform multiple writes per log event.
+//
+// Error Handling
+//
+// This package relies on the practice of wrapping or decorating loggers with
+// other loggers to provide composable pieces of functionality. It also means
+// that Logger.Log must return an error because some
+// implementations—especially those that output log data to an io.Writer—may
+// encounter errors that cannot be handled locally. This in turn means that
+// Loggers that wrap other loggers should return errors from the wrapped
+// logger up the stack.
+//
+// Fortunately, the decorator pattern also provides a way to avoid the
+// necessity to check for errors every time an application calls Logger.Log.
+// An application required to panic whenever its Logger encounters
+// an error could initialize its logger as follows.
+//
+//    fmtlogger := log.NewLogfmtLogger(log.NewSyncWriter(os.Stdout))
+//    logger := log.LoggerFunc(func(keyvals ...interface{}) error {
+//        if err := fmtlogger.Log(keyvals...); err != nil {
+//            panic(err)
+//        }
+//        return nil
+//    })
+package log
diff --git a/vendor/github.com/go-kit/kit/log/json_logger.go b/vendor/github.com/go-kit/kit/log/json_logger.go
new file mode 100644
index 0000000..66094b4
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/json_logger.go
@@ -0,0 +1,89 @@
+package log
+
+import (
+	"encoding"
+	"encoding/json"
+	"fmt"
+	"io"
+	"reflect"
+)
+
+type jsonLogger struct {
+	io.Writer
+}
+
+// NewJSONLogger returns a Logger that encodes keyvals to the Writer as a
+// single JSON object. Each log event produces no more than one call to
+// w.Write. The passed Writer must be safe for concurrent use by multiple
+// goroutines if the returned Logger will be used concurrently.
+func NewJSONLogger(w io.Writer) Logger {
+	return &jsonLogger{w}
+}
+
+func (l *jsonLogger) Log(keyvals ...interface{}) error {
+	n := (len(keyvals) + 1) / 2 // +1 to handle case when len is odd
+	m := make(map[string]interface{}, n)
+	for i := 0; i < len(keyvals); i += 2 {
+		k := keyvals[i]
+		var v interface{} = ErrMissingValue
+		if i+1 < len(keyvals) {
+			v = keyvals[i+1]
+		}
+		merge(m, k, v)
+	}
+	return json.NewEncoder(l.Writer).Encode(m)
+}
+
+func merge(dst map[string]interface{}, k, v interface{}) {
+	var key string
+	switch x := k.(type) {
+	case string:
+		key = x
+	case fmt.Stringer:
+		key = safeString(x)
+	default:
+		key = fmt.Sprint(x)
+	}
+
+	// We want json.Marshaler and encoding.TextMarshaller to take priority over
+	// err.Error() and v.String(). But json.Marshall (called later) does that by
+	// default so we force a no-op if it's one of those 2 case.
+	switch x := v.(type) {
+	case json.Marshaler:
+	case encoding.TextMarshaler:
+	case error:
+		v = safeError(x)
+	case fmt.Stringer:
+		v = safeString(x)
+	}
+
+	dst[key] = v
+}
+
+func safeString(str fmt.Stringer) (s string) {
+	defer func() {
+		if panicVal := recover(); panicVal != nil {
+			if v := reflect.ValueOf(str); v.Kind() == reflect.Ptr && v.IsNil() {
+				s = "NULL"
+			} else {
+				panic(panicVal)
+			}
+		}
+	}()
+	s = str.String()
+	return
+}
+
+func safeError(err error) (s interface{}) {
+	defer func() {
+		if panicVal := recover(); panicVal != nil {
+			if v := reflect.ValueOf(err); v.Kind() == reflect.Ptr && v.IsNil() {
+				s = nil
+			} else {
+				panic(panicVal)
+			}
+		}
+	}()
+	s = err.Error()
+	return
+}
diff --git a/vendor/github.com/go-kit/kit/log/level/doc.go b/vendor/github.com/go-kit/kit/log/level/doc.go
new file mode 100644
index 0000000..feadc4c
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/level/doc.go
@@ -0,0 +1,22 @@
+// Package level implements leveled logging on top of package log. To use the
+// level package, create a logger as per normal in your func main, and wrap it
+// with level.NewFilter.
+//
+//    var logger log.Logger
+//    logger = log.NewLogfmtLogger(os.Stderr)
+//    logger = level.NewFilter(logger, level.AllowInfo()) // <--
+//    logger = log.With(logger, "ts", log.DefaultTimestampUTC)
+//
+// Then, at the callsites, use one of the level.Debug, Info, Warn, or Error
+// helper methods to emit leveled log events.
+//
+//    logger.Log("foo", "bar") // as normal, no level
+//    level.Debug(logger).Log("request_id", reqID, "trace_data", trace.Get())
+//    if value > 100 {
+//        level.Error(logger).Log("value", value)
+//    }
+//
+// NewFilter allows precise control over what happens when a log event is
+// emitted without a level key, or if a squelched level is used. Check the
+// Option functions for details.
+package level
diff --git a/vendor/github.com/go-kit/kit/log/level/level.go b/vendor/github.com/go-kit/kit/log/level/level.go
new file mode 100644
index 0000000..dd4ef60
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/level/level.go
@@ -0,0 +1,205 @@
+package level
+
+import "github.com/go-kit/kit/log"
+
+// Error returns a logger that includes a Key/ErrorValue pair.
+func Error(logger log.Logger) log.Logger {
+	return log.WithPrefix(logger, Key(), ErrorValue())
+}
+
+// Warn returns a logger that includes a Key/WarnValue pair.
+func Warn(logger log.Logger) log.Logger {
+	return log.WithPrefix(logger, Key(), WarnValue())
+}
+
+// Info returns a logger that includes a Key/InfoValue pair.
+func Info(logger log.Logger) log.Logger {
+	return log.WithPrefix(logger, Key(), InfoValue())
+}
+
+// Debug returns a logger that includes a Key/DebugValue pair.
+func Debug(logger log.Logger) log.Logger {
+	return log.WithPrefix(logger, Key(), DebugValue())
+}
+
+// NewFilter wraps next and implements level filtering. See the commentary on
+// the Option functions for a detailed description of how to configure levels.
+// If no options are provided, all leveled log events created with Debug,
+// Info, Warn or Error helper methods are squelched and non-leveled log
+// events are passed to next unmodified.
+func NewFilter(next log.Logger, options ...Option) log.Logger {
+	l := &logger{
+		next: next,
+	}
+	for _, option := range options {
+		option(l)
+	}
+	return l
+}
+
+type logger struct {
+	next           log.Logger
+	allowed        level
+	squelchNoLevel bool
+	errNotAllowed  error
+	errNoLevel     error
+}
+
+func (l *logger) Log(keyvals ...interface{}) error {
+	var hasLevel, levelAllowed bool
+	for i := 1; i < len(keyvals); i += 2 {
+		if v, ok := keyvals[i].(*levelValue); ok {
+			hasLevel = true
+			levelAllowed = l.allowed&v.level != 0
+			break
+		}
+	}
+	if !hasLevel && l.squelchNoLevel {
+		return l.errNoLevel
+	}
+	if hasLevel && !levelAllowed {
+		return l.errNotAllowed
+	}
+	return l.next.Log(keyvals...)
+}
+
+// Option sets a parameter for the leveled logger.
+type Option func(*logger)
+
+// AllowAll is an alias for AllowDebug.
+func AllowAll() Option {
+	return AllowDebug()
+}
+
+// AllowDebug allows error, warn, info and debug level log events to pass.
+func AllowDebug() Option {
+	return allowed(levelError | levelWarn | levelInfo | levelDebug)
+}
+
+// AllowInfo allows error, warn and info level log events to pass.
+func AllowInfo() Option {
+	return allowed(levelError | levelWarn | levelInfo)
+}
+
+// AllowWarn allows error and warn level log events to pass.
+func AllowWarn() Option {
+	return allowed(levelError | levelWarn)
+}
+
+// AllowError allows only error level log events to pass.
+func AllowError() Option {
+	return allowed(levelError)
+}
+
+// AllowNone allows no leveled log events to pass.
+func AllowNone() Option {
+	return allowed(0)
+}
+
+func allowed(allowed level) Option {
+	return func(l *logger) { l.allowed = allowed }
+}
+
+// ErrNotAllowed sets the error to return from Log when it squelches a log
+// event disallowed by the configured Allow[Level] option. By default,
+// ErrNotAllowed is nil; in this case the log event is squelched with no
+// error.
+func ErrNotAllowed(err error) Option {
+	return func(l *logger) { l.errNotAllowed = err }
+}
+
+// SquelchNoLevel instructs Log to squelch log events with no level, so that
+// they don't proceed through to the wrapped logger. If SquelchNoLevel is set
+// to true and a log event is squelched in this way, the error value
+// configured with ErrNoLevel is returned to the caller.
+func SquelchNoLevel(squelch bool) Option {
+	return func(l *logger) { l.squelchNoLevel = squelch }
+}
+
+// ErrNoLevel sets the error to return from Log when it squelches a log event
+// with no level. By default, ErrNoLevel is nil; in this case the log event is
+// squelched with no error.
+func ErrNoLevel(err error) Option {
+	return func(l *logger) { l.errNoLevel = err }
+}
+
+// NewInjector wraps next and returns a logger that adds a Key/level pair to
+// the beginning of log events that don't already contain a level. In effect,
+// this gives a default level to logs without a level.
+func NewInjector(next log.Logger, level Value) log.Logger {
+	return &injector{
+		next:  next,
+		level: level,
+	}
+}
+
+type injector struct {
+	next  log.Logger
+	level interface{}
+}
+
+func (l *injector) Log(keyvals ...interface{}) error {
+	for i := 1; i < len(keyvals); i += 2 {
+		if _, ok := keyvals[i].(*levelValue); ok {
+			return l.next.Log(keyvals...)
+		}
+	}
+	kvs := make([]interface{}, len(keyvals)+2)
+	kvs[0], kvs[1] = key, l.level
+	copy(kvs[2:], keyvals)
+	return l.next.Log(kvs...)
+}
+
+// Value is the interface that each of the canonical level values implement.
+// It contains unexported methods that prevent types from other packages from
+// implementing it and guaranteeing that NewFilter can distinguish the levels
+// defined in this package from all other values.
+type Value interface {
+	String() string
+	levelVal()
+}
+
+// Key returns the unique key added to log events by the loggers in this
+// package.
+func Key() interface{} { return key }
+
+// ErrorValue returns the unique value added to log events by Error.
+func ErrorValue() Value { return errorValue }
+
+// WarnValue returns the unique value added to log events by Warn.
+func WarnValue() Value { return warnValue }
+
+// InfoValue returns the unique value added to log events by Info.
+func InfoValue() Value { return infoValue }
+
+// DebugValue returns the unique value added to log events by Warn.
+func DebugValue() Value { return debugValue }
+
+var (
+	// key is of type interfae{} so that it allocates once during package
+	// initialization and avoids allocating every time the value is added to a
+	// []interface{} later.
+	key interface{} = "level"
+
+	errorValue = &levelValue{level: levelError, name: "error"}
+	warnValue  = &levelValue{level: levelWarn, name: "warn"}
+	infoValue  = &levelValue{level: levelInfo, name: "info"}
+	debugValue = &levelValue{level: levelDebug, name: "debug"}
+)
+
+type level byte
+
+const (
+	levelDebug level = 1 << iota
+	levelInfo
+	levelWarn
+	levelError
+)
+
+type levelValue struct {
+	name string
+	level
+}
+
+func (v *levelValue) String() string { return v.name }
+func (v *levelValue) levelVal()      {}
diff --git a/vendor/github.com/go-kit/kit/log/log.go b/vendor/github.com/go-kit/kit/log/log.go
new file mode 100644
index 0000000..66a9e2f
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/log.go
@@ -0,0 +1,135 @@
+package log
+
+import "errors"
+
+// Logger is the fundamental interface for all log operations. Log creates a
+// log event from keyvals, a variadic sequence of alternating keys and values.
+// Implementations must be safe for concurrent use by multiple goroutines. In
+// particular, any implementation of Logger that appends to keyvals or
+// modifies or retains any of its elements must make a copy first.
+type Logger interface {
+	Log(keyvals ...interface{}) error
+}
+
+// ErrMissingValue is appended to keyvals slices with odd length to substitute
+// the missing value.
+var ErrMissingValue = errors.New("(MISSING)")
+
+// With returns a new contextual logger with keyvals prepended to those passed
+// to calls to Log. If logger is also a contextual logger created by With or
+// WithPrefix, keyvals is appended to the existing context.
+//
+// The returned Logger replaces all value elements (odd indexes) containing a
+// Valuer with their generated value for each call to its Log method.
+func With(logger Logger, keyvals ...interface{}) Logger {
+	if len(keyvals) == 0 {
+		return logger
+	}
+	l := newContext(logger)
+	kvs := append(l.keyvals, keyvals...)
+	if len(kvs)%2 != 0 {
+		kvs = append(kvs, ErrMissingValue)
+	}
+	return &context{
+		logger: l.logger,
+		// Limiting the capacity of the stored keyvals ensures that a new
+		// backing array is created if the slice must grow in Log or With.
+		// Using the extra capacity without copying risks a data race that
+		// would violate the Logger interface contract.
+		keyvals:   kvs[:len(kvs):len(kvs)],
+		hasValuer: l.hasValuer || containsValuer(keyvals),
+	}
+}
+
+// WithPrefix returns a new contextual logger with keyvals prepended to those
+// passed to calls to Log. If logger is also a contextual logger created by
+// With or WithPrefix, keyvals is prepended to the existing context.
+//
+// The returned Logger replaces all value elements (odd indexes) containing a
+// Valuer with their generated value for each call to its Log method.
+func WithPrefix(logger Logger, keyvals ...interface{}) Logger {
+	if len(keyvals) == 0 {
+		return logger
+	}
+	l := newContext(logger)
+	// Limiting the capacity of the stored keyvals ensures that a new
+	// backing array is created if the slice must grow in Log or With.
+	// Using the extra capacity without copying risks a data race that
+	// would violate the Logger interface contract.
+	n := len(l.keyvals) + len(keyvals)
+	if len(keyvals)%2 != 0 {
+		n++
+	}
+	kvs := make([]interface{}, 0, n)
+	kvs = append(kvs, keyvals...)
+	if len(kvs)%2 != 0 {
+		kvs = append(kvs, ErrMissingValue)
+	}
+	kvs = append(kvs, l.keyvals...)
+	return &context{
+		logger:    l.logger,
+		keyvals:   kvs,
+		hasValuer: l.hasValuer || containsValuer(keyvals),
+	}
+}
+
+// context is the Logger implementation returned by With and WithPrefix. It
+// wraps a Logger and holds keyvals that it includes in all log events. Its
+// Log method calls bindValues to generate values for each Valuer in the
+// context keyvals.
+//
+// A context must always have the same number of stack frames between calls to
+// its Log method and the eventual binding of Valuers to their value. This
+// requirement comes from the functional requirement to allow a context to
+// resolve application call site information for a Caller stored in the
+// context. To do this we must be able to predict the number of logging
+// functions on the stack when bindValues is called.
+//
+// Two implementation details provide the needed stack depth consistency.
+//
+//    1. newContext avoids introducing an additional layer when asked to
+//       wrap another context.
+//    2. With and WithPrefix avoid introducing an additional layer by
+//       returning a newly constructed context with a merged keyvals rather
+//       than simply wrapping the existing context.
+type context struct {
+	logger    Logger
+	keyvals   []interface{}
+	hasValuer bool
+}
+
+func newContext(logger Logger) *context {
+	if c, ok := logger.(*context); ok {
+		return c
+	}
+	return &context{logger: logger}
+}
+
+// Log replaces all value elements (odd indexes) containing a Valuer in the
+// stored context with their generated value, appends keyvals, and passes the
+// result to the wrapped Logger.
+func (l *context) Log(keyvals ...interface{}) error {
+	kvs := append(l.keyvals, keyvals...)
+	if len(kvs)%2 != 0 {
+		kvs = append(kvs, ErrMissingValue)
+	}
+	if l.hasValuer {
+		// If no keyvals were appended above then we must copy l.keyvals so
+		// that future log events will reevaluate the stored Valuers.
+		if len(keyvals) == 0 {
+			kvs = append([]interface{}{}, l.keyvals...)
+		}
+		bindValues(kvs[:len(l.keyvals)])
+	}
+	return l.logger.Log(kvs...)
+}
+
+// LoggerFunc is an adapter to allow use of ordinary functions as Loggers. If
+// f is a function with the appropriate signature, LoggerFunc(f) is a Logger
+// object that calls f.
+type LoggerFunc func(...interface{}) error
+
+// Log implements Logger by calling f(keyvals...).
+func (f LoggerFunc) Log(keyvals ...interface{}) error {
+	return f(keyvals...)
+}
diff --git a/vendor/github.com/go-kit/kit/log/logfmt_logger.go b/vendor/github.com/go-kit/kit/log/logfmt_logger.go
new file mode 100644
index 0000000..a003052
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/logfmt_logger.go
@@ -0,0 +1,62 @@
+package log
+
+import (
+	"bytes"
+	"io"
+	"sync"
+
+	"github.com/go-logfmt/logfmt"
+)
+
+type logfmtEncoder struct {
+	*logfmt.Encoder
+	buf bytes.Buffer
+}
+
+func (l *logfmtEncoder) Reset() {
+	l.Encoder.Reset()
+	l.buf.Reset()
+}
+
+var logfmtEncoderPool = sync.Pool{
+	New: func() interface{} {
+		var enc logfmtEncoder
+		enc.Encoder = logfmt.NewEncoder(&enc.buf)
+		return &enc
+	},
+}
+
+type logfmtLogger struct {
+	w io.Writer
+}
+
+// NewLogfmtLogger returns a logger that encodes keyvals to the Writer in
+// logfmt format. Each log event produces no more than one call to w.Write.
+// The passed Writer must be safe for concurrent use by multiple goroutines if
+// the returned Logger will be used concurrently.
+func NewLogfmtLogger(w io.Writer) Logger {
+	return &logfmtLogger{w}
+}
+
+func (l logfmtLogger) Log(keyvals ...interface{}) error {
+	enc := logfmtEncoderPool.Get().(*logfmtEncoder)
+	enc.Reset()
+	defer logfmtEncoderPool.Put(enc)
+
+	if err := enc.EncodeKeyvals(keyvals...); err != nil {
+		return err
+	}
+
+	// Add newline to the end of the buffer
+	if err := enc.EndRecord(); err != nil {
+		return err
+	}
+
+	// The Logger interface requires implementations to be safe for concurrent
+	// use by multiple goroutines. For this implementation that means making
+	// only one call to l.w.Write() for each call to Log.
+	if _, err := l.w.Write(enc.buf.Bytes()); err != nil {
+		return err
+	}
+	return nil
+}
diff --git a/vendor/github.com/go-kit/kit/log/nop_logger.go b/vendor/github.com/go-kit/kit/log/nop_logger.go
new file mode 100644
index 0000000..1047d62
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/nop_logger.go
@@ -0,0 +1,8 @@
+package log
+
+type nopLogger struct{}
+
+// NewNopLogger returns a logger that doesn't do anything.
+func NewNopLogger() Logger { return nopLogger{} }
+
+func (nopLogger) Log(...interface{}) error { return nil }
diff --git a/vendor/github.com/go-kit/kit/log/stdlib.go b/vendor/github.com/go-kit/kit/log/stdlib.go
new file mode 100644
index 0000000..ff96b5d
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/stdlib.go
@@ -0,0 +1,116 @@
+package log
+
+import (
+	"io"
+	"log"
+	"regexp"
+	"strings"
+)
+
+// StdlibWriter implements io.Writer by invoking the stdlib log.Print. It's
+// designed to be passed to a Go kit logger as the writer, for cases where
+// it's necessary to redirect all Go kit log output to the stdlib logger.
+//
+// If you have any choice in the matter, you shouldn't use this. Prefer to
+// redirect the stdlib log to the Go kit logger via NewStdlibAdapter.
+type StdlibWriter struct{}
+
+// Write implements io.Writer.
+func (w StdlibWriter) Write(p []byte) (int, error) {
+	log.Print(strings.TrimSpace(string(p)))
+	return len(p), nil
+}
+
+// StdlibAdapter wraps a Logger and allows it to be passed to the stdlib
+// logger's SetOutput. It will extract date/timestamps, filenames, and
+// messages, and place them under relevant keys.
+type StdlibAdapter struct {
+	Logger
+	timestampKey string
+	fileKey      string
+	messageKey   string
+}
+
+// StdlibAdapterOption sets a parameter for the StdlibAdapter.
+type StdlibAdapterOption func(*StdlibAdapter)
+
+// TimestampKey sets the key for the timestamp field. By default, it's "ts".
+func TimestampKey(key string) StdlibAdapterOption {
+	return func(a *StdlibAdapter) { a.timestampKey = key }
+}
+
+// FileKey sets the key for the file and line field. By default, it's "caller".
+func FileKey(key string) StdlibAdapterOption {
+	return func(a *StdlibAdapter) { a.fileKey = key }
+}
+
+// MessageKey sets the key for the actual log message. By default, it's "msg".
+func MessageKey(key string) StdlibAdapterOption {
+	return func(a *StdlibAdapter) { a.messageKey = key }
+}
+
+// NewStdlibAdapter returns a new StdlibAdapter wrapper around the passed
+// logger. It's designed to be passed to log.SetOutput.
+func NewStdlibAdapter(logger Logger, options ...StdlibAdapterOption) io.Writer {
+	a := StdlibAdapter{
+		Logger:       logger,
+		timestampKey: "ts",
+		fileKey:      "caller",
+		messageKey:   "msg",
+	}
+	for _, option := range options {
+		option(&a)
+	}
+	return a
+}
+
+func (a StdlibAdapter) Write(p []byte) (int, error) {
+	result := subexps(p)
+	keyvals := []interface{}{}
+	var timestamp string
+	if date, ok := result["date"]; ok && date != "" {
+		timestamp = date
+	}
+	if time, ok := result["time"]; ok && time != "" {
+		if timestamp != "" {
+			timestamp += " "
+		}
+		timestamp += time
+	}
+	if timestamp != "" {
+		keyvals = append(keyvals, a.timestampKey, timestamp)
+	}
+	if file, ok := result["file"]; ok && file != "" {
+		keyvals = append(keyvals, a.fileKey, file)
+	}
+	if msg, ok := result["msg"]; ok {
+		keyvals = append(keyvals, a.messageKey, msg)
+	}
+	if err := a.Logger.Log(keyvals...); err != nil {
+		return 0, err
+	}
+	return len(p), nil
+}
+
+const (
+	logRegexpDate = `(?P<date>[0-9]{4}/[0-9]{2}/[0-9]{2})?[ ]?`
+	logRegexpTime = `(?P<time>[0-9]{2}:[0-9]{2}:[0-9]{2}(\.[0-9]+)?)?[ ]?`
+	logRegexpFile = `(?P<file>.+?:[0-9]+)?`
+	logRegexpMsg  = `(: )?(?P<msg>.*)`
+)
+
+var (
+	logRegexp = regexp.MustCompile(logRegexpDate + logRegexpTime + logRegexpFile + logRegexpMsg)
+)
+
+func subexps(line []byte) map[string]string {
+	m := logRegexp.FindSubmatch(line)
+	if len(m) < len(logRegexp.SubexpNames()) {
+		return map[string]string{}
+	}
+	result := map[string]string{}
+	for i, name := range logRegexp.SubexpNames() {
+		result[name] = string(m[i])
+	}
+	return result
+}
diff --git a/vendor/github.com/go-kit/kit/log/sync.go b/vendor/github.com/go-kit/kit/log/sync.go
new file mode 100644
index 0000000..c07cdfa
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/sync.go
@@ -0,0 +1,116 @@
+package log
+
+import (
+	"io"
+	"sync"
+	"sync/atomic"
+)
+
+// SwapLogger wraps another logger that may be safely replaced while other
+// goroutines use the SwapLogger concurrently. The zero value for a SwapLogger
+// will discard all log events without error.
+//
+// SwapLogger serves well as a package global logger that can be changed by
+// importers.
+type SwapLogger struct {
+	logger atomic.Value
+}
+
+type loggerStruct struct {
+	Logger
+}
+
+// Log implements the Logger interface by forwarding keyvals to the currently
+// wrapped logger. It does not log anything if the wrapped logger is nil.
+func (l *SwapLogger) Log(keyvals ...interface{}) error {
+	s, ok := l.logger.Load().(loggerStruct)
+	if !ok || s.Logger == nil {
+		return nil
+	}
+	return s.Log(keyvals...)
+}
+
+// Swap replaces the currently wrapped logger with logger. Swap may be called
+// concurrently with calls to Log from other goroutines.
+func (l *SwapLogger) Swap(logger Logger) {
+	l.logger.Store(loggerStruct{logger})
+}
+
+// NewSyncWriter returns a new writer that is safe for concurrent use by
+// multiple goroutines. Writes to the returned writer are passed on to w. If
+// another write is already in progress, the calling goroutine blocks until
+// the writer is available.
+//
+// If w implements the following interface, so does the returned writer.
+//
+//    interface {
+//        Fd() uintptr
+//    }
+func NewSyncWriter(w io.Writer) io.Writer {
+	switch w := w.(type) {
+	case fdWriter:
+		return &fdSyncWriter{fdWriter: w}
+	default:
+		return &syncWriter{Writer: w}
+	}
+}
+
+// syncWriter synchronizes concurrent writes to an io.Writer.
+type syncWriter struct {
+	sync.Mutex
+	io.Writer
+}
+
+// Write writes p to the underlying io.Writer. If another write is already in
+// progress, the calling goroutine blocks until the syncWriter is available.
+func (w *syncWriter) Write(p []byte) (n int, err error) {
+	w.Lock()
+	n, err = w.Writer.Write(p)
+	w.Unlock()
+	return n, err
+}
+
+// fdWriter is an io.Writer that also has an Fd method. The most common
+// example of an fdWriter is an *os.File.
+type fdWriter interface {
+	io.Writer
+	Fd() uintptr
+}
+
+// fdSyncWriter synchronizes concurrent writes to an fdWriter.
+type fdSyncWriter struct {
+	sync.Mutex
+	fdWriter
+}
+
+// Write writes p to the underlying io.Writer. If another write is already in
+// progress, the calling goroutine blocks until the fdSyncWriter is available.
+func (w *fdSyncWriter) Write(p []byte) (n int, err error) {
+	w.Lock()
+	n, err = w.fdWriter.Write(p)
+	w.Unlock()
+	return n, err
+}
+
+// syncLogger provides concurrent safe logging for another Logger.
+type syncLogger struct {
+	mu     sync.Mutex
+	logger Logger
+}
+
+// NewSyncLogger returns a logger that synchronizes concurrent use of the
+// wrapped logger. When multiple goroutines use the SyncLogger concurrently
+// only one goroutine will be allowed to log to the wrapped logger at a time.
+// The other goroutines will block until the logger is available.
+func NewSyncLogger(logger Logger) Logger {
+	return &syncLogger{logger: logger}
+}
+
+// Log logs keyvals to the underlying Logger. If another log is already in
+// progress, the calling goroutine blocks until the syncLogger is available.
+func (l *syncLogger) Log(keyvals ...interface{}) error {
+	l.mu.Lock()
+	err := l.logger.Log(keyvals...)
+	l.mu.Unlock()
+	return err
+}
diff --git a/vendor/github.com/go-kit/kit/log/term/LICENSE b/vendor/github.com/go-kit/kit/log/term/LICENSE
new file mode 100644
index 0000000..f090cb4
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/term/LICENSE
@@ -0,0 +1,21 @@
+The MIT License (MIT)
+
+Copyright (c) 2014 Simon Eskildsen
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+THE SOFTWARE.
diff --git a/vendor/github.com/go-kit/kit/log/value.go b/vendor/github.com/go-kit/kit/log/value.go
new file mode 100644
index 0000000..b56f154
--- /dev/null
+++ b/vendor/github.com/go-kit/kit/log/value.go
@@ -0,0 +1,102 @@
+package log
+
+import (
+	"time"
+
+	"github.com/go-stack/stack"
+)
+
+// A Valuer generates a log value. When passed to With or WithPrefix in a
+// value element (odd indexes), it represents a dynamic value which is re-
+// evaluated with each log event.
+type Valuer func() interface{}
+
+// bindValues replaces all value elements (odd indexes) containing a Valuer
+// with their generated value.
+func bindValues(keyvals []interface{}) {
+	for i := 1; i < len(keyvals); i += 2 {
+		if v, ok := keyvals[i].(Valuer); ok {
+			keyvals[i] = v()
+		}
+	}
+}
+
+// containsValuer returns true if any of the value elements (odd indexes)
+// contain a Valuer.
+func containsValuer(keyvals []interface{}) bool {
+	for i := 1; i < len(keyvals); i += 2 {
+		if _, ok := keyvals[i].(Valuer); ok {
+			return true
+		}
+	}
+	return false
+}
+
+// Timestamp returns a timestamp Valuer. It invokes the t function to get the
+// time; unless you are doing something tricky, pass time.Now.
+//
+// Most users will want to use DefaultTimestamp or DefaultTimestampUTC, which
+// are TimestampFormats that use the RFC3339Nano format.
+func Timestamp(t func() time.Time) Valuer {
+	return func() interface{} { return t() }
+}
+
+// TimestampFormat returns a timestamp Valuer with a custom time format. It
+// invokes the t function to get the time to format; unless you are doing
+// something tricky, pass time.Now. The layout string is passed to
+// Time.Format.
+//
+// Most users will want to use DefaultTimestamp or DefaultTimestampUTC, which
+// are TimestampFormats that use the RFC3339Nano format.
+func TimestampFormat(t func() time.Time, layout string) Valuer {
+	return func() interface{} {
+		return timeFormat{
+			time:   t(),
+			layout: layout,
+		}
+	}
+}
+
+// A timeFormat represents an instant in time and a layout used when
+// marshaling to a text format.
+type timeFormat struct {
+	time   time.Time
+	layout string
+}
+
+func (tf timeFormat) String() string {
+	return tf.time.Format(tf.layout)
+}
+
+// MarshalText implements encoding.TextMarshaller.
+func (tf timeFormat) MarshalText() (text []byte, err error) {
+	// The following code adapted from the standard library time.Time.Format
+	// method. Using the same undocumented magic constant to extend the size
+	// of the buffer as seen there.
+	b := make([]byte, 0, len(tf.layout)+10)
+	b = tf.time.AppendFormat(b, tf.layout)
+	return b, nil
+}
+
+// Caller returns a Valuer that returns a file and line from a specified depth
+// in the callstack. Users will probably want to use DefaultCaller.
+func Caller(depth int) Valuer {
+	return func() interface{} { return stack.Caller(depth) }
+}
+
+var (
+	// DefaultTimestamp is a Valuer that returns the current wallclock time,
+	// respecting time zones, when bound.
+	DefaultTimestamp = TimestampFormat(time.Now, time.RFC3339Nano)
+
+	// DefaultTimestampUTC is a Valuer that returns the current time in UTC
+	// when bound.
+	DefaultTimestampUTC = TimestampFormat(
+		func() time.Time { return time.Now().UTC() },
+		time.RFC3339Nano,
+	)
+
+	// DefaultCaller is a Valuer that returns the file and line where the Log
+	// method was invoked. It can only be used with log.With.
+	DefaultCaller = Caller(3)
+)
diff --git a/vendor/github.com/go-logfmt/logfmt/.gitignore b/vendor/github.com/go-logfmt/logfmt/.gitignore
new file mode 100644
index 0000000..320e53e
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/.gitignore
@@ -0,0 +1,4 @@
+_testdata/
+_testdata2/
+logfmt-fuzz.zip
+logfmt.test.exe
diff --git a/vendor/github.com/go-logfmt/logfmt/.travis.yml b/vendor/github.com/go-logfmt/logfmt/.travis.yml
new file mode 100644
index 0000000..b599f65
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/.travis.yml
@@ -0,0 +1,15 @@
+language: go
+sudo: false
+go:
+  - 1.3
+  - 1.4
+  - 1.5
+  - 1.6
+  - tip
+
+before_install:
+  - go get github.com/mattn/goveralls
+  - go get golang.org/x/tools/cmd/cover
+
+script:
+  - goveralls -service=travis-ci
diff --git a/vendor/github.com/go-logfmt/logfmt/LICENSE b/vendor/github.com/go-logfmt/logfmt/LICENSE
new file mode 100644
index 0000000..c026508
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/LICENSE
@@ -0,0 +1,22 @@
+The MIT License (MIT)
+
+Copyright (c) 2015 go-logfmt
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
diff --git a/vendor/github.com/go-logfmt/logfmt/README.md b/vendor/github.com/go-logfmt/logfmt/README.md
new file mode 100644
index 0000000..3a8f10b
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/README.md
@@ -0,0 +1,33 @@
+[![GoDoc](https://godoc.org/github.com/go-logfmt/logfmt?status.svg)](https://godoc.org/github.com/go-logfmt/logfmt)
+[![Go Report Card](https://goreportcard.com/badge/go-logfmt/logfmt)](https://goreportcard.com/report/go-logfmt/logfmt)
+[![TravisCI](https://travis-ci.org/go-logfmt/logfmt.svg?branch=master)](https://travis-ci.org/go-logfmt/logfmt)
+[![Coverage Status](https://coveralls.io/repos/github/go-logfmt/logfmt/badge.svg?branch=master)](https://coveralls.io/github/go-logfmt/logfmt?branch=master)
+
+# logfmt
+
+Package logfmt implements utilities to marshal and unmarshal data in the [logfmt
+format](https://brandur.org/logfmt). It provides an API similar to
+[encoding/json](http://golang.org/pkg/encoding/json/) and
+[encoding/xml](http://golang.org/pkg/encoding/xml/).
+
+The logfmt format was first documented by Brandur Leach in [this
+article](https://brandur.org/logfmt). The format has not been formally
+standardized. The most authoritative public specification to date has been the
+documentation of a Go Language [package](http://godoc.org/github.com/kr/logfmt)
+written by Blake Mizerany and Keith Rarick.
+
+## Goals
+
+This project attempts to conform as closely as possible to the prior art, while
+also removing ambiguity where necessary to provide well behaved encoder and
+decoder implementations.
+
+## Non-goals
+
+This project does not attempt to formally standardize the logfmt format. In the
+event that logfmt is standardized this project would take conforming to the
+standard as a goal.
+
+## Versioning
+
+Package logfmt publishes releases via [semver](http://semver.org/) compatible Git tags prefixed with a single 'v'.
diff --git a/vendor/github.com/go-logfmt/logfmt/decode.go b/vendor/github.com/go-logfmt/logfmt/decode.go
new file mode 100644
index 0000000..04e0eff
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/decode.go
@@ -0,0 +1,237 @@
+package logfmt
+
+import (
+	"bufio"
+	"bytes"
+	"fmt"
+	"io"
+	"unicode/utf8"
+)
+
+// A Decoder reads and decodes logfmt records from an input stream.
+type Decoder struct {
+	pos     int
+	key     []byte
+	value   []byte
+	lineNum int
+	s       *bufio.Scanner
+	err     error
+}
+
+// NewDecoder returns a new decoder that reads from r.
+//
+// The decoder introduces its own buffering and may read data from r beyond
+// the logfmt records requested.
+func NewDecoder(r io.Reader) *Decoder {
+	dec := &Decoder{
+		s: bufio.NewScanner(r),
+	}
+	return dec
+}
+
+// ScanRecord advances the Decoder to the next record, which can then be
+// parsed with the ScanKeyval method. It returns false when decoding stops,
+// either by reaching the end of the input or an error. After ScanRecord
+// returns false, the Err method will return any error that occurred during
+// decoding, except that if it was io.EOF, Err will return nil.
+func (dec *Decoder) ScanRecord() bool {
+	if dec.err != nil {
+		return false
+	}
+	if !dec.s.Scan() {
+		dec.err = dec.s.Err()
+		return false
+	}
+	dec.lineNum++
+	dec.pos = 0
+	return true
+}
+
+// ScanKeyval advances the Decoder to the next key/value pair of the current
+// record, which can then be retrieved with the Key and Value methods. It
+// returns false when decoding stops, either by reaching the end of the
+// current record or an error.
+func (dec *Decoder) ScanKeyval() bool {
+	dec.key, dec.value = nil, nil
+	if dec.err != nil {
+		return false
+	}
+
+	line := dec.s.Bytes()
+
+	// garbage
+	for p, c := range line[dec.pos:] {
+		if c > ' ' {
+			dec.pos += p
+			goto key
+		}
+	}
+	dec.pos = len(line)
+	return false
+
+key:
+	const invalidKeyError = "invalid key"
+
+	start, multibyte := dec.pos, false
+	for p, c := range line[dec.pos:] {
+		switch {
+		case c == '=':
+			dec.pos += p
+			if dec.pos > start {
+				dec.key = line[start:dec.pos]
+				if multibyte && bytes.IndexRune(dec.key, utf8.RuneError) != -1 {
+					dec.syntaxError(invalidKeyError)
+					return false
+				}
+			}
+			if dec.key == nil {
+				dec.unexpectedByte(c)
+				return false
+			}
+			goto equal
+		case c == '"':
+			dec.pos += p
+			dec.unexpectedByte(c)
+			return false
+		case c <= ' ':
+			dec.pos += p
+			if dec.pos > start {
+				dec.key = line[start:dec.pos]
+				if multibyte && bytes.IndexRune(dec.key, utf8.RuneError) != -1 {
+					dec.syntaxError(invalidKeyError)
+					return false
+				}
+			}
+			return true
+		case c >= utf8.RuneSelf:
+			multibyte = true
+		}
+	}
+	dec.pos = len(line)
+	if dec.pos > start {
+		dec.key = line[start:dec.pos]
+		if multibyte && bytes.IndexRune(dec.key, utf8.RuneError) != -1 {
+			dec.syntaxError(invalidKeyError)
+			return false
+		}
+	}
+	return true
+
+equal:
+	dec.pos++
+	if dec.pos >= len(line) {
+		return true
+	}
+	switch c := line[dec.pos]; {
+	case c <= ' ':
+		return true
+	case c == '"':
+		goto qvalue
+	}
+
+	// value
+	start = dec.pos
+	for p, c := range line[dec.pos:] {
+		switch {
+		case c == '=' || c == '"':
+			dec.pos += p
+			dec.unexpectedByte(c)
+			return false
+		case c <= ' ':
+			dec.pos += p
+			if dec.pos > start {
+				dec.value = line[start:dec.pos]
+			}
+			return true
+		}
+	}
+	dec.pos = len(line)
+	if dec.pos > start {
+		dec.value = line[start:dec.pos]
+	}
+	return true
+
+qvalue:
+	const (
+		untermQuote  = "unterminated quoted value"
+		invalidQuote = "invalid quoted value"
+	)
+
+	hasEsc, esc := false, false
+	start = dec.pos
+	for p, c := range line[dec.pos+1:] {
+		switch {
+		case esc:
+			esc = false
+		case c == '\\':
+			hasEsc, esc = true, true
+		case c == '"':
+			dec.pos += p + 2
+			if hasEsc {
+				v, ok := unquoteBytes(line[start:dec.pos])
+				if !ok {
+					dec.syntaxError(invalidQuote)
+					return false
+				}
+				dec.value = v
+			} else {
+				start++
+				end := dec.pos - 1
+				if end > start {
+					dec.value = line[start:end]
+				}
+			}
+			return true
+		}
+	}
+	dec.pos = len(line)
+	dec.syntaxError(untermQuote)
+	return false
+}
+
+// Key returns the most recent key found by a call to ScanKeyval. The returned
+// slice may point to internal buffers and is only valid until the next call
+// to ScanRecord.  It does no allocation.
+func (dec *Decoder) Key() []byte {
+	return dec.key
+}
+
+// Value returns the most recent value found by a call to ScanKeyval. The
+// returned slice may point to internal buffers and is only valid until the
+// next call to ScanRecord.  It does no allocation when the value has no
+// escape sequences.
+func (dec *Decoder) Value() []byte {
+	return dec.value
+}
+
+// Err returns the first non-EOF error that was encountered by the Scanner.
+func (dec *Decoder) Err() error {
+	return dec.err
+}
+
+func (dec *Decoder) syntaxError(msg string) {
+	dec.err = &SyntaxError{
+		Msg:  msg,
+		Line: dec.lineNum,
+		Pos:  dec.pos + 1,
+	}
+}
+
+func (dec *Decoder) unexpectedByte(c byte) {
+	dec.err = &SyntaxError{
+		Msg:  fmt.Sprintf("unexpected %q", c),
+		Line: dec.lineNum,
+		Pos:  dec.pos + 1,
+	}
+}
+
+// A SyntaxError represents a syntax error in the logfmt input stream.
+type SyntaxError struct {
+	Msg  string
+	Line int
+	Pos  int
+}
+
+func (e *SyntaxError) Error() string {
+	return fmt.Sprintf("logfmt syntax error at pos %d on line %d: %s", e.Pos, e.Line, e.Msg)
+}
diff --git a/vendor/github.com/go-logfmt/logfmt/doc.go b/vendor/github.com/go-logfmt/logfmt/doc.go
new file mode 100644
index 0000000..378e9ad
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/doc.go
@@ -0,0 +1,6 @@
+// Package logfmt implements utilities to marshal and unmarshal data in the
+// logfmt format. The logfmt format records key/value pairs in a way that
+// balances readability for humans and simplicity of computer parsing. It is
+// most commonly used as a more human friendly alternative to JSON for
+// structured logging.
+package logfmt
diff --git a/vendor/github.com/go-logfmt/logfmt/encode.go b/vendor/github.com/go-logfmt/logfmt/encode.go
new file mode 100644
index 0000000..55f1603
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/encode.go
@@ -0,0 +1,321 @@
+package logfmt
+
+import (
+	"bytes"
+	"encoding"
+	"errors"
+	"fmt"
+	"io"
+	"reflect"
+	"strings"
+	"unicode/utf8"
+)
+
+// MarshalKeyvals returns the logfmt encoding of keyvals, a variadic sequence
+// of alternating keys and values.
+func MarshalKeyvals(keyvals ...interface{}) ([]byte, error) {
+	buf := &bytes.Buffer{}
+	if err := NewEncoder(buf).EncodeKeyvals(keyvals...); err != nil {
+		return nil, err
+	}
+	return buf.Bytes(), nil
+}
+
+// An Encoder writes logfmt data to an output stream.
+type Encoder struct {
+	w       io.Writer
+	scratch bytes.Buffer
+	needSep bool
+}
+
+// NewEncoder returns a new encoder that writes to w.
+func NewEncoder(w io.Writer) *Encoder {
+	return &Encoder{
+		w: w,
+	}
+}
+
+var (
+	space   = []byte(" ")
+	equals  = []byte("=")
+	newline = []byte("\n")
+	null    = []byte("null")
+)
+
+// EncodeKeyval writes the logfmt encoding of key and value to the stream. A
+// single space is written before the second and subsequent keys in a record.
+// Nothing is written if a non-nil error is returned.
+func (enc *Encoder) EncodeKeyval(key, value interface{}) error {
+	enc.scratch.Reset()
+	if enc.needSep {
+		if _, err := enc.scratch.Write(space); err != nil {
+			return err
+		}
+	}
+	if err := writeKey(&enc.scratch, key); err != nil {
+		return err
+	}
+	if _, err := enc.scratch.Write(equals); err != nil {
+		return err
+	}
+	if err := writeValue(&enc.scratch, value); err != nil {
+		return err
+	}
+	_, err := enc.w.Write(enc.scratch.Bytes())
+	enc.needSep = true
+	return err
+}
+
+// EncodeKeyvals writes the logfmt encoding of keyvals to the stream. Keyvals
+// is a variadic sequence of alternating keys and values. Keys of unsupported
+// type are skipped along with their corresponding value. Values of
+// unsupported type or that cause a MarshalerError are replaced by their error
+// but do not cause EncodeKeyvals to return an error. If a non-nil error is
+// returned some key/value pairs may not have be written.
+func (enc *Encoder) EncodeKeyvals(keyvals ...interface{}) error {
+	if len(keyvals) == 0 {
+		return nil
+	}
+	if len(keyvals)%2 == 1 {
+		keyvals = append(keyvals, nil)
+	}
+	for i := 0; i < len(keyvals); i += 2 {
+		k, v := keyvals[i], keyvals[i+1]
+		err := enc.EncodeKeyval(k, v)
+		if err == ErrUnsupportedKeyType {
+			continue
+		}
+		if _, ok := err.(*MarshalerError); ok || err == ErrUnsupportedValueType {
+			v = err
+			err = enc.EncodeKeyval(k, v)
+		}
+		if err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+// MarshalerError represents an error encountered while marshaling a value.
+type MarshalerError struct {
+	Type reflect.Type
+	Err  error
+}
+
+func (e *MarshalerError) Error() string {
+	return "error marshaling value of type " + e.Type.String() + ": " + e.Err.Error()
+}
+
+// ErrNilKey is returned by Marshal functions and Encoder methods if a key is
+// a nil interface or pointer value.
+var ErrNilKey = errors.New("nil key")
+
+// ErrInvalidKey is returned by Marshal functions and Encoder methods if a key
+// contains an invalid character.
+var ErrInvalidKey = errors.New("invalid key")
+
+// ErrUnsupportedKeyType is returned by Encoder methods if a key has an
+// unsupported type.
+var ErrUnsupportedKeyType = errors.New("unsupported key type")
+
+// ErrUnsupportedValueType is returned by Encoder methods if a value has an
+// unsupported type.
+var ErrUnsupportedValueType = errors.New("unsupported value type")
+
+func writeKey(w io.Writer, key interface{}) error {
+	if key == nil {
+		return ErrNilKey
+	}
+
+	switch k := key.(type) {
+	case string:
+		return writeStringKey(w, k)
+	case []byte:
+		if k == nil {
+			return ErrNilKey
+		}
+		return writeBytesKey(w, k)
+	case encoding.TextMarshaler:
+		kb, err := safeMarshal(k)
+		if err != nil {
+			return err
+		}
+		if kb == nil {
+			return ErrNilKey
+		}
+		return writeBytesKey(w, kb)
+	case fmt.Stringer:
+		ks, ok := safeString(k)
+		if !ok {
+			return ErrNilKey
+		}
+		return writeStringKey(w, ks)
+	default:
+		rkey := reflect.ValueOf(key)
+		switch rkey.Kind() {
+		case reflect.Array, reflect.Chan, reflect.Func, reflect.Map, reflect.Slice, reflect.Struct:
+			return ErrUnsupportedKeyType
+		case reflect.Ptr:
+			if rkey.IsNil() {
+				return ErrNilKey
+			}
+			return writeKey(w, rkey.Elem().Interface())
+		}
+		return writeStringKey(w, fmt.Sprint(k))
+	}
+}
+
+func invalidKeyRune(r rune) bool {
+	return r <= ' ' || r == '=' || r == '"' || r == utf8.RuneError
+}
+
+func invalidKeyString(key string) bool {
+	return len(key) == 0 || strings.IndexFunc(key, invalidKeyRune) != -1
+}
+
+func invalidKey(key []byte) bool {
+	return len(key) == 0 || bytes.IndexFunc(key, invalidKeyRune) != -1
+}
+
+func writeStringKey(w io.Writer, key string) error {
+	if invalidKeyString(key) {
+		return ErrInvalidKey
+	}
+	_, err := io.WriteString(w, key)
+	return err
+}
+
+func writeBytesKey(w io.Writer, key []byte) error {
+	if invalidKey(key) {
+		return ErrInvalidKey
+	}
+	_, err := w.Write(key)
+	return err
+}
+
+func writeValue(w io.Writer, value interface{}) error {
+	switch v := value.(type) {
+	case nil:
+		return writeBytesValue(w, null)
+	case string:
+		return writeStringValue(w, v, true)
+	case []byte:
+		return writeBytesValue(w, v)
+	case encoding.TextMarshaler:
+		vb, err := safeMarshal(v)
+		if err != nil {
+			return err
+		}
+		if vb == nil {
+			vb = null
+		}
+		return writeBytesValue(w, vb)
+	case error:
+		se, ok := safeError(v)
+		return writeStringValue(w, se, ok)
+	case fmt.Stringer:
+		ss, ok := safeString(v)
+		return writeStringValue(w, ss, ok)
+	default:
+		rvalue := reflect.ValueOf(value)
+		switch rvalue.Kind() {
+		case reflect.Array, reflect.Chan, reflect.Func, reflect.Map, reflect.Slice, reflect.Struct:
+			return ErrUnsupportedValueType
+		case reflect.Ptr:
+			if rvalue.IsNil() {
+				return writeBytesValue(w, null)
+			}
+			return writeValue(w, rvalue.Elem().Interface())
+		}
+		return writeStringValue(w, fmt.Sprint(v), true)
+	}
+}
+
+func needsQuotedValueRune(r rune) bool {
+	return r <= ' ' || r == '=' || r == '"' || r == utf8.RuneError
+}
+
+func writeStringValue(w io.Writer, value string, ok bool) error {
+	var err error
+	if ok && value == "null" {
+		_, err = io.WriteString(w, `"null"`)
+	} else if strings.IndexFunc(value, needsQuotedValueRune) != -1 {
+		_, err = writeQuotedString(w, value)
+	} else {
+		_, err = io.WriteString(w, value)
+	}
+	return err
+}
+
+func writeBytesValue(w io.Writer, value []byte) error {
+	var err error
+	if bytes.IndexFunc(value, needsQuotedValueRune) != -1 {
+		_, err = writeQuotedBytes(w, value)
+	} else {
+		_, err = w.Write(value)
+	}
+	return err
+}
+
+// EndRecord writes a newline character to the stream and resets the encoder
+// to the beginning of a new record.
+func (enc *Encoder) EndRecord() error {
+	_, err := enc.w.Write(newline)
+	if err == nil {
+		enc.needSep = false
+	}
+	return err
+}
+
+// Reset resets the encoder to the beginning of a new record.
+func (enc *Encoder) Reset() {
+	enc.needSep = false
+}
+
+func safeError(err error) (s string, ok bool) {
+	defer func() {
+		if panicVal := recover(); panicVal != nil {
+			if v := reflect.ValueOf(err); v.Kind() == reflect.Ptr && v.IsNil() {
+				s, ok = "null", false
+			} else {
+				panic(panicVal)
+			}
+		}
+	}()
+	s, ok = err.Error(), true
+	return
+}
+
+func safeString(str fmt.Stringer) (s string, ok bool) {
+	defer func() {
+		if panicVal := recover(); panicVal != nil {
+			if v := reflect.ValueOf(str); v.Kind() == reflect.Ptr && v.IsNil() {
+				s, ok = "null", false
+			} else {
+				panic(panicVal)
+			}
+		}
+	}()
+	s, ok = str.String(), true
+	return
+}
+
+func safeMarshal(tm encoding.TextMarshaler) (b []byte, err error) {
+	defer func() {
+		if panicVal := recover(); panicVal != nil {
+			if v := reflect.ValueOf(tm); v.Kind() == reflect.Ptr && v.IsNil() {
+				b, err = nil, nil
+			} else {
+				panic(panicVal)
+			}
+		}
+	}()
+	b, err = tm.MarshalText()
+	if err != nil {
+		return nil, &MarshalerError{
+			Type: reflect.TypeOf(tm),
+			Err:  err,
+		}
+	}
+	return
+}
diff --git a/vendor/github.com/go-logfmt/logfmt/fuzz.go b/vendor/github.com/go-logfmt/logfmt/fuzz.go
new file mode 100644
index 0000000..6553b35
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/fuzz.go
@@ -0,0 +1,126 @@
+// +build gofuzz
+
+package logfmt
+
+import (
+	"bufio"
+	"bytes"
+	"fmt"
+	"io"
+	"reflect"
+
+	kr "github.com/kr/logfmt"
+)
+
+// Fuzz checks reserialized data matches
+func Fuzz(data []byte) int {
+	parsed, err := parse(data)
+	if err != nil {
+		return 0
+	}
+	var w1 bytes.Buffer
+	if err = write(parsed, &w1); err != nil {
+		panic(err)
+	}
+	parsed, err = parse(w1.Bytes())
+	if err != nil {
+		panic(err)
+	}
+	var w2 bytes.Buffer
+	if err = write(parsed, &w2); err != nil {
+		panic(err)
+	}
+	if !bytes.Equal(w1.Bytes(), w2.Bytes()) {
+		panic(fmt.Sprintf("reserialized data does not match:\n%q\n%q\n", w1.Bytes(), w2.Bytes()))
+	}
+	return 1
+}
+
+// FuzzVsKR checks go-logfmt/logfmt against kr/logfmt
+func FuzzVsKR(data []byte) int {
+	parsed, err := parse(data)
+	parsedKR, errKR := parseKR(data)
+
+	// github.com/go-logfmt/logfmt is a stricter parser. It returns errors for
+	// more inputs than github.com/kr/logfmt. Ignore any inputs that have a
+	// stict error.
+	if err != nil {
+		return 0
+	}
+
+	// Fail if the more forgiving parser finds an error not found by the
+	// stricter parser.
+	if errKR != nil {
+		panic(fmt.Sprintf("unmatched error: %v", errKR))
+	}
+
+	if !reflect.DeepEqual(parsed, parsedKR) {
+		panic(fmt.Sprintf("parsers disagree:\n%+v\n%+v\n", parsed, parsedKR))
+	}
+	return 1
+}
+
+type kv struct {
+	k, v []byte
+}
+
+func parse(data []byte) ([][]kv, error) {
+	var got [][]kv
+	dec := NewDecoder(bytes.NewReader(data))
+	for dec.ScanRecord() {
+		var kvs []kv
+		for dec.ScanKeyval() {
+			kvs = append(kvs, kv{dec.Key(), dec.Value()})
+		}
+		got = append(got, kvs)
+	}
+	return got, dec.Err()
+}
+
+func parseKR(data []byte) ([][]kv, error) {
+	var (
+		s   = bufio.NewScanner(bytes.NewReader(data))
+		err error
+		h   saveHandler
+		got [][]kv
+	)
+	for err == nil && s.Scan() {
+		h.kvs = nil
+		err = kr.Unmarshal(s.Bytes(), &h)
+		got = append(got, h.kvs)
+	}
+	if err == nil {
+		err = s.Err()
+	}
+	return got, err
+}
+
+type saveHandler struct {
+	kvs []kv
+}
+
+func (h *saveHandler) HandleLogfmt(key, val []byte) error {
+	if len(key) == 0 {
+		key = nil
+	}
+	if len(val) == 0 {
+		val = nil
+	}
+	h.kvs = append(h.kvs, kv{key, val})
+	return nil
+}
+
+func write(recs [][]kv, w io.Writer) error {
+	enc := NewEncoder(w)
+	for _, rec := range recs {
+		for _, f := range rec {
+			if err := enc.EncodeKeyval(f.k, f.v); err != nil {
+				return err
+			}
+		}
+		if err := enc.EndRecord(); err != nil {
+			return err
+		}
+	}
+	return nil
+}
diff --git a/vendor/github.com/go-logfmt/logfmt/jsonstring.go b/vendor/github.com/go-logfmt/logfmt/jsonstring.go
new file mode 100644
index 0000000..030ac85
--- /dev/null
+++ b/vendor/github.com/go-logfmt/logfmt/jsonstring.go
@@ -0,0 +1,277 @@
+package logfmt
+
+import (
+	"bytes"
+	"io"
+	"strconv"
+	"sync"
+	"unicode"
+	"unicode/utf16"
+	"unicode/utf8"
+)
+
+// Taken from Go's encoding/json and modified for use here.
+
+// Copyright 2010 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+var hex = "0123456789abcdef"
+
+var bufferPool = sync.Pool{
+	New: func() interface{} {
+		return &bytes.Buffer{}
+	},
+}
+
+func getBuffer() *bytes.Buffer {
+	return bufferPool.Get().(*bytes.Buffer)
+}
+
+func poolBuffer(buf *bytes.Buffer) {
+	buf.Reset()
+	bufferPool.Put(buf)
+}
+
+// NOTE: keep in sync with writeQuotedBytes below.
+func writeQuotedString(w io.Writer, s string) (int, error) {
+	buf := getBuffer()
+	buf.WriteByte('"')
+	start := 0
+	for i := 0; i < len(s); {
+		if b := s[i]; b < utf8.RuneSelf {
+			if 0x20 <= b && b != '\\' && b != '"' {
+				i++
+				continue
+			}
+			if start < i {
+				buf.WriteString(s[start:i])
+			}
+			switch b {
+			case '\\', '"':
+				buf.WriteByte('\\')
+				buf.WriteByte(b)
+			case '\n':
+				buf.WriteByte('\\')
+				buf.WriteByte('n')
+			case '\r':
+				buf.WriteByte('\\')
+				buf.WriteByte('r')
+			case '\t':
+				buf.WriteByte('\\')
+				buf.WriteByte('t')
+			default:
+				// This encodes bytes < 0x20 except for \n, \r, and \t.
+				buf.WriteString(`\u00`)
+				buf.WriteByte(hex[b>>4])
+				buf.WriteByte(hex[b&0xF])
+			}
+			i++
+			start = i
+			continue
+		}
+		c, size := utf8.DecodeRuneInString(s[i:])
+		if c == utf8.RuneError {
+			if start < i {
+				buf.WriteString(s[start:i])
+			}
+			buf.WriteString(`\ufffd`)
+			i += size
+			start = i
+			continue
+		}
+		i += size
+	}
+	if start < len(s) {
+		buf.WriteString(s[start:])
+	}
+	buf.WriteByte('"')
+	n, err := w.Write(buf.Bytes())
+	poolBuffer(buf)
+	return n, err
+}
+
+// NOTE: keep in sync with writeQuoteString above.
+func writeQuotedBytes(w io.Writer, s []byte) (int, error) {
+	buf := getBuffer()
+	buf.WriteByte('"')
+	start := 0
+	for i := 0; i < len(s); {
+		if b := s[i]; b < utf8.RuneSelf {
+			if 0x20 <= b && b != '\\' && b != '"' {
+				i++
+				continue
+			}
+			if start < i {
+				buf.Write(s[start:i])
+			}
+			switch b {
+			case '\\', '"':
+				buf.WriteByte('\\')
+				buf.WriteByte(b)
+			case '\n':
+				buf.WriteByte('\\')
+				buf.WriteByte('n')
+			case '\r':
+				buf.WriteByte('\\')
+				buf.WriteByte('r')
+			case '\t':
+				buf.WriteByte('\\')
+				buf.WriteByte('t')
+			default:
+				// This encodes bytes < 0x20 except for \n, \r, and \t.
+				buf.WriteString(`\u00`)
+				buf.WriteByte(hex[b>>4])
+				buf.WriteByte(hex[b&0xF])
+			}
+			i++
+			start = i
+			continue
+		}
+		c, size := utf8.DecodeRune(s[i:])
+		if c == utf8.RuneError {
+			if start < i {
+				buf.Write(s[start:i])
+			}
+			buf.WriteString(`\ufffd`)
+			i += size
+			start = i
+			continue
+		}
+		i += size
+	}
+	if start < len(s) {
+		buf.Write(s[start:])
+	}
+	buf.WriteByte('"')
+	n, err := w.Write(buf.Bytes())
+	poolBuffer(buf)
+	return n, err
+}
+
+// getu4 decodes \uXXXX from the beginning of s, returning the hex value,
+// or it returns -1.
+func getu4(s []byte) rune {
+	if len(s) < 6 || s[0] != '\\' || s[1] != 'u' {
+		return -1
+	}
+	r, err := strconv.ParseUint(string(s[2:6]), 16, 64)
+	if err != nil {
+		return -1
+	}
+	return rune(r)
+}
+
+func unquoteBytes(s []byte) (t []byte, ok bool) {
+	if len(s) < 2 || s[0] != '"' || s[len(s)-1] != '"' {
+		return
+	}
+	s = s[1 : len(s)-1]
+
+	// Check for unusual characters. If there are none,
+	// then no unquoting is needed, so return a slice of the
+	// original bytes.
+	r := 0
+	for r < len(s) {
+		c := s[r]
+		if c == '\\' || c == '"' || c < ' ' {
+			break
+		}
+		if c < utf8.RuneSelf {
+			r++
+			continue
+		}
+		rr, size := utf8.DecodeRune(s[r:])
+		if rr == utf8.RuneError {
+			break
+		}
+		r += size
+	}
+	if r == len(s) {
+		return s, true
+	}
+
+	b := make([]byte, len(s)+2*utf8.UTFMax)
+	w := copy(b, s[0:r])
+	for r < len(s) {
+		// Out of room?  Can only happen if s is full of
+		// malformed UTF-8 and we're replacing each
+		// byte with RuneError.
+		if w >= len(b)-2*utf8.UTFMax {
+			nb := make([]byte, (len(b)+utf8.UTFMax)*2)
+			copy(nb, b[0:w])
+			b = nb
+		}
+		switch c := s[r]; {
+		case c == '\\':
+			r++
+			if r >= len(s) {
+				return
+			}
+			switch s[r] {
+			default:
+				return
+			case '"', '\\', '/', '\'':
+				b[w] = s[r]
+				r++
+				w++
+			case 'b':
+				b[w] = '\b'
+				r++
+				w++
+			case 'f':
+				b[w] = '\f'
+				r++
+				w++
+			case 'n':
+				b[w] = '\n'
+				r++
+				w++
+			case 'r':
+				b[w] = '\r'
+				r++
+				w++
+			case 't':
+				b[w] = '\t'
+				r++
+				w++
+			case 'u':
+				r--
+				rr := getu4(s[r:])
+				if rr < 0 {
+					return
+				}
+				r += 6
+				if utf16.IsSurrogate(rr) {
+					rr1 := getu4(s[r:])
+					if dec := utf16.DecodeRune(rr, rr1); dec != unicode.ReplacementChar {
+						// A valid pair; consume.
+						r += 6
+						w += utf8.EncodeRune(b[w:], dec)
+						break
+					}
+					// Invalid surrogate; fall back to replacement rune.
+					rr = unicode.ReplacementChar
+				}
+				w += utf8.EncodeRune(b[w:], rr)
+			}
+
+		// Quote, control characters are invalid.
+		case c == '"', c < ' ':
+			return
+
+		// ASCII
+		case c < utf8.RuneSelf:
+			b[w] = c
+			r++
+			w++
+
+		// Coerce to well-formed UTF-8.
+		default:
+			rr, size := utf8.DecodeRune(s[r:])
+			r += size
+			w += utf8.EncodeRune(b[w:], rr)
+		}
+	}
+	return b[0:w], true
+}
diff --git a/vendor/github.com/go-stack/stack/.travis.yml b/vendor/github.com/go-stack/stack/.travis.yml
new file mode 100644
index 0000000..e91cc9a
--- /dev/null
+++ b/vendor/github.com/go-stack/stack/.travis.yml
@@ -0,0 +1,13 @@
+language: go
+sudo: false
+go:
+  - 1.7
+  - 1.8
+  - 1.9
+  - tip
+
+before_install:
+  - go get github.com/mattn/goveralls
+
+script:
+  - goveralls -service=travis-ci
diff --git a/vendor/github.com/go-stack/stack/LICENSE.md b/vendor/github.com/go-stack/stack/LICENSE.md
new file mode 100644
index 0000000..2abf98e
--- /dev/null
+++ b/vendor/github.com/go-stack/stack/LICENSE.md
@@ -0,0 +1,21 @@
+The MIT License (MIT)
+
+Copyright (c) 2014 Chris Hines
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/vendor/github.com/go-stack/stack/README.md b/vendor/github.com/go-stack/stack/README.md
new file mode 100644
index 0000000..f11cccc
--- /dev/null
+++ b/vendor/github.com/go-stack/stack/README.md
@@ -0,0 +1,38 @@
+[![GoDoc](https://godoc.org/github.com/go-stack/stack?status.svg)](https://godoc.org/github.com/go-stack/stack)
+[![Go Report Card](https://goreportcard.com/badge/go-stack/stack)](https://goreportcard.com/report/go-stack/stack)
+[![TravisCI](https://travis-ci.org/go-stack/stack.svg?branch=master)](https://travis-ci.org/go-stack/stack)
+[![Coverage Status](https://coveralls.io/repos/github/go-stack/stack/badge.svg?branch=master)](https://coveralls.io/github/go-stack/stack?branch=master)
+
+# stack
+
+Package stack implements utilities to capture, manipulate, and format call
+stacks. It provides a simpler API than package runtime.
+
+The implementation takes care of the minutia and special cases of interpreting
+the program counter (pc) values returned by runtime.Callers.
+
+## Versioning
+
+Package stack publishes releases via [semver](http://semver.org/) compatible Git
+tags prefixed with a single 'v'. The master branch always contains the latest
+release. The develop branch contains unreleased commits.
+
+## Formatting
+
+Package stack's types implement fmt.Formatter, which provides a simple and
+flexible way to declaratively configure formatting when used with logging or
+error tracking packages.
+
+```go
+func DoTheThing() {
+    c := stack.Caller(0)
+    log.Print(c)          // "source.go:10"
+    log.Printf("%+v", c)  // "pkg/path/source.go:10"
+    log.Printf("%n", c)   // "DoTheThing"
+
+    s := stack.Trace().TrimRuntime()
+    log.Print(s)          // "[source.go:15 caller.go:42 main.go:14]"
+}
+```
+
+See the docs for all of the supported formatting options.
diff --git a/vendor/github.com/go-stack/stack/stack.go b/vendor/github.com/go-stack/stack/stack.go
new file mode 100644
index 0000000..1bc6971
--- /dev/null
+++ b/vendor/github.com/go-stack/stack/stack.go
@@ -0,0 +1,324 @@
+// +build go1.7
+
+// Package stack implements utilities to capture, manipulate, and format call
+// stacks. It provides a simpler API than package runtime.
+//
+// The implementation takes care of the minutia and special cases of
+// interpreting the program counter (pc) values returned by runtime.Callers.
+//
+// Package stack's types implement fmt.Formatter, which provides a simple and
+// flexible way to declaratively configure formatting when used with logging
+// or error tracking packages.
+package stack
+
+import (
+	"bytes"
+	"errors"
+	"fmt"
+	"io"
+	"runtime"
+	"strconv"
+	"strings"
+)
+
+// Call records a single function invocation from a goroutine stack.
+type Call struct {
+	frame runtime.Frame
+}
+
+// Caller returns a Call from the stack of the current goroutine. The argument
+// skip is the number of stack frames to ascend, with 0 identifying the
+// calling function.
+func Caller(skip int) Call {
+	// As of Go 1.9 we need room for up to three PC entries.
+	//
+	// 0. An entry for the stack frame prior to the target to check for
+	//    special handling needed if that prior entry is runtime.sigpanic.
+	// 1. A possible second entry to hold metadata about skipped inlined
+	//    functions. If inline functions were not skipped the target frame
+	//    PC will be here.
+	// 2. A third entry for the target frame PC when the second entry
+	//    is used for skipped inline functions.
+	var pcs [3]uintptr
+	n := runtime.Callers(skip+1, pcs[:])
+	frames := runtime.CallersFrames(pcs[:n])
+	frame, _ := frames.Next()
+	frame, _ = frames.Next()
+
+	return Call{
+		frame: frame,
+	}
+}
+
+// String implements fmt.Stinger. It is equivalent to fmt.Sprintf("%v", c).
+func (c Call) String() string {
+	return fmt.Sprint(c)
+}
+
+// MarshalText implements encoding.TextMarshaler. It formats the Call the same
+// as fmt.Sprintf("%v", c).
+func (c Call) MarshalText() ([]byte, error) {
+	if c.frame == (runtime.Frame{}) {
+		return nil, ErrNoFunc
+	}
+
+	buf := bytes.Buffer{}
+	fmt.Fprint(&buf, c)
+	return buf.Bytes(), nil
+}
+
+// ErrNoFunc means that the Call has a nil *runtime.Func. The most likely
+// cause is a Call with the zero value.
+var ErrNoFunc = errors.New("no call stack information")
+
+// Format implements fmt.Formatter with support for the following verbs.
+//
+//    %s    source file
+//    %d    line number
+//    %n    function name
+//    %k    last segment of the package path
+//    %v    equivalent to %s:%d
+//
+// It accepts the '+' and '#' flags for most of the verbs as follows.
+//
+//    %+s   path of source file relative to the compile time GOPATH
+//    %#s   full path of source file
+//    %+n   import path qualified function name
+//    %+k   full package path
+//    %+v   equivalent to %+s:%d
+//    %#v   equivalent to %#s:%d
+func (c Call) Format(s fmt.State, verb rune) {
+	if c.frame == (runtime.Frame{}) {
+		fmt.Fprintf(s, "%%!%c(NOFUNC)", verb)
+		return
+	}
+
+	switch verb {
+	case 's', 'v':
+		file := c.frame.File
+		switch {
+		case s.Flag('#'):
+			// done
+		case s.Flag('+'):
+			file = file[pkgIndex(file, c.frame.Function):]
+		default:
+			const sep = "/"
+			if i := strings.LastIndex(file, sep); i != -1 {
+				file = file[i+len(sep):]
+			}
+		}
+		io.WriteString(s, file)
+		if verb == 'v' {
+			buf := [7]byte{':'}
+			s.Write(strconv.AppendInt(buf[:1], int64(c.frame.Line), 10))
+		}
+
+	case 'd':
+		buf := [6]byte{}
+		s.Write(strconv.AppendInt(buf[:0], int64(c.frame.Line), 10))
+
+	case 'k':
+		name := c.frame.Function
+		const pathSep = "/"
+		start, end := 0, len(name)
+		if i := strings.LastIndex(name, pathSep); i != -1 {
+			start = i + len(pathSep)
+		}
+		const pkgSep = "."
+		if i := strings.Index(name[start:], pkgSep); i != -1 {
+			end = start + i
+		}
+		if s.Flag('+') {
+			start = 0
+		}
+		io.WriteString(s, name[start:end])
+
+	case 'n':
+		name := c.frame.Function
+		if !s.Flag('+') {
+			const pathSep = "/"
+			if i := strings.LastIndex(name, pathSep); i != -1 {
+				name = name[i+len(pathSep):]
+			}
+			const pkgSep = "."
+			if i := strings.Index(name, pkgSep); i != -1 {
+				name = name[i+len(pkgSep):]
+			}
+		}
+		io.WriteString(s, name)
+	}
+}
+
+// Frame returns the call frame infomation for the Call.
+func (c Call) Frame() runtime.Frame {
+	return c.frame
+}
+
+// PC returns the program counter for this call frame; multiple frames may
+// have the same PC value.
+//
+// Deprecated: Use Call.Frame instead.
+func (c Call) PC() uintptr {
+	return c.frame.PC
+}
+
+// CallStack records a sequence of function invocations from a goroutine
+// stack.
+type CallStack []Call
+
+// String implements fmt.Stinger. It is equivalent to fmt.Sprintf("%v", cs).
+func (cs CallStack) String() string {
+	return fmt.Sprint(cs)
+}
+
+var (
+	openBracketBytes  = []byte("[")
+	closeBracketBytes = []byte("]")
+	spaceBytes        = []byte(" ")
+)
+
+// MarshalText implements encoding.TextMarshaler. It formats the CallStack the
+// same as fmt.Sprintf("%v", cs).
+func (cs CallStack) MarshalText() ([]byte, error) {
+	buf := bytes.Buffer{}
+	buf.Write(openBracketBytes)
+	for i, pc := range cs {
+		if i > 0 {
+			buf.Write(spaceBytes)
+		}
+		fmt.Fprint(&buf, pc)
+	}
+	buf.Write(closeBracketBytes)
+	return buf.Bytes(), nil
+}
+
+// Format implements fmt.Formatter by printing the CallStack as square brackets
+// ([, ]) surrounding a space separated list of Calls each formatted with the
+// supplied verb and options.
+func (cs CallStack) Format(s fmt.State, verb rune) {
+	s.Write(openBracketBytes)
+	for i, pc := range cs {
+		if i > 0 {
+			s.Write(spaceBytes)
+		}
+		pc.Format(s, verb)
+	}
+	s.Write(closeBracketBytes)
+}
+
+// Trace returns a CallStack for the current goroutine with element 0
+// identifying the calling function.
+func Trace() CallStack {
+	var pcs [512]uintptr
+	n := runtime.Callers(1, pcs[:])
+
+	frames := runtime.CallersFrames(pcs[:n])
+	cs := make(CallStack, 0, n)
+
+	// Skip extra frame retrieved just to make sure the runtime.sigpanic
+	// special case is handled.
+	frame, more := frames.Next()
+
+	for more {
+		frame, more = frames.Next()
+		cs = append(cs, Call{frame: frame})
+	}
+
+	return cs
+}
+
+// TrimBelow returns a slice of the CallStack with all entries below c
+// removed.
+func (cs CallStack) TrimBelow(c Call) CallStack {
+	for len(cs) > 0 && cs[0] != c {
+		cs = cs[1:]
+	}
+	return cs
+}
+
+// TrimAbove returns a slice of the CallStack with all entries above c
+// removed.
+func (cs CallStack) TrimAbove(c Call) CallStack {
+	for len(cs) > 0 && cs[len(cs)-1] != c {
+		cs = cs[:len(cs)-1]
+	}
+	return cs
+}
+
+// pkgIndex returns the index that results in file[index:] being the path of
+// file relative to the compile time GOPATH, and file[:index] being the
+// $GOPATH/src/ portion of file. funcName must be the name of a function in
+// file as returned by runtime.Func.Name.
+func pkgIndex(file, funcName string) int {
+	// As of Go 1.6.2 there is no direct way to know the compile time GOPATH
+	// at runtime, but we can infer the number of path segments in the GOPATH.
+	// We note that runtime.Func.Name() returns the function name qualified by
+	// the import path, which does not include the GOPATH. Thus we can trim
+	// segments from the beginning of the file path until the number of path
+	// separators remaining is one more than the number of path separators in
+	// the function name. For example, given:
+	//
+	//    GOPATH     /home/user
+	//    file       /home/user/src/pkg/sub/file.go
+	//    fn.Name()  pkg/sub.Type.Method
+	//
+	// We want to produce:
+	//
+	//    file[:idx] == /home/user/src/
+	//    file[idx:] == pkg/sub/file.go
+	//
+	// From this we can easily see that fn.Name() has one less path separator
+	// than our desired result for file[idx:]. We count separators from the
+	// end of the file path until it finds two more than in the function name
+	// and then move one character forward to preserve the initial path
+	// segment without a leading separator.
+	const sep = "/"
+	i := len(file)
+	for n := strings.Count(funcName, sep) + 2; n > 0; n-- {
+		i = strings.LastIndex(file[:i], sep)
+		if i == -1 {
+			i = -len(sep)
+			break
+		}
+	}
+	// get back to 0 or trim the leading separator
+	return i + len(sep)
+}
+
+var runtimePath string
+
+func init() {
+	var pcs [3]uintptr
+	runtime.Callers(0, pcs[:])
+	frames := runtime.CallersFrames(pcs[:])
+	frame, _ := frames.Next()
+	file := frame.File
+
+	idx := pkgIndex(frame.File, frame.Function)
+
+	runtimePath = file[:idx]
+	if runtime.GOOS == "windows" {
+		runtimePath = strings.ToLower(runtimePath)
+	}
+}
+
+func inGoroot(c Call) bool {
+	file := c.frame.File
+	if len(file) == 0 || file[0] == '?' {
+		return true
+	}
+	if runtime.GOOS == "windows" {
+		file = strings.ToLower(file)
+	}
+	return strings.HasPrefix(file, runtimePath) || strings.HasSuffix(file, "/_testmain.go")
+}
+
+// TrimRuntime returns a slice of the CallStack with the topmost entries from
+// the go runtime removed. It considers any calls originating from unknown
+// files, files under GOROOT, or _testmain.go as part of the runtime.
+func (cs CallStack) TrimRuntime() CallStack {
+	for len(cs) > 0 && inGoroot(cs[len(cs)-1]) {
+		cs = cs[:len(cs)-1]
+	}
+	return cs
+}
diff --git a/vendor/github.com/gorilla/context/.travis.yml b/vendor/github.com/gorilla/context/.travis.yml
new file mode 100644
index 0000000..faca4da
--- /dev/null
+++ b/vendor/github.com/gorilla/context/.travis.yml
@@ -0,0 +1,19 @@
+language: go
+sudo: false
+
+matrix:
+  include:
+    - go: 1.3
+    - go: 1.4
+    - go: 1.5
+    - go: 1.6
+    - go: tip
+
+install:
+  - go get golang.org/x/tools/cmd/vet
+
+script:
+  - go get -t -v ./...
+  - diff -u <(echo -n) <(gofmt -d .)
+  - go tool vet .
+  - go test -v -race ./...
diff --git a/vendor/github.com/gorilla/context/LICENSE b/vendor/github.com/gorilla/context/LICENSE
new file mode 100644
index 0000000..0e5fb87
--- /dev/null
+++ b/vendor/github.com/gorilla/context/LICENSE
@@ -0,0 +1,27 @@
+Copyright (c) 2012 Rodrigo Moraes. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+	 * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+	 * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+	 * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/vendor/github.com/gorilla/context/README.md b/vendor/github.com/gorilla/context/README.md
new file mode 100644
index 0000000..c60a31b
--- /dev/null
+++ b/vendor/github.com/gorilla/context/README.md
@@ -0,0 +1,7 @@
+context
+=======
+[![Build Status](https://travis-ci.org/gorilla/context.png?branch=master)](https://travis-ci.org/gorilla/context)
+
+gorilla/context is a general purpose registry for global request variables.
+
+Read the full documentation here: http://www.gorillatoolkit.org/pkg/context
diff --git a/vendor/github.com/gorilla/context/context.go b/vendor/github.com/gorilla/context/context.go
new file mode 100644
index 0000000..81cb128
--- /dev/null
+++ b/vendor/github.com/gorilla/context/context.go
@@ -0,0 +1,143 @@
+// Copyright 2012 The Gorilla Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package context
+
+import (
+	"net/http"
+	"sync"
+	"time"
+)
+
+var (
+	mutex sync.RWMutex
+	data  = make(map[*http.Request]map[interface{}]interface{})
+	datat = make(map[*http.Request]int64)
+)
+
+// Set stores a value for a given key in a given request.
+func Set(r *http.Request, key, val interface{}) {
+	mutex.Lock()
+	if data[r] == nil {
+		data[r] = make(map[interface{}]interface{})
+		datat[r] = time.Now().Unix()
+	}
+	data[r][key] = val
+	mutex.Unlock()
+}
+
+// Get returns a value stored for a given key in a given request.
+func Get(r *http.Request, key interface{}) interface{} {
+	mutex.RLock()
+	if ctx := data[r]; ctx != nil {
+		value := ctx[key]
+		mutex.RUnlock()
+		return value
+	}
+	mutex.RUnlock()
+	return nil
+}
+
+// GetOk returns stored value and presence state like multi-value return of map access.
+func GetOk(r *http.Request, key interface{}) (interface{}, bool) {
+	mutex.RLock()
+	if _, ok := data[r]; ok {
+		value, ok := data[r][key]
+		mutex.RUnlock()
+		return value, ok
+	}
+	mutex.RUnlock()
+	return nil, false
+}
+
+// GetAll returns all stored values for the request as a map. Nil is returned for invalid requests.
+func GetAll(r *http.Request) map[interface{}]interface{} {
+	mutex.RLock()
+	if context, ok := data[r]; ok {
+		result := make(map[interface{}]interface{}, len(context))
+		for k, v := range context {
+			result[k] = v
+		}
+		mutex.RUnlock()
+		return result
+	}
+	mutex.RUnlock()
+	return nil
+}
+
+// GetAllOk returns all stored values for the request as a map and a boolean value that indicates if
+// the request was registered.
+func GetAllOk(r *http.Request) (map[interface{}]interface{}, bool) {
+	mutex.RLock()
+	context, ok := data[r]
+	result := make(map[interface{}]interface{}, len(context))
+	for k, v := range context {
+		result[k] = v
+	}
+	mutex.RUnlock()
+	return result, ok
+}
+
+// Delete removes a value stored for a given key in a given request.
+func Delete(r *http.Request, key interface{}) {
+	mutex.Lock()
+	if data[r] != nil {
+		delete(data[r], key)
+	}
+	mutex.Unlock()
+}
+
+// Clear removes all values stored for a given request.
+//
+// This is usually called by a handler wrapper to clean up request
+// variables at the end of a request lifetime. See ClearHandler().
+func Clear(r *http.Request) {
+	mutex.Lock()
+	clear(r)
+	mutex.Unlock()
+}
+
+// clear is Clear without the lock.
+func clear(r *http.Request) {
+	delete(data, r)
+	delete(datat, r)
+}
+
+// Purge removes request data stored for longer than maxAge, in seconds.
+// It returns the amount of requests removed.
+//
+// If maxAge <= 0, all request data is removed.
+//
+// This is only used for sanity check: in case context cleaning was not
+// properly set some request data can be kept forever, consuming an increasing
+// amount of memory. In case this is detected, Purge() must be called
+// periodically until the problem is fixed.
+func Purge(maxAge int) int {
+	mutex.Lock()
+	count := 0
+	if maxAge <= 0 {
+		count = len(data)
+		data = make(map[*http.Request]map[interface{}]interface{})
+		datat = make(map[*http.Request]int64)
+	} else {
+		min := time.Now().Unix() - int64(maxAge)
+		for r := range data {
+			if datat[r] < min {
+				clear(r)
+				count++
+			}
+		}
+	}
+	mutex.Unlock()
+	return count
+}
+
+// ClearHandler wraps an http.Handler and clears request values at the end
+// of a request lifetime.
+func ClearHandler(h http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		defer Clear(r)
+		h.ServeHTTP(w, r)
+	})
+}
diff --git a/vendor/github.com/gorilla/context/doc.go b/vendor/github.com/gorilla/context/doc.go
new file mode 100644
index 0000000..73c7400
--- /dev/null
+++ b/vendor/github.com/gorilla/context/doc.go
@@ -0,0 +1,82 @@
+// Copyright 2012 The Gorilla Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+/*
+Package context stores values shared during a request lifetime.
+
+For example, a router can set variables extracted from the URL and later
+application handlers can access those values, or it can be used to store
+sessions values to be saved at the end of a request. There are several
+others common uses.
+
+The idea was posted by Brad Fitzpatrick to the go-nuts mailing list:
+
+	http://groups.google.com/group/golang-nuts/msg/e2d679d303aa5d53
+
+Here's the basic usage: first define the keys that you will need. The key
+type is interface{} so a key can be of any type that supports equality.
+Here we define a key using a custom int type to avoid name collisions:
+
+	package foo
+
+	import (
+		"github.com/gorilla/context"
+	)
+
+	type key int
+
+	const MyKey key = 0
+
+Then set a variable. Variables are bound to an http.Request object, so you
+need a request instance to set a value:
+
+	context.Set(r, MyKey, "bar")
+
+The application can later access the variable using the same key you provided:
+
+	func MyHandler(w http.ResponseWriter, r *http.Request) {
+		// val is "bar".
+		val := context.Get(r, foo.MyKey)
+
+		// returns ("bar", true)
+		val, ok := context.GetOk(r, foo.MyKey)
+		// ...
+	}
+
+And that's all about the basic usage. We discuss some other ideas below.
+
+Any type can be stored in the context. To enforce a given type, make the key
+private and wrap Get() and Set() to accept and return values of a specific
+type:
+
+	type key int
+
+	const mykey key = 0
+
+	// GetMyKey returns a value for this package from the request values.
+	func GetMyKey(r *http.Request) SomeType {
+		if rv := context.Get(r, mykey); rv != nil {
+			return rv.(SomeType)
+		}
+		return nil
+	}
+
+	// SetMyKey sets a value for this package in the request values.
+	func SetMyKey(r *http.Request, val SomeType) {
+		context.Set(r, mykey, val)
+	}
+
+Variables must be cleared at the end of a request, to remove all values
+that were stored. This can be done in an http.Handler, after a request was
+served. Just call Clear() passing the request:
+
+	context.Clear(r)
+
+...or use ClearHandler(), which conveniently wraps an http.Handler to clear
+variables at the end of a request lifetime.
+
+The Routers from the packages gorilla/mux and gorilla/pat call Clear()
+so if you are using either of them you don't need to clear the context manually.
+*/
+package context
diff --git a/vendor/github.com/gorilla/mux/.travis.yml b/vendor/github.com/gorilla/mux/.travis.yml
new file mode 100644
index 0000000..3302233
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/.travis.yml
@@ -0,0 +1,22 @@
+language: go
+sudo: false
+
+matrix:
+  include:
+    - go: 1.5
+    - go: 1.6
+    - go: 1.7
+    - go: 1.8
+    - go: 1.9
+    - go: tip
+  allow_failures:
+    - go: tip
+
+install:
+  - # Skip
+
+script:
+  - go get -t -v ./...
+  - diff -u <(echo -n) <(gofmt -d .)
+  - go tool vet .
+  - go test -v -race ./...
diff --git a/vendor/github.com/gorilla/mux/ISSUE_TEMPLATE.md b/vendor/github.com/gorilla/mux/ISSUE_TEMPLATE.md
new file mode 100644
index 0000000..232be82
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/ISSUE_TEMPLATE.md
@@ -0,0 +1,11 @@
+**What version of Go are you running?** (Paste the output of `go version`)
+
+
+**What version of gorilla/mux are you at?** (Paste the output of `git rev-parse HEAD` inside `$GOPATH/src/github.com/gorilla/mux`)
+
+
+**Describe your problem** (and what you have tried so far)
+
+
+**Paste a minimal, runnable, reproduction of your issue below** (use backticks to format it)
+
diff --git a/vendor/github.com/gorilla/mux/LICENSE b/vendor/github.com/gorilla/mux/LICENSE
new file mode 100644
index 0000000..0e5fb87
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/LICENSE
@@ -0,0 +1,27 @@
+Copyright (c) 2012 Rodrigo Moraes. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+	 * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+	 * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+	 * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/vendor/github.com/gorilla/mux/README.md b/vendor/github.com/gorilla/mux/README.md
new file mode 100644
index 0000000..f9b3103
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/README.md
@@ -0,0 +1,560 @@
+gorilla/mux
+===
+[![GoDoc](https://godoc.org/github.com/gorilla/mux?status.svg)](https://godoc.org/github.com/gorilla/mux)
+[![Build Status](https://travis-ci.org/gorilla/mux.svg?branch=master)](https://travis-ci.org/gorilla/mux)
+[![Sourcegraph](https://sourcegraph.com/github.com/gorilla/mux/-/badge.svg)](https://sourcegraph.com/github.com/gorilla/mux?badge)
+
+![Gorilla Logo](http://www.gorillatoolkit.org/static/images/gorilla-icon-64.png)
+
+http://www.gorillatoolkit.org/pkg/mux
+
+Package `gorilla/mux` implements a request router and dispatcher for matching incoming requests to
+their respective handler.
+
+The name mux stands for "HTTP request multiplexer". Like the standard `http.ServeMux`, `mux.Router` matches incoming requests against a list of registered routes and calls a handler for the route that matches the URL or other conditions. The main features are:
+
+* It implements the `http.Handler` interface so it is compatible with the standard `http.ServeMux`.
+* Requests can be matched based on URL host, path, path prefix, schemes, header and query values, HTTP methods or using custom matchers.
+* URL hosts, paths and query values can have variables with an optional regular expression.
+* Registered URLs can be built, or "reversed", which helps maintaining references to resources.
+* Routes can be used as subrouters: nested routes are only tested if the parent route matches. This is useful to define groups of routes that share common conditions like a host, a path prefix or other repeated attributes. As a bonus, this optimizes request matching.
+
+---
+
+* [Install](#install)
+* [Examples](#examples)
+* [Matching Routes](#matching-routes)
+* [Static Files](#static-files)
+* [Registered URLs](#registered-urls)
+* [Walking Routes](#walking-routes)
+* [Graceful Shutdown](#graceful-shutdown)
+* [Middleware](#middleware)
+* [Full Example](#full-example)
+
+---
+
+## Install
+
+With a [correctly configured](https://golang.org/doc/install#testing) Go toolchain:
+
+```sh
+go get -u github.com/gorilla/mux
+```
+
+## Examples
+
+Let's start registering a couple of URL paths and handlers:
+
+```go
+func main() {
+    r := mux.NewRouter()
+    r.HandleFunc("/", HomeHandler)
+    r.HandleFunc("/products", ProductsHandler)
+    r.HandleFunc("/articles", ArticlesHandler)
+    http.Handle("/", r)
+}
+```
+
+Here we register three routes mapping URL paths to handlers. This is equivalent to how `http.HandleFunc()` works: if an incoming request URL matches one of the paths, the corresponding handler is called passing (`http.ResponseWriter`, `*http.Request`) as parameters.
+
+Paths can have variables. They are defined using the format `{name}` or `{name:pattern}`. If a regular expression pattern is not defined, the matched variable will be anything until the next slash. For example:
+
+```go
+r := mux.NewRouter()
+r.HandleFunc("/products/{key}", ProductHandler)
+r.HandleFunc("/articles/{category}/", ArticlesCategoryHandler)
+r.HandleFunc("/articles/{category}/{id:[0-9]+}", ArticleHandler)
+```
+
+The names are used to create a map of route variables which can be retrieved calling `mux.Vars()`:
+
+```go
+func ArticlesCategoryHandler(w http.ResponseWriter, r *http.Request) {
+    vars := mux.Vars(r)
+    w.WriteHeader(http.StatusOK)
+    fmt.Fprintf(w, "Category: %v\n", vars["category"])
+}
+```
+
+And this is all you need to know about the basic usage. More advanced options are explained below.
+
+### Matching Routes
+
+Routes can also be restricted to a domain or subdomain. Just define a host pattern to be matched. They can also have variables:
+
+```go
+r := mux.NewRouter()
+// Only matches if domain is "www.example.com".
+r.Host("www.example.com")
+// Matches a dynamic subdomain.
+r.Host("{subdomain:[a-z]+}.domain.com")
+```
+
+There are several other matchers that can be added. To match path prefixes:
+
+```go
+r.PathPrefix("/products/")
+```
+
+...or HTTP methods:
+
+```go
+r.Methods("GET", "POST")
+```
+
+...or URL schemes:
+
+```go
+r.Schemes("https")
+```
+
+...or header values:
+
+```go
+r.Headers("X-Requested-With", "XMLHttpRequest")
+```
+
+...or query values:
+
+```go
+r.Queries("key", "value")
+```
+
+...or to use a custom matcher function:
+
+```go
+r.MatcherFunc(func(r *http.Request, rm *RouteMatch) bool {
+    return r.ProtoMajor == 0
+})
+```
+
+...and finally, it is possible to combine several matchers in a single route:
+
+```go
+r.HandleFunc("/products", ProductsHandler).
+  Host("www.example.com").
+  Methods("GET").
+  Schemes("http")
+```
+
+Routes are tested in the order they were added to the router. If two routes match, the first one wins:
+
+```go
+r := mux.NewRouter()
+r.HandleFunc("/specific", specificHandler)
+r.PathPrefix("/").Handler(catchAllHandler)
+```
+
+Setting the same matching conditions again and again can be boring, so we have a way to group several routes that share the same requirements. We call it "subrouting".
+
+For example, let's say we have several URLs that should only match when the host is `www.example.com`. Create a route for that host and get a "subrouter" from it:
+
+```go
+r := mux.NewRouter()
+s := r.Host("www.example.com").Subrouter()
+```
+
+Then register routes in the subrouter:
+
+```go
+s.HandleFunc("/products/", ProductsHandler)
+s.HandleFunc("/products/{key}", ProductHandler)
+s.HandleFunc("/articles/{category}/{id:[0-9]+}", ArticleHandler)
+```
+
+The three URL paths we registered above will only be tested if the domain is `www.example.com`, because the subrouter is tested first. This is not only convenient, but also optimizes request matching. You can create subrouters combining any attribute matchers accepted by a route.
+
+Subrouters can be used to create domain or path "namespaces": you define subrouters in a central place and then parts of the app can register its paths relatively to a given subrouter.
+
+There's one more thing about subroutes. When a subrouter has a path prefix, the inner routes use it as base for their paths:
+
+```go
+r := mux.NewRouter()
+s := r.PathPrefix("/products").Subrouter()
+// "/products/"
+s.HandleFunc("/", ProductsHandler)
+// "/products/{key}/"
+s.HandleFunc("/{key}/", ProductHandler)
+// "/products/{key}/details"
+s.HandleFunc("/{key}/details", ProductDetailsHandler)
+```
+### Listing Routes
+
+Routes on a mux can be listed using the Router.Walk method—useful for generating documentation:
+
+```go
+package main
+
+import (
+    "fmt"
+    "net/http"
+    "strings"
+
+    "github.com/gorilla/mux"
+)
+
+func handler(w http.ResponseWriter, r *http.Request) {
+    return
+}
+
+func main() {
+    r := mux.NewRouter()
+    r.HandleFunc("/", handler)
+    r.HandleFunc("/products", handler).Methods("POST")
+    r.HandleFunc("/articles", handler).Methods("GET")
+    r.HandleFunc("/articles/{id}", handler).Methods("GET", "PUT")
+    r.HandleFunc("/authors", handler).Queries("surname", "{surname}")
+    r.Walk(func(route *mux.Route, router *mux.Router, ancestors []*mux.Route) error {
+        t, err := route.GetPathTemplate()
+        if err != nil {
+            return err
+        }
+        qt, err := route.GetQueriesTemplates()
+        if err != nil {
+            return err
+        }
+        // p will contain regular expression is compatible with regular expression in Perl, Python, and other languages.
+        // for instance the regular expression for path '/articles/{id}' will be '^/articles/(?P<v0>[^/]+)$'
+        p, err := route.GetPathRegexp()
+        if err != nil {
+            return err
+        }
+        // qr will contain a list of regular expressions with the same semantics as GetPathRegexp,
+        // just applied to the Queries pairs instead, e.g., 'Queries("surname", "{surname}") will return
+        // {"^surname=(?P<v0>.*)$}. Where each combined query pair will have an entry in the list.
+        qr, err := route.GetQueriesRegexp()
+        if err != nil {
+            return err
+        }
+        m, err := route.GetMethods()
+        if err != nil {
+            return err
+        }
+        fmt.Println(strings.Join(m, ","), strings.Join(qt, ","), strings.Join(qr, ","), t, p)
+        return nil
+    })
+    http.Handle("/", r)
+}
+```
+
+### Static Files
+
+Note that the path provided to `PathPrefix()` represents a "wildcard": calling
+`PathPrefix("/static/").Handler(...)` means that the handler will be passed any
+request that matches "/static/*". This makes it easy to serve static files with mux:
+
+```go
+func main() {
+    var dir string
+
+    flag.StringVar(&dir, "dir", ".", "the directory to serve files from. Defaults to the current dir")
+    flag.Parse()
+    r := mux.NewRouter()
+
+    // This will serve files under http://localhost:8000/static/<filename>
+    r.PathPrefix("/static/").Handler(http.StripPrefix("/static/", http.FileServer(http.Dir(dir))))
+
+    srv := &http.Server{
+        Handler:      r,
+        Addr:         "127.0.0.1:8000",
+        // Good practice: enforce timeouts for servers you create!
+        WriteTimeout: 15 * time.Second,
+        ReadTimeout:  15 * time.Second,
+    }
+
+    log.Fatal(srv.ListenAndServe())
+}
+```
+
+### Registered URLs
+
+Now let's see how to build registered URLs.
+
+Routes can be named. All routes that define a name can have their URLs built, or "reversed". We define a name calling `Name()` on a route. For example:
+
+```go
+r := mux.NewRouter()
+r.HandleFunc("/articles/{category}/{id:[0-9]+}", ArticleHandler).
+  Name("article")
+```
+
+To build a URL, get the route and call the `URL()` method, passing a sequence of key/value pairs for the route variables. For the previous route, we would do:
+
+```go
+url, err := r.Get("article").URL("category", "technology", "id", "42")
+```
+
+...and the result will be a `url.URL` with the following path:
+
+```
+"/articles/technology/42"
+```
+
+This also works for host and query value variables:
+
+```go
+r := mux.NewRouter()
+r.Host("{subdomain}.domain.com").
+  Path("/articles/{category}/{id:[0-9]+}").
+  Queries("filter", "{filter}").
+  HandlerFunc(ArticleHandler).
+  Name("article")
+
+// url.String() will be "http://news.domain.com/articles/technology/42?filter=gorilla"
+url, err := r.Get("article").URL("subdomain", "news",
+                                 "category", "technology",
+                                 "id", "42",
+                                 "filter", "gorilla")
+```
+
+All variables defined in the route are required, and their values must conform to the corresponding patterns. These requirements guarantee that a generated URL will always match a registered route -- the only exception is for explicitly defined "build-only" routes which never match.
+
+Regex support also exists for matching Headers within a route. For example, we could do:
+
+```go
+r.HeadersRegexp("Content-Type", "application/(text|json)")
+```
+
+...and the route will match both requests with a Content-Type of `application/json` as well as `application/text`
+
+There's also a way to build only the URL host or path for a route: use the methods `URLHost()` or `URLPath()` instead. For the previous route, we would do:
+
+```go
+// "http://news.domain.com/"
+host, err := r.Get("article").URLHost("subdomain", "news")
+
+// "/articles/technology/42"
+path, err := r.Get("article").URLPath("category", "technology", "id", "42")
+```
+
+And if you use subrouters, host and path defined separately can be built as well:
+
+```go
+r := mux.NewRouter()
+s := r.Host("{subdomain}.domain.com").Subrouter()
+s.Path("/articles/{category}/{id:[0-9]+}").
+  HandlerFunc(ArticleHandler).
+  Name("article")
+
+// "http://news.domain.com/articles/technology/42"
+url, err := r.Get("article").URL("subdomain", "news",
+                                 "category", "technology",
+                                 "id", "42")
+```
+
+### Walking Routes
+
+The `Walk` function on `mux.Router` can be used to visit all of the routes that are registered on a router. For example,
+the following prints all of the registered routes:
+
+```go
+r := mux.NewRouter()
+r.HandleFunc("/", handler)
+r.HandleFunc("/products", handler).Methods("POST")
+r.HandleFunc("/articles", handler).Methods("GET")
+r.HandleFunc("/articles/{id}", handler).Methods("GET", "PUT")
+r.HandleFunc("/authors", handler).Queries("surname", "{surname}")
+r.Walk(func(route *mux.Route, router *mux.Router, ancestors []*mux.Route) error {
+    t, err := route.GetPathTemplate()
+    if err != nil {
+        return err
+    }
+    qt, err := route.GetQueriesTemplates()
+    if err != nil {
+        return err
+    }
+    // p will contain a regular expression that is compatible with regular expressions in Perl, Python, and other languages.
+    // For example, the regular expression for path '/articles/{id}' will be '^/articles/(?P<v0>[^/]+)$'.
+    p, err := route.GetPathRegexp()
+    if err != nil {
+        return err
+    }
+    // qr will contain a list of regular expressions with the same semantics as GetPathRegexp,
+    // just applied to the Queries pairs instead, e.g., 'Queries("surname", "{surname}") will return
+    // {"^surname=(?P<v0>.*)$}. Where each combined query pair will have an entry in the list.
+    qr, err := route.GetQueriesRegexp()
+    if err != nil {
+        return err
+    }
+    m, err := route.GetMethods()
+    if err != nil {
+        return err
+    }
+    fmt.Println(strings.Join(m, ","), strings.Join(qt, ","), strings.Join(qr, ","), t, p)
+    return nil
+})
+```
+
+### Graceful Shutdown
+
+Go 1.8 introduced the ability to [gracefully shutdown](https://golang.org/doc/go1.8#http_shutdown) a `*http.Server`. Here's how to do that alongside `mux`:
+
+```go
+package main
+
+import (
+    "context"
+    "flag"
+    "log"
+    "net/http"
+    "os"
+    "os/signal"
+
+    "github.com/gorilla/mux"
+)
+
+func main() {
+    var wait time.Duration
+    flag.DurationVar(&wait, "graceful-timeout", time.Second * 15, "the duration for which the server gracefully wait for existing connections to finish - e.g. 15s or 1m")
+    flag.Parse()
+
+    r := mux.NewRouter()
+    // Add your routes as needed
+    
+    srv := &http.Server{
+        Addr:         "0.0.0.0:8080",
+        // Good practice to set timeouts to avoid Slowloris attacks.
+        WriteTimeout: time.Second * 15,
+        ReadTimeout:  time.Second * 15,
+        IdleTimeout:  time.Second * 60,
+        Handler: r, // Pass our instance of gorilla/mux in.
+    }
+
+    // Run our server in a goroutine so that it doesn't block.
+    go func() {
+        if err := srv.ListenAndServe(); err != nil {
+            log.Println(err)
+        }
+    }()
+    
+    c := make(chan os.Signal, 1)
+    // We'll accept graceful shutdowns when quit via SIGINT (Ctrl+C)
+    // SIGKILL, SIGQUIT or SIGTERM (Ctrl+/) will not be caught.
+    signal.Notify(c, os.Interrupt)
+
+    // Block until we receive our signal.
+    <-c
+
+    // Create a deadline to wait for.
+    ctx, cancel := context.WithTimeout(ctx, wait)
+    // Doesn't block if no connections, but will otherwise wait
+    // until the timeout deadline.
+    srv.Shutdown(ctx)
+    // Optionally, you could run srv.Shutdown in a goroutine and block on
+    // <-ctx.Done() if your application should wait for other services
+    // to finalize based on context cancellation.
+    log.Println("shutting down")
+    os.Exit(0)
+}
+```
+
+### Middleware
+
+Mux supports the addition of middlewares to a [Router](https://godoc.org/github.com/gorilla/mux#Router), which are executed in the order they are added if a match is found, including its subrouters.
+Middlewares are (typically) small pieces of code which take one request, do something with it, and pass it down to another middleware or the final handler. Some common use cases for middleware are request logging, header manipulation, or `ResponseWriter` hijacking.
+
+Mux middlewares are defined using the de facto standard type:
+
+```go
+type MiddlewareFunc func(http.Handler) http.Handler
+```
+
+Typically, the returned handler is a closure which does something with the http.ResponseWriter and http.Request passed to it, and then calls the handler passed as parameter to the MiddlewareFunc. This takes advantage of closures being able access variables from the context where they are created, while retaining the signature enforced by the receivers.
+
+A very basic middleware which logs the URI of the request being handled could be written as:
+
+```go
+func simpleMw(next http.Handler) http.Handler {
+    return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+        // Do stuff here
+        log.Println(r.RequestURI)
+        // Call the next handler, which can be another middleware in the chain, or the final handler.
+        next.ServeHTTP(w, r)
+    })
+}
+```
+
+Middlewares can be added to a router using `Router.AddMiddlewareFunc()`:
+
+```go
+r := mux.NewRouter()
+r.HandleFunc("/", handler)
+r.AddMiddleware(simpleMw)
+```
+
+A more complex authentication middleware, which maps session token to users, could be written as:
+
+```go
+// Define our struct
+type authenticationMiddleware struct {
+	tokenUsers map[string]string
+}
+
+// Initialize it somewhere
+func (amw *authenticationMiddleware) Populate() {
+	amw.tokenUsers["00000000"] = "user0"
+	amw.tokenUsers["aaaaaaaa"] = "userA"
+	amw.tokenUsers["05f717e5"] = "randomUser"
+	amw.tokenUsers["deadbeef"] = "user0"
+}
+
+// Middleware function, which will be called for each request
+func (amw *authenticationMiddleware) Middleware(next http.Handler) http.Handler {
+    return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+        token := r.Header.Get("X-Session-Token")
+        
+        if user, found := amw.tokenUsers[token]; found {
+        	// We found the token in our map
+        	log.Printf("Authenticated user %s\n", user)
+        	// Pass down the request to the next middleware (or final handler)
+        	next.ServeHTTP(w, r)
+        } else {
+        	// Write an error and stop the handler chain
+        	http.Error(w, "Forbidden", 403)
+        }
+    })
+}
+```
+
+```go
+r := mux.NewRouter()
+r.HandleFunc("/", handler)
+
+amw := authenticationMiddleware{}
+amw.Populate()
+
+r.AddMiddlewareFunc(amw.Middleware)
+```
+
+Note: The handler chain will be stopped if your middleware doesn't call `next.ServeHTTP()` with the corresponding parameters. This can be used to abort a request if the middleware writer wants to. Middlewares *should* write to `ResponseWriter` if they *are* going to terminate the request, and they *should not* write to `ResponseWriter` if they *are not* going to terminate it.
+
+## Full Example
+
+Here's a complete, runnable example of a small `mux` based server:
+
+```go
+package main
+
+import (
+    "net/http"
+    "log"
+    "github.com/gorilla/mux"
+)
+
+func YourHandler(w http.ResponseWriter, r *http.Request) {
+    w.Write([]byte("Gorilla!\n"))
+}
+
+func main() {
+    r := mux.NewRouter()
+    // Routes consist of a path and a handler function.
+    r.HandleFunc("/", YourHandler)
+
+    // Bind to a port and pass our router in
+    log.Fatal(http.ListenAndServe(":8000", r))
+}
+```
+
+## License
+
+BSD licensed. See the LICENSE file for details.
diff --git a/vendor/github.com/gorilla/mux/context_gorilla.go b/vendor/github.com/gorilla/mux/context_gorilla.go
new file mode 100644
index 0000000..d7adaa8
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/context_gorilla.go
@@ -0,0 +1,26 @@
+// +build !go1.7
+
+package mux
+
+import (
+	"net/http"
+
+	"github.com/gorilla/context"
+)
+
+func contextGet(r *http.Request, key interface{}) interface{} {
+	return context.Get(r, key)
+}
+
+func contextSet(r *http.Request, key, val interface{}) *http.Request {
+	if val == nil {
+		return r
+	}
+
+	context.Set(r, key, val)
+	return r
+}
+
+func contextClear(r *http.Request) {
+	context.Clear(r)
+}
diff --git a/vendor/github.com/gorilla/mux/context_native.go b/vendor/github.com/gorilla/mux/context_native.go
new file mode 100644
index 0000000..209cbea
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/context_native.go
@@ -0,0 +1,24 @@
+// +build go1.7
+
+package mux
+
+import (
+	"context"
+	"net/http"
+)
+
+func contextGet(r *http.Request, key interface{}) interface{} {
+	return r.Context().Value(key)
+}
+
+func contextSet(r *http.Request, key, val interface{}) *http.Request {
+	if val == nil {
+		return r
+	}
+
+	return r.WithContext(context.WithValue(r.Context(), key, val))
+}
+
+func contextClear(r *http.Request) {
+	return
+}
diff --git a/vendor/github.com/gorilla/mux/doc.go b/vendor/github.com/gorilla/mux/doc.go
new file mode 100644
index 0000000..013f088
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/doc.go
@@ -0,0 +1,307 @@
+// Copyright 2012 The Gorilla Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+/*
+Package mux implements a request router and dispatcher.
+
+The name mux stands for "HTTP request multiplexer". Like the standard
+http.ServeMux, mux.Router matches incoming requests against a list of
+registered routes and calls a handler for the route that matches the URL
+or other conditions. The main features are:
+
+	* Requests can be matched based on URL host, path, path prefix, schemes,
+	  header and query values, HTTP methods or using custom matchers.
+	* URL hosts, paths and query values can have variables with an optional
+	  regular expression.
+	* Registered URLs can be built, or "reversed", which helps maintaining
+	  references to resources.
+	* Routes can be used as subrouters: nested routes are only tested if the
+	  parent route matches. This is useful to define groups of routes that
+	  share common conditions like a host, a path prefix or other repeated
+	  attributes. As a bonus, this optimizes request matching.
+	* It implements the http.Handler interface so it is compatible with the
+	  standard http.ServeMux.
+
+Let's start registering a couple of URL paths and handlers:
+
+	func main() {
+		r := mux.NewRouter()
+		r.HandleFunc("/", HomeHandler)
+		r.HandleFunc("/products", ProductsHandler)
+		r.HandleFunc("/articles", ArticlesHandler)
+		http.Handle("/", r)
+	}
+
+Here we register three routes mapping URL paths to handlers. This is
+equivalent to how http.HandleFunc() works: if an incoming request URL matches
+one of the paths, the corresponding handler is called passing
+(http.ResponseWriter, *http.Request) as parameters.
+
+Paths can have variables. They are defined using the format {name} or
+{name:pattern}. If a regular expression pattern is not defined, the matched
+variable will be anything until the next slash. For example:
+
+	r := mux.NewRouter()
+	r.HandleFunc("/products/{key}", ProductHandler)
+	r.HandleFunc("/articles/{category}/", ArticlesCategoryHandler)
+	r.HandleFunc("/articles/{category}/{id:[0-9]+}", ArticleHandler)
+
+Groups can be used inside patterns, as long as they are non-capturing (?:re). For example:
+
+	r.HandleFunc("/articles/{category}/{sort:(?:asc|desc|new)}", ArticlesCategoryHandler)
+
+The names are used to create a map of route variables which can be retrieved
+calling mux.Vars():
+
+	vars := mux.Vars(request)
+	category := vars["category"]
+
+Note that if any capturing groups are present, mux will panic() during parsing. To prevent
+this, convert any capturing groups to non-capturing, e.g. change "/{sort:(asc|desc)}" to
+"/{sort:(?:asc|desc)}". This is a change from prior versions which behaved unpredictably
+when capturing groups were present.
+
+And this is all you need to know about the basic usage. More advanced options
+are explained below.
+
+Routes can also be restricted to a domain or subdomain. Just define a host
+pattern to be matched. They can also have variables:
+
+	r := mux.NewRouter()
+	// Only matches if domain is "www.example.com".
+	r.Host("www.example.com")
+	// Matches a dynamic subdomain.
+	r.Host("{subdomain:[a-z]+}.domain.com")
+
+There are several other matchers that can be added. To match path prefixes:
+
+	r.PathPrefix("/products/")
+
+...or HTTP methods:
+
+	r.Methods("GET", "POST")
+
+...or URL schemes:
+
+	r.Schemes("https")
+
+...or header values:
+
+	r.Headers("X-Requested-With", "XMLHttpRequest")
+
+...or query values:
+
+	r.Queries("key", "value")
+
+...or to use a custom matcher function:
+
+	r.MatcherFunc(func(r *http.Request, rm *RouteMatch) bool {
+		return r.ProtoMajor == 0
+	})
+
+...and finally, it is possible to combine several matchers in a single route:
+
+	r.HandleFunc("/products", ProductsHandler).
+	  Host("www.example.com").
+	  Methods("GET").
+	  Schemes("http")
+
+Setting the same matching conditions again and again can be boring, so we have
+a way to group several routes that share the same requirements.
+We call it "subrouting".
+
+For example, let's say we have several URLs that should only match when the
+host is "www.example.com". Create a route for that host and get a "subrouter"
+from it:
+
+	r := mux.NewRouter()
+	s := r.Host("www.example.com").Subrouter()
+
+Then register routes in the subrouter:
+
+	s.HandleFunc("/products/", ProductsHandler)
+	s.HandleFunc("/products/{key}", ProductHandler)
+	s.HandleFunc("/articles/{category}/{id:[0-9]+}"), ArticleHandler)
+
+The three URL paths we registered above will only be tested if the domain is
+"www.example.com", because the subrouter is tested first. This is not
+only convenient, but also optimizes request matching. You can create
+subrouters combining any attribute matchers accepted by a route.
+
+Subrouters can be used to create domain or path "namespaces": you define
+subrouters in a central place and then parts of the app can register its
+paths relatively to a given subrouter.
+
+There's one more thing about subroutes. When a subrouter has a path prefix,
+the inner routes use it as base for their paths:
+
+	r := mux.NewRouter()
+	s := r.PathPrefix("/products").Subrouter()
+	// "/products/"
+	s.HandleFunc("/", ProductsHandler)
+	// "/products/{key}/"
+	s.HandleFunc("/{key}/", ProductHandler)
+	// "/products/{key}/details"
+	s.HandleFunc("/{key}/details", ProductDetailsHandler)
+
+Note that the path provided to PathPrefix() represents a "wildcard": calling
+PathPrefix("/static/").Handler(...) means that the handler will be passed any
+request that matches "/static/*". This makes it easy to serve static files with mux:
+
+	func main() {
+		var dir string
+
+		flag.StringVar(&dir, "dir", ".", "the directory to serve files from. Defaults to the current dir")
+		flag.Parse()
+		r := mux.NewRouter()
+
+		// This will serve files under http://localhost:8000/static/<filename>
+		r.PathPrefix("/static/").Handler(http.StripPrefix("/static/", http.FileServer(http.Dir(dir))))
+
+		srv := &http.Server{
+			Handler:      r,
+			Addr:         "127.0.0.1:8000",
+			// Good practice: enforce timeouts for servers you create!
+			WriteTimeout: 15 * time.Second,
+			ReadTimeout:  15 * time.Second,
+		}
+
+		log.Fatal(srv.ListenAndServe())
+	}
+
+Now let's see how to build registered URLs.
+
+Routes can be named. All routes that define a name can have their URLs built,
+or "reversed". We define a name calling Name() on a route. For example:
+
+	r := mux.NewRouter()
+	r.HandleFunc("/articles/{category}/{id:[0-9]+}", ArticleHandler).
+	  Name("article")
+
+To build a URL, get the route and call the URL() method, passing a sequence of
+key/value pairs for the route variables. For the previous route, we would do:
+
+	url, err := r.Get("article").URL("category", "technology", "id", "42")
+
+...and the result will be a url.URL with the following path:
+
+	"/articles/technology/42"
+
+This also works for host and query value variables:
+
+	r := mux.NewRouter()
+	r.Host("{subdomain}.domain.com").
+	  Path("/articles/{category}/{id:[0-9]+}").
+	  Queries("filter", "{filter}").
+	  HandlerFunc(ArticleHandler).
+	  Name("article")
+
+	// url.String() will be "http://news.domain.com/articles/technology/42?filter=gorilla"
+	url, err := r.Get("article").URL("subdomain", "news",
+	                                 "category", "technology",
+	                                 "id", "42",
+	                                 "filter", "gorilla")
+
+All variables defined in the route are required, and their values must
+conform to the corresponding patterns. These requirements guarantee that a
+generated URL will always match a registered route -- the only exception is
+for explicitly defined "build-only" routes which never match.
+
+Regex support also exists for matching Headers within a route. For example, we could do:
+
+	r.HeadersRegexp("Content-Type", "application/(text|json)")
+
+...and the route will match both requests with a Content-Type of `application/json` as well as
+`application/text`
+
+There's also a way to build only the URL host or path for a route:
+use the methods URLHost() or URLPath() instead. For the previous route,
+we would do:
+
+	// "http://news.domain.com/"
+	host, err := r.Get("article").URLHost("subdomain", "news")
+
+	// "/articles/technology/42"
+	path, err := r.Get("article").URLPath("category", "technology", "id", "42")
+
+And if you use subrouters, host and path defined separately can be built
+as well:
+
+	r := mux.NewRouter()
+	s := r.Host("{subdomain}.domain.com").Subrouter()
+	s.Path("/articles/{category}/{id:[0-9]+}").
+	  HandlerFunc(ArticleHandler).
+	  Name("article")
+
+	// "http://news.domain.com/articles/technology/42"
+	url, err := r.Get("article").URL("subdomain", "news",
+	                                 "category", "technology",
+	                                 "id", "42")
+
+Since **vX.Y.Z**, mux supports the addition of middlewares to a [Router](https://godoc.org/github.com/gorilla/mux#Router), which are executed if a
+match is found (including subrouters). Middlewares are defined using the de facto standard type:
+
+	type MiddlewareFunc func(http.Handler) http.Handler
+
+Typically, the returned handler is a closure which does something with the http.ResponseWriter and http.Request passed to it, and then calls the handler passed as parameter to the MiddlewareFunc (closures can access variables from the context where they are created).
+
+A very basic middleware which logs the URI of the request being handled could be written as:
+
+	func simpleMw(next http.Handler) http.Handler {
+		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+			// Do stuff here
+			log.Println(r.RequestURI)
+			// Call the next handler, which can be another middleware in the chain, or the final handler.
+			next.ServeHTTP(w, r)
+		})
+	}
+
+Middlewares can be added to a router using `Router.Use()`:
+
+	r := mux.NewRouter()
+	r.HandleFunc("/", handler)
+	r.AddMiddleware(simpleMw)
+
+A more complex authentication middleware, which maps session token to users, could be written as:
+
+	// Define our struct
+	type authenticationMiddleware struct {
+		tokenUsers map[string]string
+	}
+
+	// Initialize it somewhere
+	func (amw *authenticationMiddleware) Populate() {
+		amw.tokenUsers["00000000"] = "user0"
+		amw.tokenUsers["aaaaaaaa"] = "userA"
+		amw.tokenUsers["05f717e5"] = "randomUser"
+		amw.tokenUsers["deadbeef"] = "user0"
+	}
+
+	// Middleware function, which will be called for each request
+	func (amw *authenticationMiddleware) Middleware(next http.Handler) http.Handler {
+		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+			token := r.Header.Get("X-Session-Token")
+
+			if user, found := amw.tokenUsers[token]; found {
+				// We found the token in our map
+				log.Printf("Authenticated user %s\n", user)
+				next.ServeHTTP(w, r)
+			} else {
+				http.Error(w, "Forbidden", 403)
+			}
+		})
+	}
+
+	r := mux.NewRouter()
+	r.HandleFunc("/", handler)
+
+	amw := authenticationMiddleware{}
+	amw.Populate()
+
+	r.Use(amw.Middleware)
+
+Note: The handler chain will be stopped if your middleware doesn't call `next.ServeHTTP()` with the corresponding parameters. This can be used to abort a request if the middleware writer wants to.
+
+*/
+package mux
diff --git a/vendor/github.com/gorilla/mux/middleware.go b/vendor/github.com/gorilla/mux/middleware.go
new file mode 100644
index 0000000..8f89867
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/middleware.go
@@ -0,0 +1,28 @@
+package mux
+
+import "net/http"
+
+// MiddlewareFunc is a function which receives an http.Handler and returns another http.Handler.
+// Typically, the returned handler is a closure which does something with the http.ResponseWriter and http.Request passed
+// to it, and then calls the handler passed as parameter to the MiddlewareFunc.
+type MiddlewareFunc func(http.Handler) http.Handler
+
+// middleware interface is anything which implements a MiddlewareFunc named Middleware.
+type middleware interface {
+	Middleware(handler http.Handler) http.Handler
+}
+
+// MiddlewareFunc also implements the middleware interface.
+func (mw MiddlewareFunc) Middleware(handler http.Handler) http.Handler {
+	return mw(handler)
+}
+
+// Use appends a MiddlewareFunc to the chain. Middleware can be used to intercept or otherwise modify requests and/or responses, and are executed in the order that they are applied to the Router.
+func (r *Router) Use(mwf MiddlewareFunc) {
+	r.middlewares = append(r.middlewares, mwf)
+}
+
+// useInterface appends a middleware to the chain. Middleware can be used to intercept or otherwise modify requests and/or responses, and are executed in the order that they are applied to the Router.
+func (r *Router) useInterface(mw middleware) {
+	r.middlewares = append(r.middlewares, mw)
+}
diff --git a/vendor/github.com/gorilla/mux/mux.go b/vendor/github.com/gorilla/mux/mux.go
new file mode 100644
index 0000000..efabd24
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/mux.go
@@ -0,0 +1,585 @@
+// Copyright 2012 The Gorilla Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package mux
+
+import (
+	"errors"
+	"fmt"
+	"net/http"
+	"path"
+	"regexp"
+)
+
+var (
+	ErrMethodMismatch = errors.New("method is not allowed")
+	ErrNotFound       = errors.New("no matching route was found")
+)
+
+// NewRouter returns a new router instance.
+func NewRouter() *Router {
+	return &Router{namedRoutes: make(map[string]*Route), KeepContext: false}
+}
+
+// Router registers routes to be matched and dispatches a handler.
+//
+// It implements the http.Handler interface, so it can be registered to serve
+// requests:
+//
+//     var router = mux.NewRouter()
+//
+//     func main() {
+//         http.Handle("/", router)
+//     }
+//
+// Or, for Google App Engine, register it in a init() function:
+//
+//     func init() {
+//         http.Handle("/", router)
+//     }
+//
+// This will send all incoming requests to the router.
+type Router struct {
+	// Configurable Handler to be used when no route matches.
+	NotFoundHandler http.Handler
+
+	// Configurable Handler to be used when the request method does not match the route.
+	MethodNotAllowedHandler http.Handler
+
+	// Parent route, if this is a subrouter.
+	parent parentRoute
+	// Routes to be matched, in order.
+	routes []*Route
+	// Routes by name for URL building.
+	namedRoutes map[string]*Route
+	// See Router.StrictSlash(). This defines the flag for new routes.
+	strictSlash bool
+	// See Router.SkipClean(). This defines the flag for new routes.
+	skipClean bool
+	// If true, do not clear the request context after handling the request.
+	// This has no effect when go1.7+ is used, since the context is stored
+	// on the request itself.
+	KeepContext bool
+	// see Router.UseEncodedPath(). This defines a flag for all routes.
+	useEncodedPath bool
+	// Slice of middlewares to be called after a match is found
+	middlewares []middleware
+}
+
+// Match attempts to match the given request against the router's registered routes.
+//
+// If the request matches a route of this router or one of its subrouters the Route,
+// Handler, and Vars fields of the the match argument are filled and this function
+// returns true.
+//
+// If the request does not match any of this router's or its subrouters' routes
+// then this function returns false. If available, a reason for the match failure
+// will be filled in the match argument's MatchErr field. If the match failure type
+// (eg: not found) has a registered handler, the handler is assigned to the Handler
+// field of the match argument.
+func (r *Router) Match(req *http.Request, match *RouteMatch) bool {
+	for _, route := range r.routes {
+		if route.Match(req, match) {
+			// Build middleware chain if no error was found
+			if match.MatchErr == nil {
+				for i := len(r.middlewares) - 1; i >= 0; i-- {
+					match.Handler = r.middlewares[i].Middleware(match.Handler)
+				}
+			}
+			return true
+		}
+	}
+
+	if match.MatchErr == ErrMethodMismatch {
+		if r.MethodNotAllowedHandler != nil {
+			match.Handler = r.MethodNotAllowedHandler
+			return true
+		} else {
+			return false
+		}
+	}
+
+	// Closest match for a router (includes sub-routers)
+	if r.NotFoundHandler != nil {
+		match.Handler = r.NotFoundHandler
+		match.MatchErr = ErrNotFound
+		return true
+	}
+
+	match.MatchErr = ErrNotFound
+	return false
+}
+
+// ServeHTTP dispatches the handler registered in the matched route.
+//
+// When there is a match, the route variables can be retrieved calling
+// mux.Vars(request).
+func (r *Router) ServeHTTP(w http.ResponseWriter, req *http.Request) {
+	if !r.skipClean {
+		path := req.URL.Path
+		if r.useEncodedPath {
+			path = req.URL.EscapedPath()
+		}
+		// Clean path to canonical form and redirect.
+		if p := cleanPath(path); p != path {
+
+			// Added 3 lines (Philip Schlump) - It was dropping the query string and #whatever from query.
+			// This matches with fix in go 1.2 r.c. 4 for same problem.  Go Issue:
+			// http://code.google.com/p/go/issues/detail?id=5252
+			url := *req.URL
+			url.Path = p
+			p = url.String()
+
+			w.Header().Set("Location", p)
+			w.WriteHeader(http.StatusMovedPermanently)
+			return
+		}
+	}
+	var match RouteMatch
+	var handler http.Handler
+	if r.Match(req, &match) {
+		handler = match.Handler
+		req = setVars(req, match.Vars)
+		req = setCurrentRoute(req, match.Route)
+	}
+
+	if handler == nil && match.MatchErr == ErrMethodMismatch {
+		handler = methodNotAllowedHandler()
+	}
+
+	if handler == nil {
+		handler = http.NotFoundHandler()
+	}
+
+	if !r.KeepContext {
+		defer contextClear(req)
+	}
+
+	handler.ServeHTTP(w, req)
+}
+
+// Get returns a route registered with the given name.
+func (r *Router) Get(name string) *Route {
+	return r.getNamedRoutes()[name]
+}
+
+// GetRoute returns a route registered with the given name. This method
+// was renamed to Get() and remains here for backwards compatibility.
+func (r *Router) GetRoute(name string) *Route {
+	return r.getNamedRoutes()[name]
+}
+
+// StrictSlash defines the trailing slash behavior for new routes. The initial
+// value is false.
+//
+// When true, if the route path is "/path/", accessing "/path" will perform a redirect
+// to the former and vice versa. In other words, your application will always
+// see the path as specified in the route.
+//
+// When false, if the route path is "/path", accessing "/path/" will not match
+// this route and vice versa.
+//
+// The re-direct is a HTTP 301 (Moved Permanently). Note that when this is set for
+// routes with a non-idempotent method (e.g. POST, PUT), the subsequent re-directed
+// request will be made as a GET by most clients. Use middleware or client settings
+// to modify this behaviour as needed.
+//
+// Special case: when a route sets a path prefix using the PathPrefix() method,
+// strict slash is ignored for that route because the redirect behavior can't
+// be determined from a prefix alone. However, any subrouters created from that
+// route inherit the original StrictSlash setting.
+func (r *Router) StrictSlash(value bool) *Router {
+	r.strictSlash = value
+	return r
+}
+
+// SkipClean defines the path cleaning behaviour for new routes. The initial
+// value is false. Users should be careful about which routes are not cleaned
+//
+// When true, if the route path is "/path//to", it will remain with the double
+// slash. This is helpful if you have a route like: /fetch/http://xkcd.com/534/
+//
+// When false, the path will be cleaned, so /fetch/http://xkcd.com/534/ will
+// become /fetch/http/xkcd.com/534
+func (r *Router) SkipClean(value bool) *Router {
+	r.skipClean = value
+	return r
+}
+
+// UseEncodedPath tells the router to match the encoded original path
+// to the routes.
+// For eg. "/path/foo%2Fbar/to" will match the path "/path/{var}/to".
+//
+// If not called, the router will match the unencoded path to the routes.
+// For eg. "/path/foo%2Fbar/to" will match the path "/path/foo/bar/to"
+func (r *Router) UseEncodedPath() *Router {
+	r.useEncodedPath = true
+	return r
+}
+
+// ----------------------------------------------------------------------------
+// parentRoute
+// ----------------------------------------------------------------------------
+
+func (r *Router) getBuildScheme() string {
+	if r.parent != nil {
+		return r.parent.getBuildScheme()
+	}
+	return ""
+}
+
+// getNamedRoutes returns the map where named routes are registered.
+func (r *Router) getNamedRoutes() map[string]*Route {
+	if r.namedRoutes == nil {
+		if r.parent != nil {
+			r.namedRoutes = r.parent.getNamedRoutes()
+		} else {
+			r.namedRoutes = make(map[string]*Route)
+		}
+	}
+	return r.namedRoutes
+}
+
+// getRegexpGroup returns regexp definitions from the parent route, if any.
+func (r *Router) getRegexpGroup() *routeRegexpGroup {
+	if r.parent != nil {
+		return r.parent.getRegexpGroup()
+	}
+	return nil
+}
+
+func (r *Router) buildVars(m map[string]string) map[string]string {
+	if r.parent != nil {
+		m = r.parent.buildVars(m)
+	}
+	return m
+}
+
+// ----------------------------------------------------------------------------
+// Route factories
+// ----------------------------------------------------------------------------
+
+// NewRoute registers an empty route.
+func (r *Router) NewRoute() *Route {
+	route := &Route{parent: r, strictSlash: r.strictSlash, skipClean: r.skipClean, useEncodedPath: r.useEncodedPath}
+	r.routes = append(r.routes, route)
+	return route
+}
+
+// Handle registers a new route with a matcher for the URL path.
+// See Route.Path() and Route.Handler().
+func (r *Router) Handle(path string, handler http.Handler) *Route {
+	return r.NewRoute().Path(path).Handler(handler)
+}
+
+// HandleFunc registers a new route with a matcher for the URL path.
+// See Route.Path() and Route.HandlerFunc().
+func (r *Router) HandleFunc(path string, f func(http.ResponseWriter,
+	*http.Request)) *Route {
+	return r.NewRoute().Path(path).HandlerFunc(f)
+}
+
+// Headers registers a new route with a matcher for request header values.
+// See Route.Headers().
+func (r *Router) Headers(pairs ...string) *Route {
+	return r.NewRoute().Headers(pairs...)
+}
+
+// Host registers a new route with a matcher for the URL host.
+// See Route.Host().
+func (r *Router) Host(tpl string) *Route {
+	return r.NewRoute().Host(tpl)
+}
+
+// MatcherFunc registers a new route with a custom matcher function.
+// See Route.MatcherFunc().
+func (r *Router) MatcherFunc(f MatcherFunc) *Route {
+	return r.NewRoute().MatcherFunc(f)
+}
+
+// Methods registers a new route with a matcher for HTTP methods.
+// See Route.Methods().
+func (r *Router) Methods(methods ...string) *Route {
+	return r.NewRoute().Methods(methods...)
+}
+
+// Path registers a new route with a matcher for the URL path.
+// See Route.Path().
+func (r *Router) Path(tpl string) *Route {
+	return r.NewRoute().Path(tpl)
+}
+
+// PathPrefix registers a new route with a matcher for the URL path prefix.
+// See Route.PathPrefix().
+func (r *Router) PathPrefix(tpl string) *Route {
+	return r.NewRoute().PathPrefix(tpl)
+}
+
+// Queries registers a new route with a matcher for URL query values.
+// See Route.Queries().
+func (r *Router) Queries(pairs ...string) *Route {
+	return r.NewRoute().Queries(pairs...)
+}
+
+// Schemes registers a new route with a matcher for URL schemes.
+// See Route.Schemes().
+func (r *Router) Schemes(schemes ...string) *Route {
+	return r.NewRoute().Schemes(schemes...)
+}
+
+// BuildVarsFunc registers a new route with a custom function for modifying
+// route variables before building a URL.
+func (r *Router) BuildVarsFunc(f BuildVarsFunc) *Route {
+	return r.NewRoute().BuildVarsFunc(f)
+}
+
+// Walk walks the router and all its sub-routers, calling walkFn for each route
+// in the tree. The routes are walked in the order they were added. Sub-routers
+// are explored depth-first.
+func (r *Router) Walk(walkFn WalkFunc) error {
+	return r.walk(walkFn, []*Route{})
+}
+
+// SkipRouter is used as a return value from WalkFuncs to indicate that the
+// router that walk is about to descend down to should be skipped.
+var SkipRouter = errors.New("skip this router")
+
+// WalkFunc is the type of the function called for each route visited by Walk.
+// At every invocation, it is given the current route, and the current router,
+// and a list of ancestor routes that lead to the current route.
+type WalkFunc func(route *Route, router *Router, ancestors []*Route) error
+
+func (r *Router) walk(walkFn WalkFunc, ancestors []*Route) error {
+	for _, t := range r.routes {
+		err := walkFn(t, r, ancestors)
+		if err == SkipRouter {
+			continue
+		}
+		if err != nil {
+			return err
+		}
+		for _, sr := range t.matchers {
+			if h, ok := sr.(*Router); ok {
+				ancestors = append(ancestors, t)
+				err := h.walk(walkFn, ancestors)
+				if err != nil {
+					return err
+				}
+				ancestors = ancestors[:len(ancestors)-1]
+			}
+		}
+		if h, ok := t.handler.(*Router); ok {
+			ancestors = append(ancestors, t)
+			err := h.walk(walkFn, ancestors)
+			if err != nil {
+				return err
+			}
+			ancestors = ancestors[:len(ancestors)-1]
+		}
+	}
+	return nil
+}
+
+// ----------------------------------------------------------------------------
+// Context
+// ----------------------------------------------------------------------------
+
+// RouteMatch stores information about a matched route.
+type RouteMatch struct {
+	Route   *Route
+	Handler http.Handler
+	Vars    map[string]string
+
+	// MatchErr is set to appropriate matching error
+	// It is set to ErrMethodMismatch if there is a mismatch in
+	// the request method and route method
+	MatchErr error
+}
+
+type contextKey int
+
+const (
+	varsKey contextKey = iota
+	routeKey
+)
+
+// Vars returns the route variables for the current request, if any.
+func Vars(r *http.Request) map[string]string {
+	if rv := contextGet(r, varsKey); rv != nil {
+		return rv.(map[string]string)
+	}
+	return nil
+}
+
+// CurrentRoute returns the matched route for the current request, if any.
+// This only works when called inside the handler of the matched route
+// because the matched route is stored in the request context which is cleared
+// after the handler returns, unless the KeepContext option is set on the
+// Router.
+func CurrentRoute(r *http.Request) *Route {
+	if rv := contextGet(r, routeKey); rv != nil {
+		return rv.(*Route)
+	}
+	return nil
+}
+
+func setVars(r *http.Request, val interface{}) *http.Request {
+	return contextSet(r, varsKey, val)
+}
+
+func setCurrentRoute(r *http.Request, val interface{}) *http.Request {
+	return contextSet(r, routeKey, val)
+}
+
+// ----------------------------------------------------------------------------
+// Helpers
+// ----------------------------------------------------------------------------
+
+// cleanPath returns the canonical path for p, eliminating . and .. elements.
+// Borrowed from the net/http package.
+func cleanPath(p string) string {
+	if p == "" {
+		return "/"
+	}
+	if p[0] != '/' {
+		p = "/" + p
+	}
+	np := path.Clean(p)
+	// path.Clean removes trailing slash except for root;
+	// put the trailing slash back if necessary.
+	if p[len(p)-1] == '/' && np != "/" {
+		np += "/"
+	}
+
+	return np
+}
+
+// uniqueVars returns an error if two slices contain duplicated strings.
+func uniqueVars(s1, s2 []string) error {
+	for _, v1 := range s1 {
+		for _, v2 := range s2 {
+			if v1 == v2 {
+				return fmt.Errorf("mux: duplicated route variable %q", v2)
+			}
+		}
+	}
+	return nil
+}
+
+// checkPairs returns the count of strings passed in, and an error if
+// the count is not an even number.
+func checkPairs(pairs ...string) (int, error) {
+	length := len(pairs)
+	if length%2 != 0 {
+		return length, fmt.Errorf(
+			"mux: number of parameters must be multiple of 2, got %v", pairs)
+	}
+	return length, nil
+}
+
+// mapFromPairsToString converts variadic string parameters to a
+// string to string map.
+func mapFromPairsToString(pairs ...string) (map[string]string, error) {
+	length, err := checkPairs(pairs...)
+	if err != nil {
+		return nil, err
+	}
+	m := make(map[string]string, length/2)
+	for i := 0; i < length; i += 2 {
+		m[pairs[i]] = pairs[i+1]
+	}
+	return m, nil
+}
+
+// mapFromPairsToRegex converts variadic string parameters to a
+// string to regex map.
+func mapFromPairsToRegex(pairs ...string) (map[string]*regexp.Regexp, error) {
+	length, err := checkPairs(pairs...)
+	if err != nil {
+		return nil, err
+	}
+	m := make(map[string]*regexp.Regexp, length/2)
+	for i := 0; i < length; i += 2 {
+		regex, err := regexp.Compile(pairs[i+1])
+		if err != nil {
+			return nil, err
+		}
+		m[pairs[i]] = regex
+	}
+	return m, nil
+}
+
+// matchInArray returns true if the given string value is in the array.
+func matchInArray(arr []string, value string) bool {
+	for _, v := range arr {
+		if v == value {
+			return true
+		}
+	}
+	return false
+}
+
+// matchMapWithString returns true if the given key/value pairs exist in a given map.
+func matchMapWithString(toCheck map[string]string, toMatch map[string][]string, canonicalKey bool) bool {
+	for k, v := range toCheck {
+		// Check if key exists.
+		if canonicalKey {
+			k = http.CanonicalHeaderKey(k)
+		}
+		if values := toMatch[k]; values == nil {
+			return false
+		} else if v != "" {
+			// If value was defined as an empty string we only check that the
+			// key exists. Otherwise we also check for equality.
+			valueExists := false
+			for _, value := range values {
+				if v == value {
+					valueExists = true
+					break
+				}
+			}
+			if !valueExists {
+				return false
+			}
+		}
+	}
+	return true
+}
+
+// matchMapWithRegex returns true if the given key/value pairs exist in a given map compiled against
+// the given regex
+func matchMapWithRegex(toCheck map[string]*regexp.Regexp, toMatch map[string][]string, canonicalKey bool) bool {
+	for k, v := range toCheck {
+		// Check if key exists.
+		if canonicalKey {
+			k = http.CanonicalHeaderKey(k)
+		}
+		if values := toMatch[k]; values == nil {
+			return false
+		} else if v != nil {
+			// If value was defined as an empty string we only check that the
+			// key exists. Otherwise we also check for equality.
+			valueExists := false
+			for _, value := range values {
+				if v.MatchString(value) {
+					valueExists = true
+					break
+				}
+			}
+			if !valueExists {
+				return false
+			}
+		}
+	}
+	return true
+}
+
+// methodNotAllowed replies to the request with an HTTP status code 405.
+func methodNotAllowed(w http.ResponseWriter, r *http.Request) {
+	w.WriteHeader(http.StatusMethodNotAllowed)
+}
+
+// methodNotAllowedHandler returns a simple request handler
+// that replies to each request with a status code 405.
+func methodNotAllowedHandler() http.Handler { return http.HandlerFunc(methodNotAllowed) }
diff --git a/vendor/github.com/gorilla/mux/regexp.go b/vendor/github.com/gorilla/mux/regexp.go
new file mode 100644
index 0000000..2b57e56
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/regexp.go
@@ -0,0 +1,332 @@
+// Copyright 2012 The Gorilla Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package mux
+
+import (
+	"bytes"
+	"fmt"
+	"net/http"
+	"net/url"
+	"regexp"
+	"strconv"
+	"strings"
+)
+
+type routeRegexpOptions struct {
+	strictSlash    bool
+	useEncodedPath bool
+}
+
+type regexpType int
+
+const (
+	regexpTypePath   regexpType = 0
+	regexpTypeHost   regexpType = 1
+	regexpTypePrefix regexpType = 2
+	regexpTypeQuery  regexpType = 3
+)
+
+// newRouteRegexp parses a route template and returns a routeRegexp,
+// used to match a host, a path or a query string.
+//
+// It will extract named variables, assemble a regexp to be matched, create
+// a "reverse" template to build URLs and compile regexps to validate variable
+// values used in URL building.
+//
+// Previously we accepted only Python-like identifiers for variable
+// names ([a-zA-Z_][a-zA-Z0-9_]*), but currently the only restriction is that
+// name and pattern can't be empty, and names can't contain a colon.
+func newRouteRegexp(tpl string, typ regexpType, options routeRegexpOptions) (*routeRegexp, error) {
+	// Check if it is well-formed.
+	idxs, errBraces := braceIndices(tpl)
+	if errBraces != nil {
+		return nil, errBraces
+	}
+	// Backup the original.
+	template := tpl
+	// Now let's parse it.
+	defaultPattern := "[^/]+"
+	if typ == regexpTypeQuery {
+		defaultPattern = ".*"
+	} else if typ == regexpTypeHost {
+		defaultPattern = "[^.]+"
+	}
+	// Only match strict slash if not matching
+	if typ != regexpTypePath {
+		options.strictSlash = false
+	}
+	// Set a flag for strictSlash.
+	endSlash := false
+	if options.strictSlash && strings.HasSuffix(tpl, "/") {
+		tpl = tpl[:len(tpl)-1]
+		endSlash = true
+	}
+	varsN := make([]string, len(idxs)/2)
+	varsR := make([]*regexp.Regexp, len(idxs)/2)
+	pattern := bytes.NewBufferString("")
+	pattern.WriteByte('^')
+	reverse := bytes.NewBufferString("")
+	var end int
+	var err error
+	for i := 0; i < len(idxs); i += 2 {
+		// Set all values we are interested in.
+		raw := tpl[end:idxs[i]]
+		end = idxs[i+1]
+		parts := strings.SplitN(tpl[idxs[i]+1:end-1], ":", 2)
+		name := parts[0]
+		patt := defaultPattern
+		if len(parts) == 2 {
+			patt = parts[1]
+		}
+		// Name or pattern can't be empty.
+		if name == "" || patt == "" {
+			return nil, fmt.Errorf("mux: missing name or pattern in %q",
+				tpl[idxs[i]:end])
+		}
+		// Build the regexp pattern.
+		fmt.Fprintf(pattern, "%s(?P<%s>%s)", regexp.QuoteMeta(raw), varGroupName(i/2), patt)
+
+		// Build the reverse template.
+		fmt.Fprintf(reverse, "%s%%s", raw)
+
+		// Append variable name and compiled pattern.
+		varsN[i/2] = name
+		varsR[i/2], err = regexp.Compile(fmt.Sprintf("^%s$", patt))
+		if err != nil {
+			return nil, err
+		}
+	}
+	// Add the remaining.
+	raw := tpl[end:]
+	pattern.WriteString(regexp.QuoteMeta(raw))
+	if options.strictSlash {
+		pattern.WriteString("[/]?")
+	}
+	if typ == regexpTypeQuery {
+		// Add the default pattern if the query value is empty
+		if queryVal := strings.SplitN(template, "=", 2)[1]; queryVal == "" {
+			pattern.WriteString(defaultPattern)
+		}
+	}
+	if typ != regexpTypePrefix {
+		pattern.WriteByte('$')
+	}
+	reverse.WriteString(raw)
+	if endSlash {
+		reverse.WriteByte('/')
+	}
+	// Compile full regexp.
+	reg, errCompile := regexp.Compile(pattern.String())
+	if errCompile != nil {
+		return nil, errCompile
+	}
+
+	// Check for capturing groups which used to work in older versions
+	if reg.NumSubexp() != len(idxs)/2 {
+		panic(fmt.Sprintf("route %s contains capture groups in its regexp. ", template) +
+			"Only non-capturing groups are accepted: e.g. (?:pattern) instead of (pattern)")
+	}
+
+	// Done!
+	return &routeRegexp{
+		template:   template,
+		regexpType: typ,
+		options:    options,
+		regexp:     reg,
+		reverse:    reverse.String(),
+		varsN:      varsN,
+		varsR:      varsR,
+	}, nil
+}
+
+// routeRegexp stores a regexp to match a host or path and information to
+// collect and validate route variables.
+type routeRegexp struct {
+	// The unmodified template.
+	template string
+	// The type of match
+	regexpType regexpType
+	// Options for matching
+	options routeRegexpOptions
+	// Expanded regexp.
+	regexp *regexp.Regexp
+	// Reverse template.
+	reverse string
+	// Variable names.
+	varsN []string
+	// Variable regexps (validators).
+	varsR []*regexp.Regexp
+}
+
+// Match matches the regexp against the URL host or path.
+func (r *routeRegexp) Match(req *http.Request, match *RouteMatch) bool {
+	if r.regexpType != regexpTypeHost {
+		if r.regexpType == regexpTypeQuery {
+			return r.matchQueryString(req)
+		}
+		path := req.URL.Path
+		if r.options.useEncodedPath {
+			path = req.URL.EscapedPath()
+		}
+		return r.regexp.MatchString(path)
+	}
+
+	return r.regexp.MatchString(getHost(req))
+}
+
+// url builds a URL part using the given values.
+func (r *routeRegexp) url(values map[string]string) (string, error) {
+	urlValues := make([]interface{}, len(r.varsN))
+	for k, v := range r.varsN {
+		value, ok := values[v]
+		if !ok {
+			return "", fmt.Errorf("mux: missing route variable %q", v)
+		}
+		if r.regexpType == regexpTypeQuery {
+			value = url.QueryEscape(value)
+		}
+		urlValues[k] = value
+	}
+	rv := fmt.Sprintf(r.reverse, urlValues...)
+	if !r.regexp.MatchString(rv) {
+		// The URL is checked against the full regexp, instead of checking
+		// individual variables. This is faster but to provide a good error
+		// message, we check individual regexps if the URL doesn't match.
+		for k, v := range r.varsN {
+			if !r.varsR[k].MatchString(values[v]) {
+				return "", fmt.Errorf(
+					"mux: variable %q doesn't match, expected %q", values[v],
+					r.varsR[k].String())
+			}
+		}
+	}
+	return rv, nil
+}
+
+// getURLQuery returns a single query parameter from a request URL.
+// For a URL with foo=bar&baz=ding, we return only the relevant key
+// value pair for the routeRegexp.
+func (r *routeRegexp) getURLQuery(req *http.Request) string {
+	if r.regexpType != regexpTypeQuery {
+		return ""
+	}
+	templateKey := strings.SplitN(r.template, "=", 2)[0]
+	for key, vals := range req.URL.Query() {
+		if key == templateKey && len(vals) > 0 {
+			return key + "=" + vals[0]
+		}
+	}
+	return ""
+}
+
+func (r *routeRegexp) matchQueryString(req *http.Request) bool {
+	return r.regexp.MatchString(r.getURLQuery(req))
+}
+
+// braceIndices returns the first level curly brace indices from a string.
+// It returns an error in case of unbalanced braces.
+func braceIndices(s string) ([]int, error) {
+	var level, idx int
+	var idxs []int
+	for i := 0; i < len(s); i++ {
+		switch s[i] {
+		case '{':
+			if level++; level == 1 {
+				idx = i
+			}
+		case '}':
+			if level--; level == 0 {
+				idxs = append(idxs, idx, i+1)
+			} else if level < 0 {
+				return nil, fmt.Errorf("mux: unbalanced braces in %q", s)
+			}
+		}
+	}
+	if level != 0 {
+		return nil, fmt.Errorf("mux: unbalanced braces in %q", s)
+	}
+	return idxs, nil
+}
+
+// varGroupName builds a capturing group name for the indexed variable.
+func varGroupName(idx int) string {
+	return "v" + strconv.Itoa(idx)
+}
+
+// ----------------------------------------------------------------------------
+// routeRegexpGroup
+// ----------------------------------------------------------------------------
+
+// routeRegexpGroup groups the route matchers that carry variables.
+type routeRegexpGroup struct {
+	host    *routeRegexp
+	path    *routeRegexp
+	queries []*routeRegexp
+}
+
+// setMatch extracts the variables from the URL once a route matches.
+func (v *routeRegexpGroup) setMatch(req *http.Request, m *RouteMatch, r *Route) {
+	// Store host variables.
+	if v.host != nil {
+		host := getHost(req)
+		matches := v.host.regexp.FindStringSubmatchIndex(host)
+		if len(matches) > 0 {
+			extractVars(host, matches, v.host.varsN, m.Vars)
+		}
+	}
+	path := req.URL.Path
+	if r.useEncodedPath {
+		path = req.URL.EscapedPath()
+	}
+	// Store path variables.
+	if v.path != nil {
+		matches := v.path.regexp.FindStringSubmatchIndex(path)
+		if len(matches) > 0 {
+			extractVars(path, matches, v.path.varsN, m.Vars)
+			// Check if we should redirect.
+			if v.path.options.strictSlash {
+				p1 := strings.HasSuffix(path, "/")
+				p2 := strings.HasSuffix(v.path.template, "/")
+				if p1 != p2 {
+					u, _ := url.Parse(req.URL.String())
+					if p1 {
+						u.Path = u.Path[:len(u.Path)-1]
+					} else {
+						u.Path += "/"
+					}
+					m.Handler = http.RedirectHandler(u.String(), 301)
+				}
+			}
+		}
+	}
+	// Store query string variables.
+	for _, q := range v.queries {
+		queryURL := q.getURLQuery(req)
+		matches := q.regexp.FindStringSubmatchIndex(queryURL)
+		if len(matches) > 0 {
+			extractVars(queryURL, matches, q.varsN, m.Vars)
+		}
+	}
+}
+
+// getHost tries its best to return the request host.
+func getHost(r *http.Request) string {
+	if r.URL.IsAbs() {
+		return r.URL.Host
+	}
+	host := r.Host
+	// Slice off any port information.
+	if i := strings.Index(host, ":"); i != -1 {
+		host = host[:i]
+	}
+	return host
+
+}
+
+func extractVars(input string, matches []int, names []string, output map[string]string) {
+	for i, name := range names {
+		output[name] = input[matches[2*i+2]:matches[2*i+3]]
+	}
+}
diff --git a/vendor/github.com/gorilla/mux/route.go b/vendor/github.com/gorilla/mux/route.go
new file mode 100644
index 0000000..4ce098d
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/route.go
@@ -0,0 +1,761 @@
+// Copyright 2012 The Gorilla Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package mux
+
+import (
+	"errors"
+	"fmt"
+	"net/http"
+	"net/url"
+	"regexp"
+	"strings"
+)
+
+// Route stores information to match a request and build URLs.
+type Route struct {
+	// Parent where the route was registered (a Router).
+	parent parentRoute
+	// Request handler for the route.
+	handler http.Handler
+	// List of matchers.
+	matchers []matcher
+	// Manager for the variables from host and path.
+	regexp *routeRegexpGroup
+	// If true, when the path pattern is "/path/", accessing "/path" will
+	// redirect to the former and vice versa.
+	strictSlash bool
+	// If true, when the path pattern is "/path//to", accessing "/path//to"
+	// will not redirect
+	skipClean bool
+	// If true, "/path/foo%2Fbar/to" will match the path "/path/{var}/to"
+	useEncodedPath bool
+	// The scheme used when building URLs.
+	buildScheme string
+	// If true, this route never matches: it is only used to build URLs.
+	buildOnly bool
+	// The name used to build URLs.
+	name string
+	// Error resulted from building a route.
+	err error
+
+	buildVarsFunc BuildVarsFunc
+}
+
+func (r *Route) SkipClean() bool {
+	return r.skipClean
+}
+
+// Match matches the route against the request.
+func (r *Route) Match(req *http.Request, match *RouteMatch) bool {
+	if r.buildOnly || r.err != nil {
+		return false
+	}
+
+	var matchErr error
+
+	// Match everything.
+	for _, m := range r.matchers {
+		if matched := m.Match(req, match); !matched {
+			if _, ok := m.(methodMatcher); ok {
+				matchErr = ErrMethodMismatch
+				continue
+			}
+			matchErr = nil
+			return false
+		}
+	}
+
+	if matchErr != nil {
+		match.MatchErr = matchErr
+		return false
+	}
+
+	if match.MatchErr == ErrMethodMismatch {
+		// We found a route which matches request method, clear MatchErr
+		match.MatchErr = nil
+		// Then override the mis-matched handler
+		match.Handler = r.handler
+	}
+
+	// Yay, we have a match. Let's collect some info about it.
+	if match.Route == nil {
+		match.Route = r
+	}
+	if match.Handler == nil {
+		match.Handler = r.handler
+	}
+	if match.Vars == nil {
+		match.Vars = make(map[string]string)
+	}
+
+	// Set variables.
+	if r.regexp != nil {
+		r.regexp.setMatch(req, match, r)
+	}
+	return true
+}
+
+// ----------------------------------------------------------------------------
+// Route attributes
+// ----------------------------------------------------------------------------
+
+// GetError returns an error resulted from building the route, if any.
+func (r *Route) GetError() error {
+	return r.err
+}
+
+// BuildOnly sets the route to never match: it is only used to build URLs.
+func (r *Route) BuildOnly() *Route {
+	r.buildOnly = true
+	return r
+}
+
+// Handler --------------------------------------------------------------------
+
+// Handler sets a handler for the route.
+func (r *Route) Handler(handler http.Handler) *Route {
+	if r.err == nil {
+		r.handler = handler
+	}
+	return r
+}
+
+// HandlerFunc sets a handler function for the route.
+func (r *Route) HandlerFunc(f func(http.ResponseWriter, *http.Request)) *Route {
+	return r.Handler(http.HandlerFunc(f))
+}
+
+// GetHandler returns the handler for the route, if any.
+func (r *Route) GetHandler() http.Handler {
+	return r.handler
+}
+
+// Name -----------------------------------------------------------------------
+
+// Name sets the name for the route, used to build URLs.
+// If the name was registered already it will be overwritten.
+func (r *Route) Name(name string) *Route {
+	if r.name != "" {
+		r.err = fmt.Errorf("mux: route already has name %q, can't set %q",
+			r.name, name)
+	}
+	if r.err == nil {
+		r.name = name
+		r.getNamedRoutes()[name] = r
+	}
+	return r
+}
+
+// GetName returns the name for the route, if any.
+func (r *Route) GetName() string {
+	return r.name
+}
+
+// ----------------------------------------------------------------------------
+// Matchers
+// ----------------------------------------------------------------------------
+
+// matcher types try to match a request.
+type matcher interface {
+	Match(*http.Request, *RouteMatch) bool
+}
+
+// addMatcher adds a matcher to the route.
+func (r *Route) addMatcher(m matcher) *Route {
+	if r.err == nil {
+		r.matchers = append(r.matchers, m)
+	}
+	return r
+}
+
+// addRegexpMatcher adds a host or path matcher and builder to a route.
+func (r *Route) addRegexpMatcher(tpl string, typ regexpType) error {
+	if r.err != nil {
+		return r.err
+	}
+	r.regexp = r.getRegexpGroup()
+	if typ == regexpTypePath || typ == regexpTypePrefix {
+		if len(tpl) > 0 && tpl[0] != '/' {
+			return fmt.Errorf("mux: path must start with a slash, got %q", tpl)
+		}
+		if r.regexp.path != nil {
+			tpl = strings.TrimRight(r.regexp.path.template, "/") + tpl
+		}
+	}
+	rr, err := newRouteRegexp(tpl, typ, routeRegexpOptions{
+		strictSlash:    r.strictSlash,
+		useEncodedPath: r.useEncodedPath,
+	})
+	if err != nil {
+		return err
+	}
+	for _, q := range r.regexp.queries {
+		if err = uniqueVars(rr.varsN, q.varsN); err != nil {
+			return err
+		}
+	}
+	if typ == regexpTypeHost {
+		if r.regexp.path != nil {
+			if err = uniqueVars(rr.varsN, r.regexp.path.varsN); err != nil {
+				return err
+			}
+		}
+		r.regexp.host = rr
+	} else {
+		if r.regexp.host != nil {
+			if err = uniqueVars(rr.varsN, r.regexp.host.varsN); err != nil {
+				return err
+			}
+		}
+		if typ == regexpTypeQuery {
+			r.regexp.queries = append(r.regexp.queries, rr)
+		} else {
+			r.regexp.path = rr
+		}
+	}
+	r.addMatcher(rr)
+	return nil
+}
+
+// Headers --------------------------------------------------------------------
+
+// headerMatcher matches the request against header values.
+type headerMatcher map[string]string
+
+func (m headerMatcher) Match(r *http.Request, match *RouteMatch) bool {
+	return matchMapWithString(m, r.Header, true)
+}
+
+// Headers adds a matcher for request header values.
+// It accepts a sequence of key/value pairs to be matched. For example:
+//
+//     r := mux.NewRouter()
+//     r.Headers("Content-Type", "application/json",
+//               "X-Requested-With", "XMLHttpRequest")
+//
+// The above route will only match if both request header values match.
+// If the value is an empty string, it will match any value if the key is set.
+func (r *Route) Headers(pairs ...string) *Route {
+	if r.err == nil {
+		var headers map[string]string
+		headers, r.err = mapFromPairsToString(pairs...)
+		return r.addMatcher(headerMatcher(headers))
+	}
+	return r
+}
+
+// headerRegexMatcher matches the request against the route given a regex for the header
+type headerRegexMatcher map[string]*regexp.Regexp
+
+func (m headerRegexMatcher) Match(r *http.Request, match *RouteMatch) bool {
+	return matchMapWithRegex(m, r.Header, true)
+}
+
+// HeadersRegexp accepts a sequence of key/value pairs, where the value has regex
+// support. For example:
+//
+//     r := mux.NewRouter()
+//     r.HeadersRegexp("Content-Type", "application/(text|json)",
+//               "X-Requested-With", "XMLHttpRequest")
+//
+// The above route will only match if both the request header matches both regular expressions.
+// If the value is an empty string, it will match any value if the key is set.
+// Use the start and end of string anchors (^ and $) to match an exact value.
+func (r *Route) HeadersRegexp(pairs ...string) *Route {
+	if r.err == nil {
+		var headers map[string]*regexp.Regexp
+		headers, r.err = mapFromPairsToRegex(pairs...)
+		return r.addMatcher(headerRegexMatcher(headers))
+	}
+	return r
+}
+
+// Host -----------------------------------------------------------------------
+
+// Host adds a matcher for the URL host.
+// It accepts a template with zero or more URL variables enclosed by {}.
+// Variables can define an optional regexp pattern to be matched:
+//
+// - {name} matches anything until the next dot.
+//
+// - {name:pattern} matches the given regexp pattern.
+//
+// For example:
+//
+//     r := mux.NewRouter()
+//     r.Host("www.example.com")
+//     r.Host("{subdomain}.domain.com")
+//     r.Host("{subdomain:[a-z]+}.domain.com")
+//
+// Variable names must be unique in a given route. They can be retrieved
+// calling mux.Vars(request).
+func (r *Route) Host(tpl string) *Route {
+	r.err = r.addRegexpMatcher(tpl, regexpTypeHost)
+	return r
+}
+
+// MatcherFunc ----------------------------------------------------------------
+
+// MatcherFunc is the function signature used by custom matchers.
+type MatcherFunc func(*http.Request, *RouteMatch) bool
+
+// Match returns the match for a given request.
+func (m MatcherFunc) Match(r *http.Request, match *RouteMatch) bool {
+	return m(r, match)
+}
+
+// MatcherFunc adds a custom function to be used as request matcher.
+func (r *Route) MatcherFunc(f MatcherFunc) *Route {
+	return r.addMatcher(f)
+}
+
+// Methods --------------------------------------------------------------------
+
+// methodMatcher matches the request against HTTP methods.
+type methodMatcher []string
+
+func (m methodMatcher) Match(r *http.Request, match *RouteMatch) bool {
+	return matchInArray(m, r.Method)
+}
+
+// Methods adds a matcher for HTTP methods.
+// It accepts a sequence of one or more methods to be matched, e.g.:
+// "GET", "POST", "PUT".
+func (r *Route) Methods(methods ...string) *Route {
+	for k, v := range methods {
+		methods[k] = strings.ToUpper(v)
+	}
+	return r.addMatcher(methodMatcher(methods))
+}
+
+// Path -----------------------------------------------------------------------
+
+// Path adds a matcher for the URL path.
+// It accepts a template with zero or more URL variables enclosed by {}. The
+// template must start with a "/".
+// Variables can define an optional regexp pattern to be matched:
+//
+// - {name} matches anything until the next slash.
+//
+// - {name:pattern} matches the given regexp pattern.
+//
+// For example:
+//
+//     r := mux.NewRouter()
+//     r.Path("/products/").Handler(ProductsHandler)
+//     r.Path("/products/{key}").Handler(ProductsHandler)
+//     r.Path("/articles/{category}/{id:[0-9]+}").
+//       Handler(ArticleHandler)
+//
+// Variable names must be unique in a given route. They can be retrieved
+// calling mux.Vars(request).
+func (r *Route) Path(tpl string) *Route {
+	r.err = r.addRegexpMatcher(tpl, regexpTypePath)
+	return r
+}
+
+// PathPrefix -----------------------------------------------------------------
+
+// PathPrefix adds a matcher for the URL path prefix. This matches if the given
+// template is a prefix of the full URL path. See Route.Path() for details on
+// the tpl argument.
+//
+// Note that it does not treat slashes specially ("/foobar/" will be matched by
+// the prefix "/foo") so you may want to use a trailing slash here.
+//
+// Also note that the setting of Router.StrictSlash() has no effect on routes
+// with a PathPrefix matcher.
+func (r *Route) PathPrefix(tpl string) *Route {
+	r.err = r.addRegexpMatcher(tpl, regexpTypePrefix)
+	return r
+}
+
+// Query ----------------------------------------------------------------------
+
+// Queries adds a matcher for URL query values.
+// It accepts a sequence of key/value pairs. Values may define variables.
+// For example:
+//
+//     r := mux.NewRouter()
+//     r.Queries("foo", "bar", "id", "{id:[0-9]+}")
+//
+// The above route will only match if the URL contains the defined queries
+// values, e.g.: ?foo=bar&id=42.
+//
+// It the value is an empty string, it will match any value if the key is set.
+//
+// Variables can define an optional regexp pattern to be matched:
+//
+// - {name} matches anything until the next slash.
+//
+// - {name:pattern} matches the given regexp pattern.
+func (r *Route) Queries(pairs ...string) *Route {
+	length := len(pairs)
+	if length%2 != 0 {
+		r.err = fmt.Errorf(
+			"mux: number of parameters must be multiple of 2, got %v", pairs)
+		return nil
+	}
+	for i := 0; i < length; i += 2 {
+		if r.err = r.addRegexpMatcher(pairs[i]+"="+pairs[i+1], regexpTypeQuery); r.err != nil {
+			return r
+		}
+	}
+
+	return r
+}
+
+// Schemes --------------------------------------------------------------------
+
+// schemeMatcher matches the request against URL schemes.
+type schemeMatcher []string
+
+func (m schemeMatcher) Match(r *http.Request, match *RouteMatch) bool {
+	return matchInArray(m, r.URL.Scheme)
+}
+
+// Schemes adds a matcher for URL schemes.
+// It accepts a sequence of schemes to be matched, e.g.: "http", "https".
+func (r *Route) Schemes(schemes ...string) *Route {
+	for k, v := range schemes {
+		schemes[k] = strings.ToLower(v)
+	}
+	if r.buildScheme == "" && len(schemes) > 0 {
+		r.buildScheme = schemes[0]
+	}
+	return r.addMatcher(schemeMatcher(schemes))
+}
+
+// BuildVarsFunc --------------------------------------------------------------
+
+// BuildVarsFunc is the function signature used by custom build variable
+// functions (which can modify route variables before a route's URL is built).
+type BuildVarsFunc func(map[string]string) map[string]string
+
+// BuildVarsFunc adds a custom function to be used to modify build variables
+// before a route's URL is built.
+func (r *Route) BuildVarsFunc(f BuildVarsFunc) *Route {
+	r.buildVarsFunc = f
+	return r
+}
+
+// Subrouter ------------------------------------------------------------------
+
+// Subrouter creates a subrouter for the route.
+//
+// It will test the inner routes only if the parent route matched. For example:
+//
+//     r := mux.NewRouter()
+//     s := r.Host("www.example.com").Subrouter()
+//     s.HandleFunc("/products/", ProductsHandler)
+//     s.HandleFunc("/products/{key}", ProductHandler)
+//     s.HandleFunc("/articles/{category}/{id:[0-9]+}"), ArticleHandler)
+//
+// Here, the routes registered in the subrouter won't be tested if the host
+// doesn't match.
+func (r *Route) Subrouter() *Router {
+	router := &Router{parent: r, strictSlash: r.strictSlash}
+	r.addMatcher(router)
+	return router
+}
+
+// ----------------------------------------------------------------------------
+// URL building
+// ----------------------------------------------------------------------------
+
+// URL builds a URL for the route.
+//
+// It accepts a sequence of key/value pairs for the route variables. For
+// example, given this route:
+//
+//     r := mux.NewRouter()
+//     r.HandleFunc("/articles/{category}/{id:[0-9]+}", ArticleHandler).
+//       Name("article")
+//
+// ...a URL for it can be built using:
+//
+//     url, err := r.Get("article").URL("category", "technology", "id", "42")
+//
+// ...which will return an url.URL with the following path:
+//
+//     "/articles/technology/42"
+//
+// This also works for host variables:
+//
+//     r := mux.NewRouter()
+//     r.Host("{subdomain}.domain.com").
+//       HandleFunc("/articles/{category}/{id:[0-9]+}", ArticleHandler).
+//       Name("article")
+//
+//     // url.String() will be "http://news.domain.com/articles/technology/42"
+//     url, err := r.Get("article").URL("subdomain", "news",
+//                                      "category", "technology",
+//                                      "id", "42")
+//
+// All variables defined in the route are required, and their values must
+// conform to the corresponding patterns.
+func (r *Route) URL(pairs ...string) (*url.URL, error) {
+	if r.err != nil {
+		return nil, r.err
+	}
+	if r.regexp == nil {
+		return nil, errors.New("mux: route doesn't have a host or path")
+	}
+	values, err := r.prepareVars(pairs...)
+	if err != nil {
+		return nil, err
+	}
+	var scheme, host, path string
+	queries := make([]string, 0, len(r.regexp.queries))
+	if r.regexp.host != nil {
+		if host, err = r.regexp.host.url(values); err != nil {
+			return nil, err
+		}
+		scheme = "http"
+		if s := r.getBuildScheme(); s != "" {
+			scheme = s
+		}
+	}
+	if r.regexp.path != nil {
+		if path, err = r.regexp.path.url(values); err != nil {
+			return nil, err
+		}
+	}
+	for _, q := range r.regexp.queries {
+		var query string
+		if query, err = q.url(values); err != nil {
+			return nil, err
+		}
+		queries = append(queries, query)
+	}
+	return &url.URL{
+		Scheme:   scheme,
+		Host:     host,
+		Path:     path,
+		RawQuery: strings.Join(queries, "&"),
+	}, nil
+}
+
+// URLHost builds the host part of the URL for a route. See Route.URL().
+//
+// The route must have a host defined.
+func (r *Route) URLHost(pairs ...string) (*url.URL, error) {
+	if r.err != nil {
+		return nil, r.err
+	}
+	if r.regexp == nil || r.regexp.host == nil {
+		return nil, errors.New("mux: route doesn't have a host")
+	}
+	values, err := r.prepareVars(pairs...)
+	if err != nil {
+		return nil, err
+	}
+	host, err := r.regexp.host.url(values)
+	if err != nil {
+		return nil, err
+	}
+	u := &url.URL{
+		Scheme: "http",
+		Host:   host,
+	}
+	if s := r.getBuildScheme(); s != "" {
+		u.Scheme = s
+	}
+	return u, nil
+}
+
+// URLPath builds the path part of the URL for a route. See Route.URL().
+//
+// The route must have a path defined.
+func (r *Route) URLPath(pairs ...string) (*url.URL, error) {
+	if r.err != nil {
+		return nil, r.err
+	}
+	if r.regexp == nil || r.regexp.path == nil {
+		return nil, errors.New("mux: route doesn't have a path")
+	}
+	values, err := r.prepareVars(pairs...)
+	if err != nil {
+		return nil, err
+	}
+	path, err := r.regexp.path.url(values)
+	if err != nil {
+		return nil, err
+	}
+	return &url.URL{
+		Path: path,
+	}, nil
+}
+
+// GetPathTemplate returns the template used to build the
+// route match.
+// This is useful for building simple REST API documentation and for instrumentation
+// against third-party services.
+// An error will be returned if the route does not define a path.
+func (r *Route) GetPathTemplate() (string, error) {
+	if r.err != nil {
+		return "", r.err
+	}
+	if r.regexp == nil || r.regexp.path == nil {
+		return "", errors.New("mux: route doesn't have a path")
+	}
+	return r.regexp.path.template, nil
+}
+
+// GetPathRegexp returns the expanded regular expression used to match route path.
+// This is useful for building simple REST API documentation and for instrumentation
+// against third-party services.
+// An error will be returned if the route does not define a path.
+func (r *Route) GetPathRegexp() (string, error) {
+	if r.err != nil {
+		return "", r.err
+	}
+	if r.regexp == nil || r.regexp.path == nil {
+		return "", errors.New("mux: route does not have a path")
+	}
+	return r.regexp.path.regexp.String(), nil
+}
+
+// GetQueriesRegexp returns the expanded regular expressions used to match the
+// route queries.
+// This is useful for building simple REST API documentation and for instrumentation
+// against third-party services.
+// An empty list will be returned if the route does not have queries.
+func (r *Route) GetQueriesRegexp() ([]string, error) {
+	if r.err != nil {
+		return nil, r.err
+	}
+	if r.regexp == nil || r.regexp.queries == nil {
+		return nil, errors.New("mux: route doesn't have queries")
+	}
+	var queries []string
+	for _, query := range r.regexp.queries {
+		queries = append(queries, query.regexp.String())
+	}
+	return queries, nil
+}
+
+// GetQueriesTemplates returns the templates used to build the
+// query matching.
+// This is useful for building simple REST API documentation and for instrumentation
+// against third-party services.
+// An empty list will be returned if the route does not define queries.
+func (r *Route) GetQueriesTemplates() ([]string, error) {
+	if r.err != nil {
+		return nil, r.err
+	}
+	if r.regexp == nil || r.regexp.queries == nil {
+		return nil, errors.New("mux: route doesn't have queries")
+	}
+	var queries []string
+	for _, query := range r.regexp.queries {
+		queries = append(queries, query.template)
+	}
+	return queries, nil
+}
+
+// GetMethods returns the methods the route matches against
+// This is useful for building simple REST API documentation and for instrumentation
+// against third-party services.
+// An empty list will be returned if route does not have methods.
+func (r *Route) GetMethods() ([]string, error) {
+	if r.err != nil {
+		return nil, r.err
+	}
+	for _, m := range r.matchers {
+		if methods, ok := m.(methodMatcher); ok {
+			return []string(methods), nil
+		}
+	}
+	return nil, nil
+}
+
+// GetHostTemplate returns the template used to build the
+// route match.
+// This is useful for building simple REST API documentation and for instrumentation
+// against third-party services.
+// An error will be returned if the route does not define a host.
+func (r *Route) GetHostTemplate() (string, error) {
+	if r.err != nil {
+		return "", r.err
+	}
+	if r.regexp == nil || r.regexp.host == nil {
+		return "", errors.New("mux: route doesn't have a host")
+	}
+	return r.regexp.host.template, nil
+}
+
+// prepareVars converts the route variable pairs into a map. If the route has a
+// BuildVarsFunc, it is invoked.
+func (r *Route) prepareVars(pairs ...string) (map[string]string, error) {
+	m, err := mapFromPairsToString(pairs...)
+	if err != nil {
+		return nil, err
+	}
+	return r.buildVars(m), nil
+}
+
+func (r *Route) buildVars(m map[string]string) map[string]string {
+	if r.parent != nil {
+		m = r.parent.buildVars(m)
+	}
+	if r.buildVarsFunc != nil {
+		m = r.buildVarsFunc(m)
+	}
+	return m
+}
+
+// ----------------------------------------------------------------------------
+// parentRoute
+// ----------------------------------------------------------------------------
+
+// parentRoute allows routes to know about parent host and path definitions.
+type parentRoute interface {
+	getBuildScheme() string
+	getNamedRoutes() map[string]*Route
+	getRegexpGroup() *routeRegexpGroup
+	buildVars(map[string]string) map[string]string
+}
+
+func (r *Route) getBuildScheme() string {
+	if r.buildScheme != "" {
+		return r.buildScheme
+	}
+	if r.parent != nil {
+		return r.parent.getBuildScheme()
+	}
+	return ""
+}
+
+// getNamedRoutes returns the map where named routes are registered.
+func (r *Route) getNamedRoutes() map[string]*Route {
+	if r.parent == nil {
+		// During tests router is not always set.
+		r.parent = NewRouter()
+	}
+	return r.parent.getNamedRoutes()
+}
+
+// getRegexpGroup returns regexp definitions from this route.
+func (r *Route) getRegexpGroup() *routeRegexpGroup {
+	if r.regexp == nil {
+		if r.parent == nil {
+			// During tests router is not always set.
+			r.parent = NewRouter()
+		}
+		regexp := r.parent.getRegexpGroup()
+		if regexp == nil {
+			r.regexp = new(routeRegexpGroup)
+		} else {
+			// Copy.
+			r.regexp = &routeRegexpGroup{
+				host:    regexp.host,
+				path:    regexp.path,
+				queries: regexp.queries,
+			}
+		}
+	}
+	return r.regexp
+}
diff --git a/vendor/github.com/gorilla/mux/test_helpers.go b/vendor/github.com/gorilla/mux/test_helpers.go
new file mode 100644
index 0000000..8b2c4a4
--- /dev/null
+++ b/vendor/github.com/gorilla/mux/test_helpers.go
@@ -0,0 +1,18 @@
+// Copyright 2012 The Gorilla Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package mux
+
+import "net/http"
+
+// SetURLVars sets the URL variables for the given request, to be accessed via
+// mux.Vars for testing route behaviour.
+//
+// This API should only be used for testing purposes; it provides a way to
+// inject variables into the request context. Alternatively, URL variables
+// can be set by making a route that captures the required variables,
+// starting a server and sending the request to that server.
+func SetURLVars(r *http.Request, val map[string]string) *http.Request {
+	return setVars(r, val)
+}
diff --git a/vendor/github.com/justinas/alice/.travis.yml b/vendor/github.com/justinas/alice/.travis.yml
new file mode 100644
index 0000000..dc6bea6
--- /dev/null
+++ b/vendor/github.com/justinas/alice/.travis.yml
@@ -0,0 +1,17 @@
+language: go
+
+matrix:
+  include:
+    - go: 1.0.x
+    - go: 1.1.x
+    - go: 1.2.x
+    - go: 1.3.x
+    - go: 1.4.x
+    - go: 1.5.x
+    - go: 1.6.x
+    - go: 1.7.x
+    - go: 1.8.x
+    - go: 1.9.x
+    - go: tip
+  allow_failures:
+    - go: tip
diff --git a/vendor/github.com/justinas/alice/LICENSE b/vendor/github.com/justinas/alice/LICENSE
new file mode 100644
index 0000000..0d0d352
--- /dev/null
+++ b/vendor/github.com/justinas/alice/LICENSE
@@ -0,0 +1,20 @@
+The MIT License (MIT)
+
+Copyright (c) 2014 Justinas Stankevicius
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of
+this software and associated documentation files (the "Software"), to deal in
+the Software without restriction, including without limitation the rights to
+use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
+the Software, and to permit persons to whom the Software is furnished to do so,
+subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
+FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
+COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
diff --git a/vendor/github.com/justinas/alice/README.md b/vendor/github.com/justinas/alice/README.md
new file mode 100644
index 0000000..e4f9157
--- /dev/null
+++ b/vendor/github.com/justinas/alice/README.md
@@ -0,0 +1,98 @@
+# Alice 
+
+[![GoDoc](https://godoc.org/github.com/golang/gddo?status.svg)](http://godoc.org/github.com/justinas/alice)
+[![Build Status](https://travis-ci.org/justinas/alice.svg?branch=master)](https://travis-ci.org/justinas/alice)
+[![Coverage](http://gocover.io/_badge/github.com/justinas/alice)](http://gocover.io/github.com/justinas/alice)
+
+Alice provides a convenient way to chain 
+your HTTP middleware functions and the app handler.
+
+In short, it transforms
+
+```go
+Middleware1(Middleware2(Middleware3(App)))
+```
+
+to
+
+```go
+alice.New(Middleware1, Middleware2, Middleware3).Then(App)
+```
+
+### Why?
+
+None of the other middleware chaining solutions
+behaves exactly like Alice.
+Alice is as minimal as it gets:
+in essence, it's just a for loop that does the wrapping for you.
+
+Check out [this blog post](http://justinas.org/alice-painless-middleware-chaining-for-go/)
+for explanation how Alice is different from other chaining solutions.
+
+### Usage
+
+Your middleware constructors should have the form of
+
+```go
+func (http.Handler) http.Handler
+```
+
+Some middleware provide this out of the box.
+For ones that don't, it's trivial to write one yourself.
+
+```go
+func myStripPrefix(h http.Handler) http.Handler {
+    return http.StripPrefix("/old", h)
+}
+```
+
+This complete example shows the full power of Alice.
+
+```go
+package main
+
+import (
+    "net/http"
+    "time"
+
+    "github.com/throttled/throttled"
+    "github.com/justinas/alice"
+    "github.com/justinas/nosurf"
+)
+
+func timeoutHandler(h http.Handler) http.Handler {
+    return http.TimeoutHandler(h, 1*time.Second, "timed out")
+}
+
+func myApp(w http.ResponseWriter, r *http.Request) {
+    w.Write([]byte("Hello world!"))
+}
+
+func main() {
+    th := throttled.Interval(throttled.PerSec(10), 1, &throttled.VaryBy{Path: true}, 50)
+    myHandler := http.HandlerFunc(myApp)
+
+    chain := alice.New(th.Throttle, timeoutHandler, nosurf.NewPure).Then(myHandler)
+    http.ListenAndServe(":8000", chain)
+}
+```
+
+Here, the request will pass [throttled](https://github.com/PuerkitoBio/throttled) first,
+then an http.TimeoutHandler we've set up,
+then [nosurf](https://github.com/justinas/nosurf)
+and will finally reach our handler.
+
+Note that Alice makes **no guarantees** for
+how one or another piece of  middleware will behave.
+Once it passes the execution to the outer layer of middleware,
+it has no saying in whether middleware will execute the inner handlers.
+This is intentional behavior.
+
+Alice works with Go 1.0 and higher.
+
+### Contributing
+
+0. Find an issue that bugs you / open a new one.
+1. Discuss.
+2. Branch off, commit, test.
+3. Make a pull request / attach the commits to the issue.
diff --git a/vendor/github.com/justinas/alice/chain.go b/vendor/github.com/justinas/alice/chain.go
new file mode 100644
index 0000000..da0e2b5
--- /dev/null
+++ b/vendor/github.com/justinas/alice/chain.go
@@ -0,0 +1,112 @@
+// Package alice provides a convenient way to chain http handlers.
+package alice
+
+import "net/http"
+
+// A constructor for a piece of middleware.
+// Some middleware use this constructor out of the box,
+// so in most cases you can just pass somepackage.New
+type Constructor func(http.Handler) http.Handler
+
+// Chain acts as a list of http.Handler constructors.
+// Chain is effectively immutable:
+// once created, it will always hold
+// the same set of constructors in the same order.
+type Chain struct {
+	constructors []Constructor
+}
+
+// New creates a new chain,
+// memorizing the given list of middleware constructors.
+// New serves no other function,
+// constructors are only called upon a call to Then().
+func New(constructors ...Constructor) Chain {
+	return Chain{append(([]Constructor)(nil), constructors...)}
+}
+
+// Then chains the middleware and returns the final http.Handler.
+//     New(m1, m2, m3).Then(h)
+// is equivalent to:
+//     m1(m2(m3(h)))
+// When the request comes in, it will be passed to m1, then m2, then m3
+// and finally, the given handler
+// (assuming every middleware calls the following one).
+//
+// A chain can be safely reused by calling Then() several times.
+//     stdStack := alice.New(ratelimitHandler, csrfHandler)
+//     indexPipe = stdStack.Then(indexHandler)
+//     authPipe = stdStack.Then(authHandler)
+// Note that constructors are called on every call to Then()
+// and thus several instances of the same middleware will be created
+// when a chain is reused in this way.
+// For proper middleware, this should cause no problems.
+//
+// Then() treats nil as http.DefaultServeMux.
+func (c Chain) Then(h http.Handler) http.Handler {
+	if h == nil {
+		h = http.DefaultServeMux
+	}
+
+	for i := range c.constructors {
+		h = c.constructors[len(c.constructors)-1-i](h)
+	}
+
+	return h
+}
+
+// ThenFunc works identically to Then, but takes
+// a HandlerFunc instead of a Handler.
+//
+// The following two statements are equivalent:
+//     c.Then(http.HandlerFunc(fn))
+//     c.ThenFunc(fn)
+//
+// ThenFunc provides all the guarantees of Then.
+func (c Chain) ThenFunc(fn http.HandlerFunc) http.Handler {
+	if fn == nil {
+		return c.Then(nil)
+	}
+	return c.Then(fn)
+}
+
+// Append extends a chain, adding the specified constructors
+// as the last ones in the request flow.
+//
+// Append returns a new chain, leaving the original one untouched.
+//
+//     stdChain := alice.New(m1, m2)
+//     extChain := stdChain.Append(m3, m4)
+//     // requests in stdChain go m1 -> m2
+//     // requests in extChain go m1 -> m2 -> m3 -> m4
+func (c Chain) Append(constructors ...Constructor) Chain {
+	newCons := make([]Constructor, 0, len(c.constructors)+len(constructors))
+	newCons = append(newCons, c.constructors...)
+	newCons = append(newCons, constructors...)
+
+	return Chain{newCons}
+}
+
+// Extend extends a chain by adding the specified chain
+// as the last one in the request flow.
+//
+// Extend returns a new chain, leaving the original one untouched.
+//
+//     stdChain := alice.New(m1, m2)
+//     ext1Chain := alice.New(m3, m4)
+//     ext2Chain := stdChain.Extend(ext1Chain)
+//     // requests in stdChain go  m1 -> m2
+//     // requests in ext1Chain go m3 -> m4
+//     // requests in ext2Chain go m1 -> m2 -> m3 -> m4
+//
+// Another example:
+//  aHtmlAfterNosurf := alice.New(m2)
+// 	aHtml := alice.New(m1, func(h http.Handler) http.Handler {
+// 		csrf := nosurf.New(h)
+// 		csrf.SetFailureHandler(aHtmlAfterNosurf.ThenFunc(csrfFail))
+// 		return csrf
+// 	}).Extend(aHtmlAfterNosurf)
+//		// requests to aHtml hitting nosurfs success handler go m1 -> nosurf -> m2 -> target-handler
+//		// requests to aHtml hitting nosurfs failure handler go m1 -> nosurf -> m2 -> csrfFail
+func (c Chain) Extend(chain Chain) Chain {
+	return c.Append(chain.constructors...)
+}
diff --git a/vendor/github.com/kr/logfmt/.gitignore b/vendor/github.com/kr/logfmt/.gitignore
new file mode 100644
index 0000000..8e524f6
--- /dev/null
+++ b/vendor/github.com/kr/logfmt/.gitignore
@@ -0,0 +1,3 @@
+*.test
+*.swp
+*.prof
diff --git a/vendor/github.com/kr/logfmt/Readme b/vendor/github.com/kr/logfmt/Readme
new file mode 100644
index 0000000..1865a11
--- /dev/null
+++ b/vendor/github.com/kr/logfmt/Readme
@@ -0,0 +1,12 @@
+Go package for parsing (and, eventually, generating)
+log lines in the logfmt style.
+
+See http://godoc.org/github.com/kr/logfmt for format, and other documentation and examples.
+
+Copyright (C) 2013 Keith Rarick, Blake Mizerany
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
diff --git a/vendor/github.com/kr/logfmt/decode.go b/vendor/github.com/kr/logfmt/decode.go
new file mode 100644
index 0000000..1397fb7
--- /dev/null
+++ b/vendor/github.com/kr/logfmt/decode.go
@@ -0,0 +1,184 @@
+// Package implements the decoding of logfmt key-value pairs.
+//
+// Example logfmt message:
+//
+//	foo=bar a=14 baz="hello kitty" cool%story=bro f %^asdf
+//
+// Example result in JSON:
+//
+//	{ "foo": "bar", "a": 14, "baz": "hello kitty", "cool%story": "bro", "f": true, "%^asdf": true }
+//
+// EBNFish:
+//
+// 	ident_byte = any byte greater than ' ', excluding '=' and '"'
+// 	string_byte = any byte excluding '"' and '\'
+// 	garbage = !ident_byte
+// 	ident = ident_byte, { ident byte }
+// 	key = ident
+// 	value = ident | '"', { string_byte | '\', '"' }, '"'
+// 	pair = key, '=', value | key, '=' | key
+// 	message = { garbage, pair }, garbage
+package logfmt
+
+import (
+	"reflect"
+	"strconv"
+	"strings"
+	"time"
+)
+
+// Handler is the interface implemented by objects that accept logfmt
+// key-value pairs. HandleLogfmt must copy the logfmt data if it
+// wishes to retain the data after returning.
+type Handler interface {
+	HandleLogfmt(key, val []byte) error
+}
+
+// The HandlerFunc type is an adapter to allow the use of ordinary functions as
+// logfmt handlers. If f is a function with the appropriate signature,
+// HandlerFunc(f) is a Handler object that calls f.
+type HandlerFunc func(key, val []byte) error
+
+func (f HandlerFunc) HandleLogfmt(key, val []byte) error {
+	return f(key, val)
+}
+
+// Unmarshal parses the logfmt encoding data and stores the result in the value
+// pointed to by v. If v is an Handler, HandleLogfmt will be called for each
+// key-value pair.
+//
+// If v is not a Handler, it will pass v to NewStructHandler and use the
+// returned StructHandler for decoding.
+func Unmarshal(data []byte, v interface{}) (err error) {
+	h, ok := v.(Handler)
+	if !ok {
+		h, err = NewStructHandler(v)
+		if err != nil {
+			return err
+		}
+	}
+	return gotoScanner(data, h)
+}
+
+// StructHandler unmarshals logfmt into a struct. It matches incoming keys to
+// the the struct's fields (either the struct field name or its tag, preferring
+// an exact match but also accepting a case-insensitive match.
+//
+// Field types supported by StructHandler are:
+//
+// 	all numeric types (e.g. float32, int, etc.)
+// 	[]byte
+// 	string
+// 	bool - true if key is present, false otherwise (the value is ignored).
+//	time.Duration - uses time.ParseDuration
+//
+// If a field is a pointer to an above type, and a matching key is not present
+// in the logfmt data, the pointer will be untouched.
+//
+// If v is not a pointer to an Handler or struct, Unmarshal will return an
+// error.
+type StructHandler struct {
+	rv reflect.Value
+}
+
+func NewStructHandler(v interface{}) (Handler, error) {
+	rv := reflect.ValueOf(v)
+	if rv.Kind() != reflect.Ptr || rv.IsNil() {
+		return nil, &InvalidUnmarshalError{reflect.TypeOf(v)}
+	}
+	return &StructHandler{rv: rv}, nil
+}
+
+func (h *StructHandler) HandleLogfmt(key, val []byte) error {
+	el := h.rv.Elem()
+	skey := string(key)
+	for i := 0; i < el.NumField(); i++ {
+		fv := el.Field(i)
+		ft := el.Type().Field(i)
+		switch {
+		case ft.Name == skey:
+		case ft.Tag.Get("logfmt") == skey:
+		case strings.EqualFold(ft.Name, skey):
+		default:
+			continue
+		}
+		if fv.Kind() == reflect.Ptr {
+			if fv.IsNil() {
+				t := fv.Type().Elem()
+				v := reflect.New(t)
+				fv.Set(v)
+				fv = v
+			}
+			fv = fv.Elem()
+		}
+		switch fv.Interface().(type) {
+		case time.Duration:
+			d, err := time.ParseDuration(string(val))
+			if err != nil {
+				return &UnmarshalTypeError{string(val), fv.Type()}
+			}
+			fv.Set(reflect.ValueOf(d))
+		case string:
+			fv.SetString(string(val))
+		case []byte:
+			b := make([]byte, len(val))
+			copy(b, val)
+			fv.SetBytes(b)
+		case bool:
+			fv.SetBool(true)
+		default:
+			switch {
+			case reflect.Int <= fv.Kind() && fv.Kind() <= reflect.Int64:
+				v, err := strconv.ParseInt(string(val), 10, 64)
+				if err != nil {
+					return err
+				}
+				fv.SetInt(v)
+			case reflect.Uint32 <= fv.Kind() && fv.Kind() <= reflect.Uint64:
+				v, err := strconv.ParseUint(string(val), 10, 64)
+				if err != nil {
+					return err
+				}
+				fv.SetUint(v)
+			case reflect.Float32 <= fv.Kind() && fv.Kind() <= reflect.Float64:
+				v, err := strconv.ParseFloat(string(val), 10)
+				if err != nil {
+					return err
+				}
+				fv.SetFloat(v)
+			default:
+				return &UnmarshalTypeError{string(val), fv.Type()}
+			}
+		}
+
+	}
+	return nil
+}
+
+// An InvalidUnmarshalError describes an invalid argument passed to Unmarshal.
+// (The argument to Unmarshal must be a non-nil pointer.)
+type InvalidUnmarshalError struct {
+	Type reflect.Type
+}
+
+func (e *InvalidUnmarshalError) Error() string {
+	if e.Type == nil {
+		return "logfmt: Unmarshal(nil)"
+	}
+
+	if e.Type.Kind() != reflect.Ptr {
+		return "logfmt: Unmarshal(non-pointer " + e.Type.String() + ")"
+	}
+	return "logfmt: Unmarshal(nil " + e.Type.String() + ")"
+}
+
+// An UnmarshalTypeError describes a logfmt value that was
+// not appropriate for a value of a specific Go type.
+type UnmarshalTypeError struct {
+	Value string       // the logfmt value
+	Type  reflect.Type // type of Go value it could not be assigned to
+}
+
+func (e *UnmarshalTypeError) Error() string {
+	return "logfmt: cannot unmarshal " + e.Value + " into Go value of type " + e.Type.String()
+}
diff --git a/vendor/github.com/kr/logfmt/scanner.go b/vendor/github.com/kr/logfmt/scanner.go
new file mode 100644
index 0000000..095221f
--- /dev/null
+++ b/vendor/github.com/kr/logfmt/scanner.go
@@ -0,0 +1,149 @@
+package logfmt
+
+import (
+	"errors"
+	"fmt"
+)
+
+var ErrUnterminatedString = errors.New("logfmt: unterminated string")
+
+func gotoScanner(data []byte, h Handler) (err error) {
+	saveError := func(e error) {
+		if err == nil {
+			err = e
+		}
+	}
+
+	var c byte
+	var i int
+	var m int
+	var key []byte
+	var val []byte
+	var ok bool
+	var esc bool
+
+garbage:
+	if i == len(data) {
+		return
+	}
+
+	c = data[i]
+	switch {
+	case c > ' ' && c != '"' && c != '=':
+		key, val = nil, nil
+		m = i
+		i++
+		goto key
+	default:
+		i++
+		goto garbage
+	}
+
+key:
+	if i >= len(data) {
+		if m >= 0 {
+			key = data[m:i]
+			saveError(h.HandleLogfmt(key, nil))
+		}
+		return
+	}
+
+	c = data[i]
+	switch {
+	case c > ' ' && c != '"' && c != '=':
+		i++
+		goto key
+	case c == '=':
+		key = data[m:i]
+		i++
+		goto equal
+	default:
+		key = data[m:i]
+		i++
+		saveError(h.HandleLogfmt(key, nil))
+		goto garbage
+	}
+
+equal:
+	if i >= len(data) {
+		if m >= 0 {
+			i--
+			key = data[m:i]
+			saveError(h.HandleLogfmt(key, nil))
+		}
+		return
+	}
+
+	c = data[i]
+	switch {
+	case c > ' ' && c != '"' && c != '=':
+		m = i
+		i++
+		goto ivalue
+	case c == '"':
+		m = i
+		i++
+		esc = false
+		goto qvalue
+	default:
+		if key != nil {
+			saveError(h.HandleLogfmt(key, val))
+		}
+		i++
+		goto garbage
+	}
+
+ivalue:
+	if i >= len(data) {
+		if m >= 0 {
+			val = data[m:i]
+			saveError(h.HandleLogfmt(key, val))
+		}
+		return
+	}
+
+	c = data[i]
+	switch {
+	case c > ' ' && c != '"' && c != '=':
+		i++
+		goto ivalue
+	default:
+		val = data[m:i]
+		saveError(h.HandleLogfmt(key, val))
+		i++
+		goto garbage
+	}
+
+qvalue:
+	if i >= len(data) {
+		if m >= 0 {
+			saveError(ErrUnterminatedString)
+		}
+		return
+	}
+
+	c = data[i]
+	switch c {
+	case '\\':
+		i += 2
+		esc = true
+		goto qvalue
+	case '"':
+		i++
+		val = data[m:i]
+		if esc {
+			val, ok = unquoteBytes(val)
+			if !ok {
+				saveError(fmt.Errorf("logfmt: error unquoting bytes %q", string(val)))
+				goto garbage
+			}
+		} else {
+			val = val[1 : len(val)-1]
+		}
+		saveError(h.HandleLogfmt(key, val))
+		goto garbage
+	default:
+		i++
+		goto qvalue
+	}
+}
diff --git a/vendor/github.com/kr/logfmt/unquote.go b/vendor/github.com/kr/logfmt/unquote.go
new file mode 100644
index 0000000..fb088a4
--- /dev/null
+++ b/vendor/github.com/kr/logfmt/unquote.go
@@ -0,0 +1,149 @@
+package logfmt
+
+import (
+	"strconv"
+	"unicode"
+	"unicode/utf16"
+	"unicode/utf8"
+)
+
+// Taken from Go's encoding/json
+
+// Copyright 2010 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// getu4 decodes \uXXXX from the beginning of s, returning the hex value,
+// or it returns -1.
+func getu4(s []byte) rune {
+	if len(s) < 6 || s[0] != '\\' || s[1] != 'u' {
+		return -1
+	}
+	r, err := strconv.ParseUint(string(s[2:6]), 16, 64)
+	if err != nil {
+		return -1
+	}
+	return rune(r)
+}
+
+// unquote converts a quoted JSON string literal s into an actual string t.
+// The rules are different than for Go, so cannot use strconv.Unquote.
+func unquote(s []byte) (t string, ok bool) {
+	s, ok = unquoteBytes(s)
+	t = string(s)
+	return
+}
+
+func unquoteBytes(s []byte) (t []byte, ok bool) {
+	if len(s) < 2 || s[0] != '"' || s[len(s)-1] != '"' {
+		return
+	}
+	s = s[1 : len(s)-1]
+
+	// Check for unusual characters. If there are none,
+	// then no unquoting is needed, so return a slice of the
+	// original bytes.
+	r := 0
+	for r < len(s) {
+		c := s[r]
+		if c == '\\' || c == '"' || c < ' ' {
+			break
+		}
+		if c < utf8.RuneSelf {
+			r++
+			continue
+		}
+		rr, size := utf8.DecodeRune(s[r:])
+		if rr == utf8.RuneError && size == 1 {
+			break
+		}
+		r += size
+	}
+	if r == len(s) {
+		return s, true
+	}
+
+	b := make([]byte, len(s)+2*utf8.UTFMax)
+	w := copy(b, s[0:r])
+	for r < len(s) {
+		// Out of room?  Can only happen if s is full of
+		// malformed UTF-8 and we're replacing each
+		// byte with RuneError.
+		if w >= len(b)-2*utf8.UTFMax {
+			nb := make([]byte, (len(b)+utf8.UTFMax)*2)
+			copy(nb, b[0:w])
+			b = nb
+		}
+		switch c := s[r]; {
+		case c == '\\':
+			r++
+			if r >= len(s) {
+				return
+			}
+			switch s[r] {
+			default:
+				return
+			case '"', '\\', '/', '\'':
+				b[w] = s[r]
+				r++
+				w++
+			case 'b':
+				b[w] = '\b'
+				r++
+				w++
+			case 'f':
+				b[w] = '\f'
+				r++
+				w++
+			case 'n':
+				b[w] = '\n'
+				r++
+				w++
+			case 'r':
+				b[w] = '\r'
+				r++
+				w++
+			case 't':
+				b[w] = '\t'
+				r++
+				w++
+			case 'u':
+				r--
+				rr := getu4(s[r:])
+				if rr < 0 {
+					return
+				}
+				r += 6
+				if utf16.IsSurrogate(rr) {
+					rr1 := getu4(s[r:])
+					if dec := utf16.DecodeRune(rr, rr1); dec != unicode.ReplacementChar {
+						// A valid pair; consume.
+						r += 6
+						w += utf8.EncodeRune(b[w:], dec)
+						break
+					}
+					// Invalid surrogate; fall back to replacement rune.
+					rr = unicode.ReplacementChar
+				}
+				w += utf8.EncodeRune(b[w:], rr)
+			}
+
+		// Quote, control characters are invalid.
+		case c == '"', c < ' ':
+			return
+
+		// ASCII
+		case c < utf8.RuneSelf:
+			b[w] = c
+			r++
+			w++
+
+		// Coerce to well-formed UTF-8.
+		default:
+			rr, size := utf8.DecodeRune(s[r:])
+			r += size
+			w += utf8.EncodeRune(b[w:], rr)
+		}
+	}
+	return b[0:w], true
+}
diff --git a/vendor/github.com/namsral/flag/LICENSE b/vendor/github.com/namsral/flag/LICENSE
new file mode 100644
index 0000000..0f2498b
--- /dev/null
+++ b/vendor/github.com/namsral/flag/LICENSE
@@ -0,0 +1,27 @@
+Copyright (c) 2012 The Go Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+   * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+   * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+   * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
\ No newline at end of file
diff --git a/vendor/github.com/namsral/flag/README.md b/vendor/github.com/namsral/flag/README.md
new file mode 100644
index 0000000..fba5546
--- /dev/null
+++ b/vendor/github.com/namsral/flag/README.md
@@ -0,0 +1,200 @@
+Flag
+===
+
+Flag is a drop in replacement for Go's flag package with the addition to parse files and environment variables. If you support the [twelve-factor app methodology][], Flag complies with the third factor; "Store config in the environment".
+
+[twelve-factor app methodology]: http://12factor.net
+
+An example using a gopher:
+
+```go
+$ cat > gopher.go
+    package main
+
+    import (
+        "fmt"
+    	"github.com/namsral/flag"
+	)
+    
+    var age int
+    
+    flag.IntVar(&age, "age", 0, "age of gopher")
+    flag.Parse()
+    
+    fmt.Print("age:", age)
+
+$ go run gopher.go -age 1
+age: 1
+```
+
+Same code but using an environment variable:
+
+```go
+$ export AGE=2
+$ go run gopher.go
+age: 2
+```
+    
+
+Same code but using a configuration file:
+
+```go
+$ cat > gopher.conf
+age 3
+
+$ go run gopher.go -config gopher.conf
+age: 3
+```
+
+The following table shows how flags are translated to environment variables and configuration files:
+
+| Type   | Flag          | Environment  | File         |
+| ------ | :------------ |:------------ |:------------ |
+| int    | -age 2        | AGE=2        | age 2        |
+| bool   | -female       | FEMALE=true  | female true  |
+| float  | -length 175.5 | LENGTH=175.5 | length 175.5 |
+| string | -name Gloria  | NAME=Gloria  | name Gloria  |
+
+This package is a port of Go's [flag][] package from the standard library with the addition of two functions `ParseEnv` and `ParseFile`.
+
+[flag]: http://golang.org/src/pkg/flagconfiguration
+
+
+Goals
+-----
+
+- Compatability with the original `flag` package
+- Support the [twelve-factor app methodology][]
+- Uniform user experience between the three input methods
+
+
+Why?
+---
+
+Why not use one of the many INI, JSON or YAML parsers?
+
+I find it best practice to have simple configuration options to control the behaviour of an applications when it starts up. Use basic types like ints, floats and strings for configuration options and store more complex data structures in the "datastore" layer.
+
+
+Usage
+---
+
+It's intended for projects which require a simple configuration made available through command-line flags, configuration files and shell environments. It's similar to the original `flag` package.
+
+Example:
+
+```go
+import "github.com/namsral/flag"
+
+flag.String(flag.DefaultConfigFlagname, "", "path to config file")
+flag.Int("age", 24, "help message for age")
+
+flag.Parse()
+```
+
+Order of precedence:
+
+1. Command line options
+2. Environment variables
+3. Configuration file
+4. Default values
+
+
+#### Parsing Configuration Files
+
+Create a configuration file:
+
+```go
+$ cat > ./gopher.conf
+# empty newlines and lines beginning with a "#" character are ignored.
+name bob
+
+# keys and values can also be separated by the "=" character
+age=20
+
+# booleans can be empty, set with 0, 1, true, false, etc
+hacker
+```
+
+Add a "config" flag:
+
+```go
+flag.String(flag.DefaultConfigFlagname, "", "path to config file")
+```
+
+Run the command:
+
+```go
+$ go run ./gopher.go -config ./gopher.conf
+```
+
+The default flag name for the configuration file is "config" and can be changed
+by setting `flag.DefaultConfigFlagname`:
+
+```go
+flag.DefaultConfigFlagname = "conf"
+flag.Parse()
+```
+
+#### Parsing Environment Variables
+
+Environment variables are parsed 1-on-1 with defined flags:
+
+```go
+$ export AGE=44
+$ go run ./gopher.go
+age=44
+```
+
+
+You can also parse prefixed environment variables by setting a prefix name when creating a new empty flag set:
+
+```go
+fs := flag.NewFlagSetWithEnvPrefix(os.Args[0], "GO", 0)
+fs.Int("age", 24, "help message for age")
+fs.Parse(os.Args[1:])
+...
+$ go export GO_AGE=33
+$ go run ./gopher.go
+age=33
+```
+
+
+For more examples see the [examples][] directory in the project repository.
+
+[examples]: https://github.com/namsral/flag/tree/master/examples
+
+That's it.
+
+
+License
+---
+
+
+Copyright (c) 2012 The Go Authors. All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+
+   * Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+   * Redistributions in binary form must reproduce the above
+copyright notice, this list of conditions and the following disclaimer
+in the documentation and/or other materials provided with the
+distribution.
+   * Neither the name of Google Inc. nor the names of its
+contributors may be used to endorse or promote products derived from
+this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
diff --git a/vendor/github.com/namsral/flag/flag.go b/vendor/github.com/namsral/flag/flag.go
new file mode 100644
index 0000000..9ee1d9c
--- /dev/null
+++ b/vendor/github.com/namsral/flag/flag.go
@@ -0,0 +1,1068 @@
+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+/*Package flag implements command-line flag parsing.
+okdaokddadok
+Usage:
+
+Define flags using flag.String(), Bool(), Int(), etc.
+
+This declares an integer flag, -flagname, stored in the pointer ip, with type *int.
+    import "flag"
+    var ip = flag.Int("flagname", 1234, "help message for flagname")
+If you like, you can bind the flag to a variable using the Var() functions.
+    var flagvar int
+    func init() {
+        flag.IntVar(&flagvar, "flagname", 1234, "help message for flagname")
+    }
+Or you can create custom flags that satisfy the Value interface (with
+pointer receivers) and couple them to flag parsing by
+    flag.Var(&flagVal, "name", "help message for flagname")
+For such flags, the default value is just the initial value of the variable.
+
+After all flags are defined, call
+    flag.Parse()
+to parse the command line into the defined flags.
+
+Flags may then be used directly. If you're using the flags themselves,
+they are all pointers; if you bind to variables, they're values.
+    fmt.Println("ip has value ", *ip)
+    fmt.Println("flagvar has value ", flagvar)
+
+After parsing, the arguments after the flag are available as the
+slice flag.Args() or individually as flag.Arg(i).
+The arguments are indexed from 0 through flag.NArg()-1.
+
+Command line flag syntax:
+    -flag
+    -flag=x
+    -flag x  // non-boolean flags only
+One or two minus signs may be used; they are equivalent.
+The last form is not permitted for boolean flags because the
+meaning of the command
+    cmd -x *
+will change if there is a file called 0, false, etc.  You must
+use the -flag=false form to turn off a boolean flag.
+
+Flag parsing stops just before the first non-flag argument
+("-" is a non-flag argument) or after the terminator "--".
+
+Integer flags accept 1234, 0664, 0x1234 and may be negative.
+Boolean flags may be 1, 0, t, f, true, false, TRUE, FALSE, True, False.
+Duration flags accept any input valid for time.ParseDuration.
+
+The default set of command-line flags is controlled by
+top-level functions.  The FlagSet type allows one to define
+independent sets of flags, such as to implement subcommands
+in a command-line interface. The methods of FlagSet are
+analogous to the top-level functions for the command-line
+flag set.
+*/
+package flag
+
+import (
+	"bufio"
+	"errors"
+	"fmt"
+	"io"
+	"os"
+	"sort"
+	"strconv"
+	"strings"
+	"time"
+)
+
+// EnvironmentPrefix defines a string that will be implicitely prefixed to a
+// flag name before looking it up in the environment variables.
+var EnvironmentPrefix = ""
+
+// DefaultConfigFlagname defines the flag name of the optional config file
+// path. Used to lookup and parse the config file when a default is set and
+// available on disk.
+var DefaultConfigFlagname = "config"
+
+// ErrHelp is the error returned if the flag -help is invoked but no such flag is defined.
+var ErrHelp = errors.New("flag: help requested")
+
+// -- bool Value
+type boolValue bool
+
+func newBoolValue(val bool, p *bool) *boolValue {
+	*p = val
+	return (*boolValue)(p)
+}
+
+func (b *boolValue) Set(s string) error {
+	v, err := strconv.ParseBool(s)
+	*b = boolValue(v)
+	return err
+}
+
+func (b *boolValue) Get() interface{} { return bool(*b) }
+
+func (b *boolValue) String() string { return fmt.Sprintf("%v", *b) }
+
+func (b *boolValue) IsBoolFlag() bool { return true }
+
+// optional interface to indicate boolean flags that can be
+// supplied without "=value" text
+type boolFlag interface {
+	Value
+	IsBoolFlag() bool
+}
+
+// -- int Value
+type intValue int
+
+func newIntValue(val int, p *int) *intValue {
+	*p = val
+	return (*intValue)(p)
+}
+
+func (i *intValue) Set(s string) error {
+	v, err := strconv.ParseInt(s, 0, 64)
+	*i = intValue(v)
+	return err
+}
+
+func (i *intValue) Get() interface{} { return int(*i) }
+
+func (i *intValue) String() string { return fmt.Sprintf("%v", *i) }
+
+// -- int64 Value
+type int64Value int64
+
+func newInt64Value(val int64, p *int64) *int64Value {
+	*p = val
+	return (*int64Value)(p)
+}
+
+func (i *int64Value) Set(s string) error {
+	v, err := strconv.ParseInt(s, 0, 64)
+	*i = int64Value(v)
+	return err
+}
+
+func (i *int64Value) Get() interface{} { return int64(*i) }
+
+func (i *int64Value) String() string { return fmt.Sprintf("%v", *i) }
+
+// -- uint Value
+type uintValue uint
+
+func newUintValue(val uint, p *uint) *uintValue {
+	*p = val
+	return (*uintValue)(p)
+}
+
+func (i *uintValue) Set(s string) error {
+	v, err := strconv.ParseUint(s, 0, 64)
+	*i = uintValue(v)
+	return err
+}
+
+func (i *uintValue) Get() interface{} { return uint(*i) }
+
+func (i *uintValue) String() string { return fmt.Sprintf("%v", *i) }
+
+// -- uint64 Value
+type uint64Value uint64
+
+func newUint64Value(val uint64, p *uint64) *uint64Value {
+	*p = val
+	return (*uint64Value)(p)
+}
+
+func (i *uint64Value) Set(s string) error {
+	v, err := strconv.ParseUint(s, 0, 64)
+	*i = uint64Value(v)
+	return err
+}
+
+func (i *uint64Value) Get() interface{} { return uint64(*i) }
+
+func (i *uint64Value) String() string { return fmt.Sprintf("%v", *i) }
+
+// -- string Value
+type stringValue string
+
+func newStringValue(val string, p *string) *stringValue {
+	*p = val
+	return (*stringValue)(p)
+}
+
+func (s *stringValue) Set(val string) error {
+	*s = stringValue(val)
+	return nil
+}
+
+func (s *stringValue) Get() interface{} { return string(*s) }
+
+func (s *stringValue) String() string { return fmt.Sprintf("%s", *s) }
+
+// -- float64 Value
+type float64Value float64
+
+func newFloat64Value(val float64, p *float64) *float64Value {
+	*p = val
+	return (*float64Value)(p)
+}
+
+func (f *float64Value) Set(s string) error {
+	v, err := strconv.ParseFloat(s, 64)
+	*f = float64Value(v)
+	return err
+}
+
+func (f *float64Value) Get() interface{} { return float64(*f) }
+
+func (f *float64Value) String() string { return fmt.Sprintf("%v", *f) }
+
+// -- time.Duration Value
+type durationValue time.Duration
+
+func newDurationValue(val time.Duration, p *time.Duration) *durationValue {
+	*p = val
+	return (*durationValue)(p)
+}
+
+func (d *durationValue) Set(s string) error {
+	v, err := time.ParseDuration(s)
+	*d = durationValue(v)
+	return err
+}
+
+func (d *durationValue) Get() interface{} { return time.Duration(*d) }
+
+func (d *durationValue) String() string { return (*time.Duration)(d).String() }
+
+// Value is the interface to the dynamic value stored in a flag.
+// (The default value is represented as a string.)
+//
+// If a Value has an IsBoolFlag() bool method returning true,
+// the command-line parser makes -name equivalent to -name=true
+// rather than using the next command-line argument.
+type Value interface {
+	String() string
+	Set(string) error
+}
+
+// Getter is an interface that allows the contents of a Value to be retrieved.
+// It wraps the Value interface, rather than being part of it, because it
+// appeared after Go 1 and its compatibility rules. All Value types provided
+// by this package satisfy the Getter interface.
+type Getter interface {
+	Value
+	Get() interface{}
+}
+
+// ErrorHandling defines how to handle flag parsing errors.
+type ErrorHandling int
+
+const (
+	ContinueOnError ErrorHandling = iota
+	ExitOnError
+	PanicOnError
+)
+
+// A FlagSet represents a set of defined flags.  The zero value of a FlagSet
+// has no name and has ContinueOnError error handling.
+type FlagSet struct {
+	// Usage is the function called when an error occurs while parsing flags.
+	// The field is a function (not a method) that may be changed to point to
+	// a custom error handler.
+	Usage func()
+
+	name          string
+	parsed        bool
+	actual        map[string]*Flag
+	formal        map[string]*Flag
+	envPrefix     string   // prefix to all env variable names
+	args          []string // arguments after flags
+	exitOnError   bool     // does the program exit if there's an error?
+	errorHandling ErrorHandling
+	output        io.Writer // nil means stderr; use out() accessor
+}
+
+// A Flag represents the state of a flag.
+type Flag struct {
+	Name     string // name as it appears on command line
+	Usage    string // help message
+	Value    Value  // value as set
+	DefValue string // default value (as text); for usage message
+}
+
+// sortFlags returns the flags as a slice in lexicographical sorted order.
+func sortFlags(flags map[string]*Flag) []*Flag {
+	list := make(sort.StringSlice, len(flags))
+	i := 0
+	for _, f := range flags {
+		list[i] = f.Name
+		i++
+	}
+	list.Sort()
+	result := make([]*Flag, len(list))
+	for i, name := range list {
+		result[i] = flags[name]
+	}
+	return result
+}
+
+func (f *FlagSet) out() io.Writer {
+	if f.output == nil {
+		return os.Stderr
+	}
+	return f.output
+}
+
+// SetOutput sets the destination for usage and error messages.
+// If output is nil, os.Stderr is used.
+func (f *FlagSet) SetOutput(output io.Writer) {
+	f.output = output
+}
+
+// VisitAll visits the flags in lexicographical order, calling fn for each.
+// It visits all flags, even those not set.
+func (f *FlagSet) VisitAll(fn func(*Flag)) {
+	for _, flag := range sortFlags(f.formal) {
+		fn(flag)
+	}
+}
+
+// VisitAll visits the command-line flags in lexicographical order, calling
+// fn for each.  It visits all flags, even those not set.
+func VisitAll(fn func(*Flag)) {
+	CommandLine.VisitAll(fn)
+}
+
+// Visit visits the flags in lexicographical order, calling fn for each.
+// It visits only those flags that have been set.
+func (f *FlagSet) Visit(fn func(*Flag)) {
+	for _, flag := range sortFlags(f.actual) {
+		fn(flag)
+	}
+}
+
+// Visit visits the command-line flags in lexicographical order, calling fn
+// for each.  It visits only those flags that have been set.
+func Visit(fn func(*Flag)) {
+	CommandLine.Visit(fn)
+}
+
+// Lookup returns the Flag structure of the named flag, returning nil if none exists.
+func (f *FlagSet) Lookup(name string) *Flag {
+	return f.formal[name]
+}
+
+// Lookup returns the Flag structure of the named command-line flag,
+// returning nil if none exists.
+func Lookup(name string) *Flag {
+	return CommandLine.formal[name]
+}
+
+// Set sets the value of the named flag.
+func (f *FlagSet) Set(name, value string) error {
+	flag, ok := f.formal[name]
+	if !ok {
+		return fmt.Errorf("no such flag -%v", name)
+	}
+	err := flag.Value.Set(value)
+	if err != nil {
+		return err
+	}
+	if f.actual == nil {
+		f.actual = make(map[string]*Flag)
+	}
+	f.actual[name] = flag
+	return nil
+}
+
+// Set sets the value of the named command-line flag.
+func Set(name, value string) error {
+	return CommandLine.Set(name, value)
+}
+
+// PrintDefaults prints, to standard error unless configured
+// otherwise, the default values of all defined flags in the set.
+func (f *FlagSet) PrintDefaults() {
+	f.VisitAll(func(flag *Flag) {
+		format := "  -%s=%s: %s\n"
+		if _, ok := flag.Value.(*stringValue); ok {
+			// put quotes on the value
+			format = "  -%s=%q: %s\n"
+		}
+		fmt.Fprintf(f.out(), format, flag.Name, flag.DefValue, flag.Usage)
+	})
+}
+
+// PrintDefaults prints to standard error the default values of all defined command-line flags.
+func PrintDefaults() {
+	CommandLine.PrintDefaults()
+}
+
+// defaultUsage is the default function to print a usage message.
+func defaultUsage(f *FlagSet) {
+	if f.name == "" {
+		fmt.Fprintf(f.out(), "Usage:\n")
+	} else {
+		fmt.Fprintf(f.out(), "Usage of %s:\n", f.name)
+	}
+	f.PrintDefaults()
+}
+
+// NOTE: Usage is not just defaultUsage(CommandLine)
+// because it serves (via godoc flag Usage) as the example
+// for how to write your own usage function.
+
+// Usage prints to standard error a usage message documenting all defined command-line flags.
+// The function is a variable that may be changed to point to a custom function.
+var Usage = func() {
+	fmt.Fprintf(os.Stderr, "Usage of %s:\n", os.Args[0])
+	PrintDefaults()
+}
+
+// NFlag returns the number of flags that have been set.
+func (f *FlagSet) NFlag() int { return len(f.actual) }
+
+// NFlag returns the number of command-line flags that have been set.
+func NFlag() int { return len(CommandLine.actual) }
+
+// Arg returns the i'th argument.  Arg(0) is the first remaining argument
+// after flags have been processed.
+func (f *FlagSet) Arg(i int) string {
+	if i < 0 || i >= len(f.args) {
+		return ""
+	}
+	return f.args[i]
+}
+
+// Arg returns the i'th command-line argument.  Arg(0) is the first remaining argument
+// after flags have been processed.
+func Arg(i int) string {
+	return CommandLine.Arg(i)
+}
+
+// NArg is the number of arguments remaining after flags have been processed.
+func (f *FlagSet) NArg() int { return len(f.args) }
+
+// NArg is the number of arguments remaining after flags have been processed.
+func NArg() int { return len(CommandLine.args) }
+
+// Args returns the non-flag arguments.
+func (f *FlagSet) Args() []string { return f.args }
+
+// Args returns the non-flag command-line arguments.
+func Args() []string { return CommandLine.args }
+
+// BoolVar defines a bool flag with specified name, default value, and usage string.
+// The argument p points to a bool variable in which to store the value of the flag.
+func (f *FlagSet) BoolVar(p *bool, name string, value bool, usage string) {
+	f.Var(newBoolValue(value, p), name, usage)
+}
+
+// BoolVar defines a bool flag with specified name, default value, and usage string.
+// The argument p points to a bool variable in which to store the value of the flag.
+func BoolVar(p *bool, name string, value bool, usage string) {
+	CommandLine.Var(newBoolValue(value, p), name, usage)
+}
+
+// Bool defines a bool flag with specified name, default value, and usage string.
+// The return value is the address of a bool variable that stores the value of the flag.
+func (f *FlagSet) Bool(name string, value bool, usage string) *bool {
+	p := new(bool)
+	f.BoolVar(p, name, value, usage)
+	return p
+}
+
+// Bool defines a bool flag with specified name, default value, and usage string.
+// The return value is the address of a bool variable that stores the value of the flag.
+func Bool(name string, value bool, usage string) *bool {
+	return CommandLine.Bool(name, value, usage)
+}
+
+// IntVar defines an int flag with specified name, default value, and usage string.
+// The argument p points to an int variable in which to store the value of the flag.
+func (f *FlagSet) IntVar(p *int, name string, value int, usage string) {
+	f.Var(newIntValue(value, p), name, usage)
+}
+
+// IntVar defines an int flag with specified name, default value, and usage string.
+// The argument p points to an int variable in which to store the value of the flag.
+func IntVar(p *int, name string, value int, usage string) {
+	CommandLine.Var(newIntValue(value, p), name, usage)
+}
+
+// Int defines an int flag with specified name, default value, and usage string.
+// The return value is the address of an int variable that stores the value of the flag.
+func (f *FlagSet) Int(name string, value int, usage string) *int {
+	p := new(int)
+	f.IntVar(p, name, value, usage)
+	return p
+}
+
+// Int defines an int flag with specified name, default value, and usage string.
+// The return value is the address of an int variable that stores the value of the flag.
+func Int(name string, value int, usage string) *int {
+	return CommandLine.Int(name, value, usage)
+}
+
+// Int64Var defines an int64 flag with specified name, default value, and usage string.
+// The argument p points to an int64 variable in which to store the value of the flag.
+func (f *FlagSet) Int64Var(p *int64, name string, value int64, usage string) {
+	f.Var(newInt64Value(value, p), name, usage)
+}
+
+// Int64Var defines an int64 flag with specified name, default value, and usage string.
+// The argument p points to an int64 variable in which to store the value of the flag.
+func Int64Var(p *int64, name string, value int64, usage string) {
+	CommandLine.Var(newInt64Value(value, p), name, usage)
+}
+
+// Int64 defines an int64 flag with specified name, default value, and usage string.
+// The return value is the address of an int64 variable that stores the value of the flag.
+func (f *FlagSet) Int64(name string, value int64, usage string) *int64 {
+	p := new(int64)
+	f.Int64Var(p, name, value, usage)
+	return p
+}
+
+// Int64 defines an int64 flag with specified name, default value, and usage string.
+// The return value is the address of an int64 variable that stores the value of the flag.
+func Int64(name string, value int64, usage string) *int64 {
+	return CommandLine.Int64(name, value, usage)
+}
+
+// UintVar defines a uint flag with specified name, default value, and usage string.
+// The argument p points to a uint variable in which to store the value of the flag.
+func (f *FlagSet) UintVar(p *uint, name string, value uint, usage string) {
+	f.Var(newUintValue(value, p), name, usage)
+}
+
+// UintVar defines a uint flag with specified name, default value, and usage string.
+// The argument p points to a uint  variable in which to store the value of the flag.
+func UintVar(p *uint, name string, value uint, usage string) {
+	CommandLine.Var(newUintValue(value, p), name, usage)
+}
+
+// Uint defines a uint flag with specified name, default value, and usage string.
+// The return value is the address of a uint  variable that stores the value of the flag.
+func (f *FlagSet) Uint(name string, value uint, usage string) *uint {
+	p := new(uint)
+	f.UintVar(p, name, value, usage)
+	return p
+}
+
+// Uint defines a uint flag with specified name, default value, and usage string.
+// The return value is the address of a uint  variable that stores the value of the flag.
+func Uint(name string, value uint, usage string) *uint {
+	return CommandLine.Uint(name, value, usage)
+}
+
+// Uint64Var defines a uint64 flag with specified name, default value, and usage string.
+// The argument p points to a uint64 variable in which to store the value of the flag.
+func (f *FlagSet) Uint64Var(p *uint64, name string, value uint64, usage string) {
+	f.Var(newUint64Value(value, p), name, usage)
+}
+
+// Uint64Var defines a uint64 flag with specified name, default value, and usage string.
+// The argument p points to a uint64 variable in which to store the value of the flag.
+func Uint64Var(p *uint64, name string, value uint64, usage string) {
+	CommandLine.Var(newUint64Value(value, p), name, usage)
+}
+
+// Uint64 defines a uint64 flag with specified name, default value, and usage string.
+// The return value is the address of a uint64 variable that stores the value of the flag.
+func (f *FlagSet) Uint64(name string, value uint64, usage string) *uint64 {
+	p := new(uint64)
+	f.Uint64Var(p, name, value, usage)
+	return p
+}
+
+// Uint64 defines a uint64 flag with specified name, default value, and usage string.
+// The return value is the address of a uint64 variable that stores the value of the flag.
+func Uint64(name string, value uint64, usage string) *uint64 {
+	return CommandLine.Uint64(name, value, usage)
+}
+
+// StringVar defines a string flag with specified name, default value, and usage string.
+// The argument p points to a string variable in which to store the value of the flag.
+func (f *FlagSet) StringVar(p *string, name string, value string, usage string) {
+	f.Var(newStringValue(value, p), name, usage)
+}
+
+// StringVar defines a string flag with specified name, default value, and usage string.
+// The argument p points to a string variable in which to store the value of the flag.
+func StringVar(p *string, name string, value string, usage string) {
+	CommandLine.Var(newStringValue(value, p), name, usage)
+}
+
+// String defines a string flag with specified name, default value, and usage string.
+// The return value is the address of a string variable that stores the value of the flag.
+func (f *FlagSet) String(name string, value string, usage string) *string {
+	p := new(string)
+	f.StringVar(p, name, value, usage)
+	return p
+}
+
+// String defines a string flag with specified name, default value, and usage string.
+// The return value is the address of a string variable that stores the value of the flag.
+func String(name string, value string, usage string) *string {
+	return CommandLine.String(name, value, usage)
+}
+
+// Float64Var defines a float64 flag with specified name, default value, and usage string.
+// The argument p points to a float64 variable in which to store the value of the flag.
+func (f *FlagSet) Float64Var(p *float64, name string, value float64, usage string) {
+	f.Var(newFloat64Value(value, p), name, usage)
+}
+
+// Float64Var defines a float64 flag with specified name, default value, and usage string.
+// The argument p points to a float64 variable in which to store the value of the flag.
+func Float64Var(p *float64, name string, value float64, usage string) {
+	CommandLine.Var(newFloat64Value(value, p), name, usage)
+}
+
+// Float64 defines a float64 flag with specified name, default value, and usage string.
+// The return value is the address of a float64 variable that stores the value of the flag.
+func (f *FlagSet) Float64(name string, value float64, usage string) *float64 {
+	p := new(float64)
+	f.Float64Var(p, name, value, usage)
+	return p
+}
+
+// Float64 defines a float64 flag with specified name, default value, and usage string.
+// The return value is the address of a float64 variable that stores the value of the flag.
+func Float64(name string, value float64, usage string) *float64 {
+	return CommandLine.Float64(name, value, usage)
+}
+
+// DurationVar defines a time.Duration flag with specified name, default value, and usage string.
+// The argument p points to a time.Duration variable in which to store the value of the flag.
+func (f *FlagSet) DurationVar(p *time.Duration, name string, value time.Duration, usage string) {
+	f.Var(newDurationValue(value, p), name, usage)
+}
+
+// DurationVar defines a time.Duration flag with specified name, default value, and usage string.
+// The argument p points to a time.Duration variable in which to store the value of the flag.
+func DurationVar(p *time.Duration, name string, value time.Duration, usage string) {
+	CommandLine.Var(newDurationValue(value, p), name, usage)
+}
+
+// Duration defines a time.Duration flag with specified name, default value, and usage string.
+// The return value is the address of a time.Duration variable that stores the value of the flag.
+func (f *FlagSet) Duration(name string, value time.Duration, usage string) *time.Duration {
+	p := new(time.Duration)
+	f.DurationVar(p, name, value, usage)
+	return p
+}
+
+// Duration defines a time.Duration flag with specified name, default value, and usage string.
+// The return value is the address of a time.Duration variable that stores the value of the flag.
+func Duration(name string, value time.Duration, usage string) *time.Duration {
+	return CommandLine.Duration(name, value, usage)
+}
+
+// Var defines a flag with the specified name and usage string. The type and
+// value of the flag are represented by the first argument, of type Value, which
+// typically holds a user-defined implementation of Value. For instance, the
+// caller could create a flag that turns a comma-separated string into a slice
+// of strings by giving the slice the methods of Value; in particular, Set would
+// decompose the comma-separated string into the slice.
+func (f *FlagSet) Var(value Value, name string, usage string) {
+	// Remember the default value as a string; it won't change.
+	flag := &Flag{name, usage, value, value.String()}
+	_, alreadythere := f.formal[name]
+	if alreadythere {
+		var msg string
+		if f.name == "" {
+			msg = fmt.Sprintf("flag redefined: %s", name)
+		} else {
+			msg = fmt.Sprintf("%s flag redefined: %s", f.name, name)
+		}
+		fmt.Fprintln(f.out(), msg)
+		panic(msg) // Happens only if flags are declared with identical names
+	}
+	if f.formal == nil {
+		f.formal = make(map[string]*Flag)
+	}
+	f.formal[name] = flag
+}
+
+// Var defines a flag with the specified name and usage string. The type and
+// value of the flag are represented by the first argument, of type Value, which
+// typically holds a user-defined implementation of Value. For instance, the
+// caller could create a flag that turns a comma-separated string into a slice
+// of strings by giving the slice the methods of Value; in particular, Set would
+// decompose the comma-separated string into the slice.
+func Var(value Value, name string, usage string) {
+	CommandLine.Var(value, name, usage)
+}
+
+// failf prints to standard error a formatted error and usage message and
+// returns the error.
+func (f *FlagSet) failf(format string, a ...interface{}) error {
+	err := fmt.Errorf(format, a...)
+	fmt.Fprintln(f.out(), err)
+	f.usage()
+	return err
+}
+
+// usage calls the Usage method for the flag set, or the usage function if
+// the flag set is CommandLine.
+func (f *FlagSet) usage() {
+	if f == CommandLine {
+		Usage()
+	} else if f.Usage == nil {
+		defaultUsage(f)
+	} else {
+		f.Usage()
+	}
+}
+
+// parseOne parses one flag. It reports whether a flag was seen.
+func (f *FlagSet) parseOne() (bool, error) {
+	if len(f.args) == 0 {
+		return false, nil
+	}
+	s := f.args[0]
+	if len(s) == 0 || s[0] != '-' || len(s) == 1 {
+		return false, nil
+	}
+	numMinuses := 1
+	if s[1] == '-' {
+		numMinuses++
+		if len(s) == 2 { // "--" terminates the flags
+			f.args = f.args[1:]
+			return false, nil
+		}
+	}
+	name := s[numMinuses:]
+	if len(name) == 0 || name[0] == '-' || name[0] == '=' {
+		return false, f.failf("bad flag syntax: %s", s)
+	}
+
+	// ignore go test flags
+	if strings.HasPrefix(name, "test.") {
+		return false, nil
+	}
+
+	// it's a flag. does it have an argument?
+	f.args = f.args[1:]
+	hasValue := false
+	value := ""
+	for i := 1; i < len(name); i++ { // equals cannot be first
+		if name[i] == '=' {
+			value = name[i+1:]
+			hasValue = true
+			name = name[0:i]
+			break
+		}
+	}
+	m := f.formal
+	flag, alreadythere := m[name] // BUG
+	if !alreadythere {
+		if name == "help" || name == "h" { // special case for nice help message.
+			f.usage()
+			return false, ErrHelp
+		}
+		return false, f.failf("flag provided but not defined: -%s", name)
+	}
+	if fv, ok := flag.Value.(boolFlag); ok && fv.IsBoolFlag() { // special case: doesn't need an arg
+		if hasValue {
+			if err := fv.Set(value); err != nil {
+				return false, f.failf("invalid boolean value %q for  -%s: %v", value, name, err)
+			}
+		} else {
+			fv.Set("true")
+		}
+	} else {
+		// It must have a value, which might be the next argument.
+		if !hasValue && len(f.args) > 0 {
+			// value is the next arg
+			hasValue = true
+			value, f.args = f.args[0], f.args[1:]
+		}
+		if !hasValue {
+			return false, f.failf("flag needs an argument: -%s", name)
+		}
+		if err := flag.Value.Set(value); err != nil {
+			return false, f.failf("invalid value %q for flag -%s: %v", value, name, err)
+		}
+	}
+	if f.actual == nil {
+		f.actual = make(map[string]*Flag)
+	}
+	f.actual[name] = flag
+	return true, nil
+}
+
+// Parse parses flag definitions from the argument list, which should not
+// include the command name.  Must be called after all flags in the FlagSet
+// are defined and before flags are accessed by the program.
+// The return value will be ErrHelp if -help was set but not defined.
+func (f *FlagSet) Parse(arguments []string) error {
+	f.parsed = true
+	f.args = arguments
+	for {
+		seen, err := f.parseOne()
+		if seen {
+			continue
+		}
+		if err == nil {
+			break
+		}
+		switch f.errorHandling {
+		case ContinueOnError:
+			return err
+		case ExitOnError:
+			os.Exit(2)
+		case PanicOnError:
+			panic(err)
+		}
+	}
+
+	// Parse environment variables
+	if err := f.ParseEnv(os.Environ()); err != nil {
+		switch f.errorHandling {
+		case ContinueOnError:
+			return err
+		case ExitOnError:
+			os.Exit(2)
+		case PanicOnError:
+			panic(err)
+		}
+		return err
+	}
+
+	// Parse configuration from file
+	var cFile string
+	if cf := f.formal[DefaultConfigFlagname]; cf != nil {
+		cFile = cf.Value.String()
+	}
+	if cf := f.actual[DefaultConfigFlagname]; cf != nil {
+		cFile = cf.Value.String()
+	}
+	if cFile != "" {
+		if err := f.ParseFile(cFile); err != nil {
+			switch f.errorHandling {
+			case ContinueOnError:
+				return err
+			case ExitOnError:
+				os.Exit(2)
+			case PanicOnError:
+				panic(err)
+			}
+			return err
+		}
+	}
+
+	return nil
+}
+
+// ParseEnv parses flags from environment variables.
+// Flags already set will be ignored.
+func (f *FlagSet) ParseEnv(environ []string) error {
+
+	m := f.formal
+
+	env := make(map[string]string)
+	for _, s := range environ {
+		i := strings.Index(s, "=")
+		if i < 1 {
+			continue
+		}
+		env[s[0:i]] = s[i+1 : len(s)]
+	}
+
+	for _, flag := range m {
+		name := flag.Name
+		_, set := f.actual[name]
+		if set {
+			continue
+		}
+
+		flag, alreadythere := m[name]
+		if !alreadythere {
+			if name == "help" || name == "h" { // special case for nice help message.
+				f.usage()
+				return ErrHelp
+			}
+			return f.failf("environment variable provided but not defined: %s", name)
+		}
+
+		envKey := strings.ToUpper(flag.Name)
+		if f.envPrefix != "" {
+			envKey = f.envPrefix + "_" + envKey
+		}
+		envKey = strings.Replace(envKey, "-", "_", -1)
+
+		value, isSet := env[envKey]
+		if !isSet {
+			continue
+		}
+
+		hasValue := false
+		if len(value) > 0 {
+			hasValue = true
+		}
+
+		if fv, ok := flag.Value.(boolFlag); ok && fv.IsBoolFlag() { // special case: doesn't need an arg
+			if hasValue {
+				if err := fv.Set(value); err != nil {
+					return f.failf("invalid boolean value %q for environment variable %s: %v", value, name, err)
+				}
+			} else {
+				// flag without value is regarded a bool
+				fv.Set("true")
+			}
+		} else {
+			if err := flag.Value.Set(value); err != nil {
+				return f.failf("invalid value %q for environment variable %s: %v", value, name, err)
+			}
+		}
+
+		// update f.actual
+		if f.actual == nil {
+			f.actual = make(map[string]*Flag)
+		}
+		f.actual[name] = flag
+
+	}
+	return nil
+}
+
+// ParseFile parses flags from the file in path.
+// Same format as commandline argumens, newlines and lines beginning with a
+// "#" charater are ignored. Flags already set will be ignored.
+func (f *FlagSet) ParseFile(path string) error {
+
+	// Extract arguments from file
+	fp, err := os.Open(path)
+	if err != nil {
+		return err
+	}
+	defer fp.Close()
+
+	scanner := bufio.NewScanner(fp)
+	for scanner.Scan() {
+		line := scanner.Text()
+
+		// Ignore empty lines
+		if len(line) == 0 {
+			continue
+		}
+
+		// Ignore comments
+		if line[:1] == "#" {
+			continue
+		}
+
+		// Match `key=value` and `key value`
+		var name, value string
+		hasValue := false
+		for i, v := range line {
+			if v == '=' || v == ' ' {
+				hasValue = true
+				name, value = line[:i], line[i+1:]
+				break
+			}
+		}
+
+		if hasValue == false {
+			name = line
+		}
+
+		// Ignore flag when already set; arguments have precedence over file
+		if f.actual[name] != nil {
+			continue
+		}
+
+		m := f.formal
+		flag, alreadythere := m[name]
+		if !alreadythere {
+			if name == "help" || name == "h" { // special case for nice help message.
+				f.usage()
+				return ErrHelp
+			}
+			return f.failf("configuration variable provided but not defined: %s", name)
+		}
+
+		if fv, ok := flag.Value.(boolFlag); ok && fv.IsBoolFlag() { // special case: doesn't need an arg
+			if hasValue {
+				if err := fv.Set(value); err != nil {
+					return f.failf("invalid boolean value %q for configuration variable %s: %v", value, name, err)
+				}
+			} else {
+				// flag without value is regarded a bool
+				fv.Set("true")
+			}
+		} else {
+			if err := flag.Value.Set(value); err != nil {
+				return f.failf("invalid value %q for configuration variable %s: %v", value, name, err)
+			}
+		}
+
+		// update f.actual
+		if f.actual == nil {
+			f.actual = make(map[string]*Flag)
+		}
+		f.actual[name] = flag
+	}
+
+	if err := scanner.Err(); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+// Parsed reports whether f.Parse has been called.
+func (f *FlagSet) Parsed() bool {
+	return f.parsed
+}
+
+// Parse parses the command-line flags from os.Args[1:].  Must be called
+// after all flags are defined and before flags are accessed by the program.
+func Parse() {
+	// Ignore errors; CommandLine is set for ExitOnError.
+	CommandLine.Parse(os.Args[1:])
+}
+
+// Parsed returns true if the command-line flags have been parsed.
+func Parsed() bool {
+	return CommandLine.Parsed()
+}
+
+// CommandLine is the default set of command-line flags, parsed from os.Args.
+// The top-level functions such as BoolVar, Arg, and on are wrappers for the
+// methods of CommandLine.
+var CommandLine = NewFlagSet(os.Args[0], ExitOnError)
+
+// NewFlagSet returns a new, empty flag set with the specified name and
+// error handling property.
+func NewFlagSet(name string, errorHandling ErrorHandling) *FlagSet {
+	f := &FlagSet{
+		name:          name,
+		errorHandling: errorHandling,
+	}
+	return f
+}
+
+// NewFlagSetWithEnvPrefix returns a new empty flag set with the specified name,
+// environment variable prefix, and error handling property.
+func NewFlagSetWithEnvPrefix(name string, prefix string, errorHandling ErrorHandling) *FlagSet {
+	f := NewFlagSet(name, errorHandling)
+	f.envPrefix = prefix
+	return f
+}
+
+// Init sets the name, environment name prefix, and error handling property
+// for a flag set.
+// By default, the zero FlagSet uses an empty name, EnvironmentPrefix, and the
+// ContinueOnError error handling policy.
+func (f *FlagSet) Init(name string, errorHandling ErrorHandling) {
+	f.name = name
+	f.envPrefix = EnvironmentPrefix
+	f.errorHandling = errorHandling
+}