diff --git a/route/matcher.go b/route/matcher.go
index 710be013c..3a7274c66 100644
--- a/route/matcher.go
+++ b/route/matcher.go
@@ -1,8 +1,6 @@
 package route
 
 import (
-	"log"
-	"path"
 	"strings"
 )
 
@@ -21,12 +19,7 @@ func prefixMatcher(uri string, r *Route) bool {
 	return strings.HasPrefix(uri, r.Path)
 }
 
-// globMatcher matches path to the routes' path using globbing.
+// globMatcher matches path to the routes' path using gobwas/glob.
 func globMatcher(uri string, r *Route) bool {
-	var hasMatch, err = path.Match(r.Path, uri)
-	if err != nil {
-		log.Printf("[ERROR] Glob matching error %s for path %s route %s", err, uri, r.Path)
-		return false
-	}
-	return hasMatch
+	return r.Glob.Match(uri)
 }
diff --git a/route/matcher_test.go b/route/matcher_test.go
index 592d7d610..59a068c90 100644
--- a/route/matcher_test.go
+++ b/route/matcher_test.go
@@ -2,6 +2,8 @@ package route
 
 import (
 	"testing"
+
+	"github.com/gobwas/glob"
 )
 
 func TestPrefixMatcher(t *testing.T) {
@@ -38,16 +40,20 @@ func TestGlobMatcher(t *testing.T) {
 		{uri: "/fools", matches: true, route: &Route{Path: "/foo*"}},
 		{uri: "/fools", matches: true, route: &Route{Path: "/foo*"}},
 		{uri: "/foo/x/bar", matches: true, route: &Route{Path: "/foo/*/bar"}},
+		{uri: "/foo/x/y/z/w/bar", matches: true, route: &Route{Path: "/foo/**"}},
+		{uri: "/foo/x/y/z/w/bar", matches: true, route: &Route{Path: "/foo/**/bar"}},
 
 		// error flows
 		{uri: "/fo", matches: false, route: &Route{Path: "/foo"}},
 		{uri: "/fools", matches: false, route: &Route{Path: "/foo"}},
 		{uri: "/fo", matches: false, route: &Route{Path: "/foo*"}},
 		{uri: "/fools", matches: false, route: &Route{Path: "/foo.*"}},
+		{uri: "/foo/x/y/z/w/baz", matches: false, route: &Route{Path: "/foo/**/bar"}},
 	}
 
 	for _, tt := range tests {
 		t.Run(tt.uri, func(t *testing.T) {
+			tt.route.Glob = glob.MustCompile(tt.route.Path)
 			if got, want := globMatcher(tt.uri, tt.route), tt.matches; got != want {
 				t.Fatalf("got %v want %v", got, want)
 			}
diff --git a/route/route.go b/route/route.go
index 32dce6a60..0f9e35dc9 100644
--- a/route/route.go
+++ b/route/route.go
@@ -10,6 +10,7 @@ import (
 	"strings"
 
 	"github.com/fabiolb/fabio/metrics"
+	"github.com/gobwas/glob"
 )
 
 // Route maps a path prefix to one or more target URLs.
@@ -36,6 +37,9 @@ type Route struct {
 	// total contains the total number of requests for this route.
 	// Used by the RRPicker
 	total uint64
+
+	// Glob represents compiled pattern.
+	Glob glob.Glob
 }
 
 func (r *Route) addTarget(service string, targetURL *url.URL, fixedWeight float64, tags []string, opts map[string]string) {
diff --git a/route/table.go b/route/table.go
index 6a23021d4..8e74bded1 100644
--- a/route/table.go
+++ b/route/table.go
@@ -13,7 +13,7 @@ import (
 	"sync/atomic"
 
 	"github.com/fabiolb/fabio/metrics"
-	"github.com/ryanuber/go-glob"
+	"github.com/gobwas/glob"
 )
 
 var errInvalidPrefix = errors.New("route: prefix must not be empty")
@@ -153,13 +153,21 @@ func (t Table) addRoute(d *RouteDef) error {
 	switch {
 	// add new host
 	case t[host] == nil:
-		r := &Route{Host: host, Path: path}
+		g, err := glob.Compile(path)
+		if err != nil {
+			return err
+		}
+		r := &Route{Host: host, Path: path, Glob: g}
 		r.addTarget(d.Service, targetURL, d.Weight, d.Tags, d.Opts)
 		t[host] = Routes{r}
 
 	// add new route to existing host
 	case t[host].find(path) == nil:
-		r := &Route{Host: host, Path: path}
+		g, err := glob.Compile(path)
+		if err != nil {
+			return err
+		}
+		r := &Route{Host: host, Path: path, Glob: g}
 		r.addTarget(d.Service, targetURL, d.Weight, d.Tags, d.Opts)
 		t[host] = append(t[host], r)
 		sort.Sort(t[host])
@@ -290,7 +298,8 @@ func (t Table) matchingHosts(req *http.Request) (hosts []string) {
 	host := normalizeHost(req.Host, req.TLS != nil)
 	for pattern := range t {
 		normpat := normalizeHost(pattern, req.TLS != nil)
-		if glob.Glob(normpat, host) {
+		g := glob.MustCompile(normpat)
+		if g.Match(host) {
 			hosts = append(hosts, pattern)
 		}
 	}
diff --git a/vendor/github.com/ryanuber/go-glob/LICENSE b/vendor/github.com/gobwas/glob/LICENSE
similarity index 95%
rename from vendor/github.com/ryanuber/go-glob/LICENSE
rename to vendor/github.com/gobwas/glob/LICENSE
index bdfbd9514..9d4735cad 100644
--- a/vendor/github.com/ryanuber/go-glob/LICENSE
+++ b/vendor/github.com/gobwas/glob/LICENSE
@@ -1,6 +1,6 @@
 The MIT License (MIT)
 
-Copyright (c) 2014 Ryan Uber
+Copyright (c) 2016 Sergey Kamardin
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal
@@ -18,4 +18,4 @@ 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.
+SOFTWARE.
\ No newline at end of file
diff --git a/vendor/github.com/gobwas/glob/bench.sh b/vendor/github.com/gobwas/glob/bench.sh
new file mode 100755
index 000000000..804cf22e6
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/bench.sh
@@ -0,0 +1,26 @@
+#! /bin/bash
+
+bench() {
+    filename="/tmp/$1-$2.bench"
+    if test -e "${filename}";
+    then
+        echo "Already exists ${filename}"
+    else
+        backup=`git rev-parse --abbrev-ref HEAD`
+        git checkout $1
+        echo -n "Creating ${filename}... "
+        go test ./... -run=NONE -bench=$2 > "${filename}" -benchmem
+        echo "OK"
+        git checkout ${backup}
+        sleep 5
+    fi
+}
+
+
+to=$1
+current=`git rev-parse --abbrev-ref HEAD`
+
+bench ${to} $2
+bench ${current} $2
+
+benchcmp $3 "/tmp/${to}-$2.bench" "/tmp/${current}-$2.bench"
diff --git a/vendor/github.com/gobwas/glob/compiler/compiler.go b/vendor/github.com/gobwas/glob/compiler/compiler.go
new file mode 100644
index 000000000..02e7de80a
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/compiler/compiler.go
@@ -0,0 +1,525 @@
+package compiler
+
+// TODO use constructor with all matchers, and to their structs private
+// TODO glue multiple Text nodes (like after QuoteMeta)
+
+import (
+	"fmt"
+	"reflect"
+
+	"github.com/gobwas/glob/match"
+	"github.com/gobwas/glob/syntax/ast"
+	"github.com/gobwas/glob/util/runes"
+)
+
+func optimizeMatcher(matcher match.Matcher) match.Matcher {
+	switch m := matcher.(type) {
+
+	case match.Any:
+		if len(m.Separators) == 0 {
+			return match.NewSuper()
+		}
+
+	case match.AnyOf:
+		if len(m.Matchers) == 1 {
+			return m.Matchers[0]
+		}
+
+		return m
+
+	case match.List:
+		if m.Not == false && len(m.List) == 1 {
+			return match.NewText(string(m.List))
+		}
+
+		return m
+
+	case match.BTree:
+		m.Left = optimizeMatcher(m.Left)
+		m.Right = optimizeMatcher(m.Right)
+
+		r, ok := m.Value.(match.Text)
+		if !ok {
+			return m
+		}
+
+		var (
+			leftNil  = m.Left == nil
+			rightNil = m.Right == nil
+		)
+		if leftNil && rightNil {
+			return match.NewText(r.Str)
+		}
+
+		_, leftSuper := m.Left.(match.Super)
+		lp, leftPrefix := m.Left.(match.Prefix)
+		la, leftAny := m.Left.(match.Any)
+
+		_, rightSuper := m.Right.(match.Super)
+		rs, rightSuffix := m.Right.(match.Suffix)
+		ra, rightAny := m.Right.(match.Any)
+
+		switch {
+		case leftSuper && rightSuper:
+			return match.NewContains(r.Str, false)
+
+		case leftSuper && rightNil:
+			return match.NewSuffix(r.Str)
+
+		case rightSuper && leftNil:
+			return match.NewPrefix(r.Str)
+
+		case leftNil && rightSuffix:
+			return match.NewPrefixSuffix(r.Str, rs.Suffix)
+
+		case rightNil && leftPrefix:
+			return match.NewPrefixSuffix(lp.Prefix, r.Str)
+
+		case rightNil && leftAny:
+			return match.NewSuffixAny(r.Str, la.Separators)
+
+		case leftNil && rightAny:
+			return match.NewPrefixAny(r.Str, ra.Separators)
+		}
+
+		return m
+	}
+
+	return matcher
+}
+
+func compileMatchers(matchers []match.Matcher) (match.Matcher, error) {
+	if len(matchers) == 0 {
+		return nil, fmt.Errorf("compile error: need at least one matcher")
+	}
+	if len(matchers) == 1 {
+		return matchers[0], nil
+	}
+	if m := glueMatchers(matchers); m != nil {
+		return m, nil
+	}
+
+	idx := -1
+	maxLen := -1
+	var val match.Matcher
+	for i, matcher := range matchers {
+		if l := matcher.Len(); l != -1 && l >= maxLen {
+			maxLen = l
+			idx = i
+			val = matcher
+		}
+	}
+
+	if val == nil { // not found matcher with static length
+		r, err := compileMatchers(matchers[1:])
+		if err != nil {
+			return nil, err
+		}
+		return match.NewBTree(matchers[0], nil, r), nil
+	}
+
+	left := matchers[:idx]
+	var right []match.Matcher
+	if len(matchers) > idx+1 {
+		right = matchers[idx+1:]
+	}
+
+	var l, r match.Matcher
+	var err error
+	if len(left) > 0 {
+		l, err = compileMatchers(left)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	if len(right) > 0 {
+		r, err = compileMatchers(right)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	return match.NewBTree(val, l, r), nil
+}
+
+func glueMatchers(matchers []match.Matcher) match.Matcher {
+	if m := glueMatchersAsEvery(matchers); m != nil {
+		return m
+	}
+	if m := glueMatchersAsRow(matchers); m != nil {
+		return m
+	}
+	return nil
+}
+
+func glueMatchersAsRow(matchers []match.Matcher) match.Matcher {
+	if len(matchers) <= 1 {
+		return nil
+	}
+
+	var (
+		c []match.Matcher
+		l int
+	)
+	for _, matcher := range matchers {
+		if ml := matcher.Len(); ml == -1 {
+			return nil
+		} else {
+			c = append(c, matcher)
+			l += ml
+		}
+	}
+	return match.NewRow(l, c...)
+}
+
+func glueMatchersAsEvery(matchers []match.Matcher) match.Matcher {
+	if len(matchers) <= 1 {
+		return nil
+	}
+
+	var (
+		hasAny    bool
+		hasSuper  bool
+		hasSingle bool
+		min       int
+		separator []rune
+	)
+
+	for i, matcher := range matchers {
+		var sep []rune
+
+		switch m := matcher.(type) {
+		case match.Super:
+			sep = []rune{}
+			hasSuper = true
+
+		case match.Any:
+			sep = m.Separators
+			hasAny = true
+
+		case match.Single:
+			sep = m.Separators
+			hasSingle = true
+			min++
+
+		case match.List:
+			if !m.Not {
+				return nil
+			}
+			sep = m.List
+			hasSingle = true
+			min++
+
+		default:
+			return nil
+		}
+
+		// initialize
+		if i == 0 {
+			separator = sep
+		}
+
+		if runes.Equal(sep, separator) {
+			continue
+		}
+
+		return nil
+	}
+
+	if hasSuper && !hasAny && !hasSingle {
+		return match.NewSuper()
+	}
+
+	if hasAny && !hasSuper && !hasSingle {
+		return match.NewAny(separator)
+	}
+
+	if (hasAny || hasSuper) && min > 0 && len(separator) == 0 {
+		return match.NewMin(min)
+	}
+
+	every := match.NewEveryOf()
+
+	if min > 0 {
+		every.Add(match.NewMin(min))
+
+		if !hasAny && !hasSuper {
+			every.Add(match.NewMax(min))
+		}
+	}
+
+	if len(separator) > 0 {
+		every.Add(match.NewContains(string(separator), true))
+	}
+
+	return every
+}
+
+func minimizeMatchers(matchers []match.Matcher) []match.Matcher {
+	var done match.Matcher
+	var left, right, count int
+
+	for l := 0; l < len(matchers); l++ {
+		for r := len(matchers); r > l; r-- {
+			if glued := glueMatchers(matchers[l:r]); glued != nil {
+				var swap bool
+
+				if done == nil {
+					swap = true
+				} else {
+					cl, gl := done.Len(), glued.Len()
+					swap = cl > -1 && gl > -1 && gl > cl
+					swap = swap || count < r-l
+				}
+
+				if swap {
+					done = glued
+					left = l
+					right = r
+					count = r - l
+				}
+			}
+		}
+	}
+
+	if done == nil {
+		return matchers
+	}
+
+	next := append(append([]match.Matcher{}, matchers[:left]...), done)
+	if right < len(matchers) {
+		next = append(next, matchers[right:]...)
+	}
+
+	if len(next) == len(matchers) {
+		return next
+	}
+
+	return minimizeMatchers(next)
+}
+
+// minimizeAnyOf tries to apply some heuristics to minimize number of nodes in given tree
+func minimizeTree(tree *ast.Node) *ast.Node {
+	switch tree.Kind {
+	case ast.KindAnyOf:
+		return minimizeTreeAnyOf(tree)
+	default:
+		return nil
+	}
+}
+
+// minimizeAnyOf tries to find common children of given node of AnyOf pattern
+// it searches for common children from left and from right
+// if any common children are found – then it returns new optimized ast tree
+// else it returns nil
+func minimizeTreeAnyOf(tree *ast.Node) *ast.Node {
+	if !areOfSameKind(tree.Children, ast.KindPattern) {
+		return nil
+	}
+
+	commonLeft, commonRight := commonChildren(tree.Children)
+	commonLeftCount, commonRightCount := len(commonLeft), len(commonRight)
+	if commonLeftCount == 0 && commonRightCount == 0 { // there are no common parts
+		return nil
+	}
+
+	var result []*ast.Node
+	if commonLeftCount > 0 {
+		result = append(result, ast.NewNode(ast.KindPattern, nil, commonLeft...))
+	}
+
+	var anyOf []*ast.Node
+	for _, child := range tree.Children {
+		reuse := child.Children[commonLeftCount : len(child.Children)-commonRightCount]
+		var node *ast.Node
+		if len(reuse) == 0 {
+			// this pattern is completely reduced by commonLeft and commonRight patterns
+			// so it become nothing
+			node = ast.NewNode(ast.KindNothing, nil)
+		} else {
+			node = ast.NewNode(ast.KindPattern, nil, reuse...)
+		}
+		anyOf = appendIfUnique(anyOf, node)
+	}
+	switch {
+	case len(anyOf) == 1 && anyOf[0].Kind != ast.KindNothing:
+		result = append(result, anyOf[0])
+	case len(anyOf) > 1:
+		result = append(result, ast.NewNode(ast.KindAnyOf, nil, anyOf...))
+	}
+
+	if commonRightCount > 0 {
+		result = append(result, ast.NewNode(ast.KindPattern, nil, commonRight...))
+	}
+
+	return ast.NewNode(ast.KindPattern, nil, result...)
+}
+
+func commonChildren(nodes []*ast.Node) (commonLeft, commonRight []*ast.Node) {
+	if len(nodes) <= 1 {
+		return
+	}
+
+	// find node that has least number of children
+	idx := leastChildren(nodes)
+	if idx == -1 {
+		return
+	}
+	tree := nodes[idx]
+	treeLength := len(tree.Children)
+
+	// allocate max able size for rightCommon slice
+	// to get ability insert elements in reverse order (from end to start)
+	// without sorting
+	commonRight = make([]*ast.Node, treeLength)
+	lastRight := treeLength // will use this to get results as commonRight[lastRight:]
+
+	var (
+		breakLeft   bool
+		breakRight  bool
+		commonTotal int
+	)
+	for i, j := 0, treeLength-1; commonTotal < treeLength && j >= 0 && !(breakLeft && breakRight); i, j = i+1, j-1 {
+		treeLeft := tree.Children[i]
+		treeRight := tree.Children[j]
+
+		for k := 0; k < len(nodes) && !(breakLeft && breakRight); k++ {
+			// skip least children node
+			if k == idx {
+				continue
+			}
+
+			restLeft := nodes[k].Children[i]
+			restRight := nodes[k].Children[j+len(nodes[k].Children)-treeLength]
+
+			breakLeft = breakLeft || !treeLeft.Equal(restLeft)
+
+			// disable searching for right common parts, if left part is already overlapping
+			breakRight = breakRight || (!breakLeft && j <= i)
+			breakRight = breakRight || !treeRight.Equal(restRight)
+		}
+
+		if !breakLeft {
+			commonTotal++
+			commonLeft = append(commonLeft, treeLeft)
+		}
+		if !breakRight {
+			commonTotal++
+			lastRight = j
+			commonRight[j] = treeRight
+		}
+	}
+
+	commonRight = commonRight[lastRight:]
+
+	return
+}
+
+func appendIfUnique(target []*ast.Node, val *ast.Node) []*ast.Node {
+	for _, n := range target {
+		if reflect.DeepEqual(n, val) {
+			return target
+		}
+	}
+	return append(target, val)
+}
+
+func areOfSameKind(nodes []*ast.Node, kind ast.Kind) bool {
+	for _, n := range nodes {
+		if n.Kind != kind {
+			return false
+		}
+	}
+	return true
+}
+
+func leastChildren(nodes []*ast.Node) int {
+	min := -1
+	idx := -1
+	for i, n := range nodes {
+		if idx == -1 || (len(n.Children) < min) {
+			min = len(n.Children)
+			idx = i
+		}
+	}
+	return idx
+}
+
+func compileTreeChildren(tree *ast.Node, sep []rune) ([]match.Matcher, error) {
+	var matchers []match.Matcher
+	for _, desc := range tree.Children {
+		m, err := compile(desc, sep)
+		if err != nil {
+			return nil, err
+		}
+		matchers = append(matchers, optimizeMatcher(m))
+	}
+	return matchers, nil
+}
+
+func compile(tree *ast.Node, sep []rune) (m match.Matcher, err error) {
+	switch tree.Kind {
+	case ast.KindAnyOf:
+		// todo this could be faster on pattern_alternatives_combine_lite (see glob_test.go)
+		if n := minimizeTree(tree); n != nil {
+			return compile(n, sep)
+		}
+		matchers, err := compileTreeChildren(tree, sep)
+		if err != nil {
+			return nil, err
+		}
+		return match.NewAnyOf(matchers...), nil
+
+	case ast.KindPattern:
+		if len(tree.Children) == 0 {
+			return match.NewNothing(), nil
+		}
+		matchers, err := compileTreeChildren(tree, sep)
+		if err != nil {
+			return nil, err
+		}
+		m, err = compileMatchers(minimizeMatchers(matchers))
+		if err != nil {
+			return nil, err
+		}
+
+	case ast.KindAny:
+		m = match.NewAny(sep)
+
+	case ast.KindSuper:
+		m = match.NewSuper()
+
+	case ast.KindSingle:
+		m = match.NewSingle(sep)
+
+	case ast.KindNothing:
+		m = match.NewNothing()
+
+	case ast.KindList:
+		l := tree.Value.(ast.List)
+		m = match.NewList([]rune(l.Chars), l.Not)
+
+	case ast.KindRange:
+		r := tree.Value.(ast.Range)
+		m = match.NewRange(r.Lo, r.Hi, r.Not)
+
+	case ast.KindText:
+		t := tree.Value.(ast.Text)
+		m = match.NewText(t.Text)
+
+	default:
+		return nil, fmt.Errorf("could not compile tree: unknown node type")
+	}
+
+	return optimizeMatcher(m), nil
+}
+
+func Compile(tree *ast.Node, sep []rune) (match.Matcher, error) {
+	m, err := compile(tree, sep)
+	if err != nil {
+		return nil, err
+	}
+
+	return m, nil
+}
diff --git a/vendor/github.com/gobwas/glob/glob.go b/vendor/github.com/gobwas/glob/glob.go
new file mode 100644
index 000000000..2afde343a
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/glob.go
@@ -0,0 +1,80 @@
+package glob
+
+import (
+	"github.com/gobwas/glob/compiler"
+	"github.com/gobwas/glob/syntax"
+)
+
+// Glob represents compiled glob pattern.
+type Glob interface {
+	Match(string) bool
+}
+
+// Compile creates Glob for given pattern and strings (if any present after pattern) as separators.
+// The pattern syntax is:
+//
+//    pattern:
+//        { term }
+//
+//    term:
+//        `*`         matches any sequence of non-separator characters
+//        `**`        matches any sequence of characters
+//        `?`         matches any single non-separator character
+//        `[` [ `!` ] { character-range } `]`
+//                    character class (must be non-empty)
+//        `{` pattern-list `}`
+//                    pattern alternatives
+//        c           matches character c (c != `*`, `**`, `?`, `\`, `[`, `{`, `}`)
+//        `\` c       matches character c
+//
+//    character-range:
+//        c           matches character c (c != `\\`, `-`, `]`)
+//        `\` c       matches character c
+//        lo `-` hi   matches character c for lo <= c <= hi
+//
+//    pattern-list:
+//        pattern { `,` pattern }
+//                    comma-separated (without spaces) patterns
+//
+func Compile(pattern string, separators ...rune) (Glob, error) {
+	ast, err := syntax.Parse(pattern)
+	if err != nil {
+		return nil, err
+	}
+
+	matcher, err := compiler.Compile(ast, separators)
+	if err != nil {
+		return nil, err
+	}
+
+	return matcher, nil
+}
+
+// MustCompile is the same as Compile, except that if Compile returns error, this will panic
+func MustCompile(pattern string, separators ...rune) Glob {
+	g, err := Compile(pattern, separators...)
+	if err != nil {
+		panic(err)
+	}
+
+	return g
+}
+
+// QuoteMeta returns a string that quotes all glob pattern meta characters
+// inside the argument text; For example, QuoteMeta(`{foo*}`) returns `\[foo\*\]`.
+func QuoteMeta(s string) string {
+	b := make([]byte, 2*len(s))
+
+	// a byte loop is correct because all meta characters are ASCII
+	j := 0
+	for i := 0; i < len(s); i++ {
+		if syntax.Special(s[i]) {
+			b[j] = '\\'
+			j++
+		}
+		b[j] = s[i]
+		j++
+	}
+
+	return string(b[0:j])
+}
diff --git a/vendor/github.com/gobwas/glob/match/any.go b/vendor/github.com/gobwas/glob/match/any.go
new file mode 100644
index 000000000..514a9a5c4
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/any.go
@@ -0,0 +1,45 @@
+package match
+
+import (
+	"fmt"
+	"github.com/gobwas/glob/util/strings"
+)
+
+type Any struct {
+	Separators []rune
+}
+
+func NewAny(s []rune) Any {
+	return Any{s}
+}
+
+func (self Any) Match(s string) bool {
+	return strings.IndexAnyRunes(s, self.Separators) == -1
+}
+
+func (self Any) Index(s string) (int, []int) {
+	found := strings.IndexAnyRunes(s, self.Separators)
+	switch found {
+	case -1:
+	case 0:
+		return 0, segments0
+	default:
+		s = s[:found]
+	}
+
+	segments := acquireSegments(len(s))
+	for i := range s {
+		segments = append(segments, i)
+	}
+	segments = append(segments, len(s))
+
+	return 0, segments
+}
+
+func (self Any) Len() int {
+	return lenNo
+}
+
+func (self Any) String() string {
+	return fmt.Sprintf("<any:![%s]>", string(self.Separators))
+}
diff --git a/vendor/github.com/gobwas/glob/match/any_of.go b/vendor/github.com/gobwas/glob/match/any_of.go
new file mode 100644
index 000000000..8e65356cd
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/any_of.go
@@ -0,0 +1,82 @@
+package match
+
+import "fmt"
+
+type AnyOf struct {
+	Matchers Matchers
+}
+
+func NewAnyOf(m ...Matcher) AnyOf {
+	return AnyOf{Matchers(m)}
+}
+
+func (self *AnyOf) Add(m Matcher) error {
+	self.Matchers = append(self.Matchers, m)
+	return nil
+}
+
+func (self AnyOf) Match(s string) bool {
+	for _, m := range self.Matchers {
+		if m.Match(s) {
+			return true
+		}
+	}
+
+	return false
+}
+
+func (self AnyOf) Index(s string) (int, []int) {
+	index := -1
+
+	segments := acquireSegments(len(s))
+	for _, m := range self.Matchers {
+		idx, seg := m.Index(s)
+		if idx == -1 {
+			continue
+		}
+
+		if index == -1 || idx < index {
+			index = idx
+			segments = append(segments[:0], seg...)
+			continue
+		}
+
+		if idx > index {
+			continue
+		}
+
+		// here idx == index
+		segments = appendMerge(segments, seg)
+	}
+
+	if index == -1 {
+		releaseSegments(segments)
+		return -1, nil
+	}
+
+	return index, segments
+}
+
+func (self AnyOf) Len() (l int) {
+	l = -1
+	for _, m := range self.Matchers {
+		ml := m.Len()
+		switch {
+		case l == -1:
+			l = ml
+			continue
+
+		case ml == -1:
+			return -1
+
+		case l != ml:
+			return -1
+		}
+	}
+
+	return
+}
+
+func (self AnyOf) String() string {
+	return fmt.Sprintf("<any_of:[%s]>", self.Matchers)
+}
diff --git a/vendor/github.com/gobwas/glob/match/btree.go b/vendor/github.com/gobwas/glob/match/btree.go
new file mode 100644
index 000000000..8302bf824
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/btree.go
@@ -0,0 +1,185 @@
+package match
+
+import (
+	"fmt"
+	"unicode/utf8"
+)
+
+type BTree struct {
+	Value            Matcher
+	Left             Matcher
+	Right            Matcher
+	ValueLengthRunes int
+	LeftLengthRunes  int
+	RightLengthRunes int
+	LengthRunes      int
+}
+
+func NewBTree(Value, Left, Right Matcher) (tree BTree) {
+	tree.Value = Value
+	tree.Left = Left
+	tree.Right = Right
+
+	lenOk := true
+	if tree.ValueLengthRunes = Value.Len(); tree.ValueLengthRunes == -1 {
+		lenOk = false
+	}
+
+	if Left != nil {
+		if tree.LeftLengthRunes = Left.Len(); tree.LeftLengthRunes == -1 {
+			lenOk = false
+		}
+	}
+
+	if Right != nil {
+		if tree.RightLengthRunes = Right.Len(); tree.RightLengthRunes == -1 {
+			lenOk = false
+		}
+	}
+
+	if lenOk {
+		tree.LengthRunes = tree.LeftLengthRunes + tree.ValueLengthRunes + tree.RightLengthRunes
+	} else {
+		tree.LengthRunes = -1
+	}
+
+	return tree
+}
+
+func (self BTree) Len() int {
+	return self.LengthRunes
+}
+
+// todo?
+func (self BTree) Index(s string) (index int, segments []int) {
+	//inputLen := len(s)
+	//// try to cut unnecessary parts
+	//// by knowledge of length of right and left part
+	//offset, limit := self.offsetLimit(inputLen)
+	//for offset < limit {
+	//	// search for matching part in substring
+	//	vi, segments := self.Value.Index(s[offset:limit])
+	//	if index == -1 {
+	//		return -1, nil
+	//	}
+	//	if self.Left == nil {
+	//		if index != offset {
+	//			return -1, nil
+	//		}
+	//	} else {
+	//		left := s[:offset+vi]
+	//		i := self.Left.IndexSuffix(left)
+	//		if i == -1 {
+	//			return -1, nil
+	//		}
+	//		index = i
+	//	}
+	//	if self.Right != nil {
+	//		for _, seg := range segments {
+	//			right := s[:offset+vi+seg]
+	//		}
+	//	}
+
+	//	l := s[:offset+index]
+	//	var left bool
+	//	if self.Left != nil {
+	//		left = self.Left.Index(l)
+	//	} else {
+	//		left = l == ""
+	//	}
+	//}
+
+	return -1, nil
+}
+
+func (self BTree) Match(s string) bool {
+	inputLen := len(s)
+	// try to cut unnecessary parts
+	// by knowledge of length of right and left part
+	offset, limit := self.offsetLimit(inputLen)
+
+	for offset < limit {
+		// search for matching part in substring
+		index, segments := self.Value.Index(s[offset:limit])
+		if index == -1 {
+			releaseSegments(segments)
+			return false
+		}
+
+		l := s[:offset+index]
+		var left bool
+		if self.Left != nil {
+			left = self.Left.Match(l)
+		} else {
+			left = l == ""
+		}
+
+		if left {
+			for i := len(segments) - 1; i >= 0; i-- {
+				length := segments[i]
+
+				var right bool
+				var r string
+				// if there is no string for the right branch
+				if inputLen <= offset+index+length {
+					r = ""
+				} else {
+					r = s[offset+index+length:]
+				}
+
+				if self.Right != nil {
+					right = self.Right.Match(r)
+				} else {
+					right = r == ""
+				}
+
+				if right {
+					releaseSegments(segments)
+					return true
+				}
+			}
+		}
+
+		_, step := utf8.DecodeRuneInString(s[offset+index:])
+		offset += index + step
+
+		releaseSegments(segments)
+	}
+
+	return false
+}
+
+func (self BTree) offsetLimit(inputLen int) (offset int, limit int) {
+	// self.Length, self.RLen and self.LLen are values meaning the length of runes for each part
+	// here we manipulating byte length for better optimizations
+	// but these checks still works, cause minLen of 1-rune string is 1 byte.
+	if self.LengthRunes != -1 && self.LengthRunes > inputLen {
+		return 0, 0
+	}
+	if self.LeftLengthRunes >= 0 {
+		offset = self.LeftLengthRunes
+	}
+	if self.RightLengthRunes >= 0 {
+		limit = inputLen - self.RightLengthRunes
+	} else {
+		limit = inputLen
+	}
+	return offset, limit
+}
+
+func (self BTree) String() string {
+	const n string = "<nil>"
+	var l, r string
+	if self.Left == nil {
+		l = n
+	} else {
+		l = self.Left.String()
+	}
+	if self.Right == nil {
+		r = n
+	} else {
+		r = self.Right.String()
+	}
+
+	return fmt.Sprintf("<btree:[%s<-%s->%s]>", l, self.Value, r)
+}
diff --git a/vendor/github.com/gobwas/glob/match/contains.go b/vendor/github.com/gobwas/glob/match/contains.go
new file mode 100644
index 000000000..0998e95b0
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/contains.go
@@ -0,0 +1,58 @@
+package match
+
+import (
+	"fmt"
+	"strings"
+)
+
+type Contains struct {
+	Needle string
+	Not    bool
+}
+
+func NewContains(needle string, not bool) Contains {
+	return Contains{needle, not}
+}
+
+func (self Contains) Match(s string) bool {
+	return strings.Contains(s, self.Needle) != self.Not
+}
+
+func (self Contains) Index(s string) (int, []int) {
+	var offset int
+
+	idx := strings.Index(s, self.Needle)
+
+	if !self.Not {
+		if idx == -1 {
+			return -1, nil
+		}
+
+		offset = idx + len(self.Needle)
+		if len(s) <= offset {
+			return 0, []int{offset}
+		}
+		s = s[offset:]
+	} else if idx != -1 {
+		s = s[:idx]
+	}
+
+	segments := acquireSegments(len(s) + 1)
+	for i := range s {
+		segments = append(segments, offset+i)
+	}
+
+	return 0, append(segments, offset+len(s))
+}
+
+func (self Contains) Len() int {
+	return lenNo
+}
+
+func (self Contains) String() string {
+	var not string
+	if self.Not {
+		not = "!"
+	}
+	return fmt.Sprintf("<contains:%s[%s]>", not, self.Needle)
+}
diff --git a/vendor/github.com/gobwas/glob/match/every_of.go b/vendor/github.com/gobwas/glob/match/every_of.go
new file mode 100644
index 000000000..7c968ee36
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/every_of.go
@@ -0,0 +1,99 @@
+package match
+
+import (
+	"fmt"
+)
+
+type EveryOf struct {
+	Matchers Matchers
+}
+
+func NewEveryOf(m ...Matcher) EveryOf {
+	return EveryOf{Matchers(m)}
+}
+
+func (self *EveryOf) Add(m Matcher) error {
+	self.Matchers = append(self.Matchers, m)
+	return nil
+}
+
+func (self EveryOf) Len() (l int) {
+	for _, m := range self.Matchers {
+		if ml := m.Len(); l > 0 {
+			l += ml
+		} else {
+			return -1
+		}
+	}
+
+	return
+}
+
+func (self EveryOf) Index(s string) (int, []int) {
+	var index int
+	var offset int
+
+	// make `in` with cap as len(s),
+	// cause it is the maximum size of output segments values
+	next := acquireSegments(len(s))
+	current := acquireSegments(len(s))
+
+	sub := s
+	for i, m := range self.Matchers {
+		idx, seg := m.Index(sub)
+		if idx == -1 {
+			releaseSegments(next)
+			releaseSegments(current)
+			return -1, nil
+		}
+
+		if i == 0 {
+			// we use copy here instead of `current = seg`
+			// cause seg is a slice from reusable buffer `in`
+			// and it could be overwritten in next iteration
+			current = append(current, seg...)
+		} else {
+			// clear the next
+			next = next[:0]
+
+			delta := index - (idx + offset)
+			for _, ex := range current {
+				for _, n := range seg {
+					if ex+delta == n {
+						next = append(next, n)
+					}
+				}
+			}
+
+			if len(next) == 0 {
+				releaseSegments(next)
+				releaseSegments(current)
+				return -1, nil
+			}
+
+			current = append(current[:0], next...)
+		}
+
+		index = idx + offset
+		sub = s[index:]
+		offset += idx
+	}
+
+	releaseSegments(next)
+
+	return index, current
+}
+
+func (self EveryOf) Match(s string) bool {
+	for _, m := range self.Matchers {
+		if !m.Match(s) {
+			return false
+		}
+	}
+
+	return true
+}
+
+func (self EveryOf) String() string {
+	return fmt.Sprintf("<every_of:[%s]>", self.Matchers)
+}
diff --git a/vendor/github.com/gobwas/glob/match/list.go b/vendor/github.com/gobwas/glob/match/list.go
new file mode 100644
index 000000000..7fd763ecd
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/list.go
@@ -0,0 +1,49 @@
+package match
+
+import (
+	"fmt"
+	"github.com/gobwas/glob/util/runes"
+	"unicode/utf8"
+)
+
+type List struct {
+	List []rune
+	Not  bool
+}
+
+func NewList(list []rune, not bool) List {
+	return List{list, not}
+}
+
+func (self List) Match(s string) bool {
+	r, w := utf8.DecodeRuneInString(s)
+	if len(s) > w {
+		return false
+	}
+
+	inList := runes.IndexRune(self.List, r) != -1
+	return inList == !self.Not
+}
+
+func (self List) Len() int {
+	return lenOne
+}
+
+func (self List) Index(s string) (int, []int) {
+	for i, r := range s {
+		if self.Not == (runes.IndexRune(self.List, r) == -1) {
+			return i, segmentsByRuneLength[utf8.RuneLen(r)]
+		}
+	}
+
+	return -1, nil
+}
+
+func (self List) String() string {
+	var not string
+	if self.Not {
+		not = "!"
+	}
+
+	return fmt.Sprintf("<list:%s[%s]>", not, string(self.List))
+}
diff --git a/vendor/github.com/gobwas/glob/match/match.go b/vendor/github.com/gobwas/glob/match/match.go
new file mode 100644
index 000000000..f80e007fb
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/match.go
@@ -0,0 +1,81 @@
+package match
+
+// todo common table of rune's length
+
+import (
+	"fmt"
+	"strings"
+)
+
+const lenOne = 1
+const lenZero = 0
+const lenNo = -1
+
+type Matcher interface {
+	Match(string) bool
+	Index(string) (int, []int)
+	Len() int
+	String() string
+}
+
+type Matchers []Matcher
+
+func (m Matchers) String() string {
+	var s []string
+	for _, matcher := range m {
+		s = append(s, fmt.Sprint(matcher))
+	}
+
+	return fmt.Sprintf("%s", strings.Join(s, ","))
+}
+
+// appendMerge merges and sorts given already SORTED and UNIQUE segments.
+func appendMerge(target, sub []int) []int {
+	lt, ls := len(target), len(sub)
+	out := make([]int, 0, lt+ls)
+
+	for x, y := 0, 0; x < lt || y < ls; {
+		if x >= lt {
+			out = append(out, sub[y:]...)
+			break
+		}
+
+		if y >= ls {
+			out = append(out, target[x:]...)
+			break
+		}
+
+		xValue := target[x]
+		yValue := sub[y]
+
+		switch {
+
+		case xValue == yValue:
+			out = append(out, xValue)
+			x++
+			y++
+
+		case xValue < yValue:
+			out = append(out, xValue)
+			x++
+
+		case yValue < xValue:
+			out = append(out, yValue)
+			y++
+
+		}
+	}
+
+	target = append(target[:0], out...)
+
+	return target
+}
+
+func reverseSegments(input []int) {
+	l := len(input)
+	m := l / 2
+
+	for i := 0; i < m; i++ {
+		input[i], input[l-i-1] = input[l-i-1], input[i]
+	}
+}
diff --git a/vendor/github.com/gobwas/glob/match/max.go b/vendor/github.com/gobwas/glob/match/max.go
new file mode 100644
index 000000000..d72f69eff
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/max.go
@@ -0,0 +1,49 @@
+package match
+
+import (
+	"fmt"
+	"unicode/utf8"
+)
+
+type Max struct {
+	Limit int
+}
+
+func NewMax(l int) Max {
+	return Max{l}
+}
+
+func (self Max) Match(s string) bool {
+	var l int
+	for range s {
+		l += 1
+		if l > self.Limit {
+			return false
+		}
+	}
+
+	return true
+}
+
+func (self Max) Index(s string) (int, []int) {
+	segments := acquireSegments(self.Limit + 1)
+	segments = append(segments, 0)
+	var count int
+	for i, r := range s {
+		count++
+		if count > self.Limit {
+			break
+		}
+		segments = append(segments, i+utf8.RuneLen(r))
+	}
+
+	return 0, segments
+}
+
+func (self Max) Len() int {
+	return lenNo
+}
+
+func (self Max) String() string {
+	return fmt.Sprintf("<max:%d>", self.Limit)
+}
diff --git a/vendor/github.com/gobwas/glob/match/min.go b/vendor/github.com/gobwas/glob/match/min.go
new file mode 100644
index 000000000..db57ac8eb
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/min.go
@@ -0,0 +1,57 @@
+package match
+
+import (
+	"fmt"
+	"unicode/utf8"
+)
+
+type Min struct {
+	Limit int
+}
+
+func NewMin(l int) Min {
+	return Min{l}
+}
+
+func (self Min) Match(s string) bool {
+	var l int
+	for range s {
+		l += 1
+		if l >= self.Limit {
+			return true
+		}
+	}
+
+	return false
+}
+
+func (self Min) Index(s string) (int, []int) {
+	var count int
+
+	c := len(s) - self.Limit + 1
+	if c <= 0 {
+		return -1, nil
+	}
+
+	segments := acquireSegments(c)
+	for i, r := range s {
+		count++
+		if count >= self.Limit {
+			segments = append(segments, i+utf8.RuneLen(r))
+		}
+	}
+
+	if len(segments) == 0 {
+		return -1, nil
+	}
+
+	return 0, segments
+}
+
+func (self Min) Len() int {
+	return lenNo
+}
+
+func (self Min) String() string {
+	return fmt.Sprintf("<min:%d>", self.Limit)
+}
diff --git a/vendor/github.com/gobwas/glob/match/nothing.go b/vendor/github.com/gobwas/glob/match/nothing.go
new file mode 100644
index 000000000..0d4ecd36b
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/nothing.go
@@ -0,0 +1,27 @@
+package match
+
+import (
+	"fmt"
+)
+
+type Nothing struct{}
+
+func NewNothing() Nothing {
+	return Nothing{}
+}
+
+func (self Nothing) Match(s string) bool {
+	return len(s) == 0
+}
+
+func (self Nothing) Index(s string) (int, []int) {
+	return 0, segments0
+}
+
+func (self Nothing) Len() int {
+	return lenZero
+}
+
+func (self Nothing) String() string {
+	return fmt.Sprintf("<nothing>")
+}
diff --git a/vendor/github.com/gobwas/glob/match/prefix.go b/vendor/github.com/gobwas/glob/match/prefix.go
new file mode 100644
index 000000000..a7347250e
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/prefix.go
@@ -0,0 +1,50 @@
+package match
+
+import (
+	"fmt"
+	"strings"
+	"unicode/utf8"
+)
+
+type Prefix struct {
+	Prefix string
+}
+
+func NewPrefix(p string) Prefix {
+	return Prefix{p}
+}
+
+func (self Prefix) Index(s string) (int, []int) {
+	idx := strings.Index(s, self.Prefix)
+	if idx == -1 {
+		return -1, nil
+	}
+
+	length := len(self.Prefix)
+	var sub string
+	if len(s) > idx+length {
+		sub = s[idx+length:]
+	} else {
+		sub = ""
+	}
+
+	segments := acquireSegments(len(sub) + 1)
+	segments = append(segments, length)
+	for i, r := range sub {
+		segments = append(segments, length+i+utf8.RuneLen(r))
+	}
+
+	return idx, segments
+}
+
+func (self Prefix) Len() int {
+	return lenNo
+}
+
+func (self Prefix) Match(s string) bool {
+	return strings.HasPrefix(s, self.Prefix)
+}
+
+func (self Prefix) String() string {
+	return fmt.Sprintf("<prefix:%s>", self.Prefix)
+}
diff --git a/vendor/github.com/gobwas/glob/match/prefix_any.go b/vendor/github.com/gobwas/glob/match/prefix_any.go
new file mode 100644
index 000000000..8ee58fe1b
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/prefix_any.go
@@ -0,0 +1,55 @@
+package match
+
+import (
+	"fmt"
+	"strings"
+	"unicode/utf8"
+
+	sutil "github.com/gobwas/glob/util/strings"
+)
+
+type PrefixAny struct {
+	Prefix     string
+	Separators []rune
+}
+
+func NewPrefixAny(s string, sep []rune) PrefixAny {
+	return PrefixAny{s, sep}
+}
+
+func (self PrefixAny) Index(s string) (int, []int) {
+	idx := strings.Index(s, self.Prefix)
+	if idx == -1 {
+		return -1, nil
+	}
+
+	n := len(self.Prefix)
+	sub := s[idx+n:]
+	i := sutil.IndexAnyRunes(sub, self.Separators)
+	if i > -1 {
+		sub = sub[:i]
+	}
+
+	seg := acquireSegments(len(sub) + 1)
+	seg = append(seg, n)
+	for i, r := range sub {
+		seg = append(seg, n+i+utf8.RuneLen(r))
+	}
+
+	return idx, seg
+}
+
+func (self PrefixAny) Len() int {
+	return lenNo
+}
+
+func (self PrefixAny) Match(s string) bool {
+	if !strings.HasPrefix(s, self.Prefix) {
+		return false
+	}
+	return sutil.IndexAnyRunes(s[len(self.Prefix):], self.Separators) == -1
+}
+
+func (self PrefixAny) String() string {
+	return fmt.Sprintf("<prefix_any:%s![%s]>", self.Prefix, string(self.Separators))
+}
diff --git a/vendor/github.com/gobwas/glob/match/prefix_suffix.go b/vendor/github.com/gobwas/glob/match/prefix_suffix.go
new file mode 100644
index 000000000..8208085a1
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/prefix_suffix.go
@@ -0,0 +1,62 @@
+package match
+
+import (
+	"fmt"
+	"strings"
+)
+
+type PrefixSuffix struct {
+	Prefix, Suffix string
+}
+
+func NewPrefixSuffix(p, s string) PrefixSuffix {
+	return PrefixSuffix{p, s}
+}
+
+func (self PrefixSuffix) Index(s string) (int, []int) {
+	prefixIdx := strings.Index(s, self.Prefix)
+	if prefixIdx == -1 {
+		return -1, nil
+	}
+
+	suffixLen := len(self.Suffix)
+	if suffixLen <= 0 {
+		return prefixIdx, []int{len(s) - prefixIdx}
+	}
+
+	if (len(s) - prefixIdx) <= 0 {
+		return -1, nil
+	}
+
+	segments := acquireSegments(len(s) - prefixIdx)
+	for sub := s[prefixIdx:]; ; {
+		suffixIdx := strings.LastIndex(sub, self.Suffix)
+		if suffixIdx == -1 {
+			break
+		}
+
+		segments = append(segments, suffixIdx+suffixLen)
+		sub = sub[:suffixIdx]
+	}
+
+	if len(segments) == 0 {
+		releaseSegments(segments)
+		return -1, nil
+	}
+
+	reverseSegments(segments)
+
+	return prefixIdx, segments
+}
+
+func (self PrefixSuffix) Len() int {
+	return lenNo
+}
+
+func (self PrefixSuffix) Match(s string) bool {
+	return strings.HasPrefix(s, self.Prefix) && strings.HasSuffix(s, self.Suffix)
+}
+
+func (self PrefixSuffix) String() string {
+	return fmt.Sprintf("<prefix_suffix:[%s,%s]>", self.Prefix, self.Suffix)
+}
diff --git a/vendor/github.com/gobwas/glob/match/range.go b/vendor/github.com/gobwas/glob/match/range.go
new file mode 100644
index 000000000..ce30245a4
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/range.go
@@ -0,0 +1,48 @@
+package match
+
+import (
+	"fmt"
+	"unicode/utf8"
+)
+
+type Range struct {
+	Lo, Hi rune
+	Not    bool
+}
+
+func NewRange(lo, hi rune, not bool) Range {
+	return Range{lo, hi, not}
+}
+
+func (self Range) Len() int {
+	return lenOne
+}
+
+func (self Range) Match(s string) bool {
+	r, w := utf8.DecodeRuneInString(s)
+	if len(s) > w {
+		return false
+	}
+
+	inRange := r >= self.Lo && r <= self.Hi
+
+	return inRange == !self.Not
+}
+
+func (self Range) Index(s string) (int, []int) {
+	for i, r := range s {
+		if self.Not != (r >= self.Lo && r <= self.Hi) {
+			return i, segmentsByRuneLength[utf8.RuneLen(r)]
+		}
+	}
+
+	return -1, nil
+}
+
+func (self Range) String() string {
+	var not string
+	if self.Not {
+		not = "!"
+	}
+	return fmt.Sprintf("<range:%s[%s,%s]>", not, string(self.Lo), string(self.Hi))
+}
diff --git a/vendor/github.com/gobwas/glob/match/row.go b/vendor/github.com/gobwas/glob/match/row.go
new file mode 100644
index 000000000..4379042e4
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/row.go
@@ -0,0 +1,77 @@
+package match
+
+import (
+	"fmt"
+)
+
+type Row struct {
+	Matchers    Matchers
+	RunesLength int
+	Segments    []int
+}
+
+func NewRow(len int, m ...Matcher) Row {
+	return Row{
+		Matchers:    Matchers(m),
+		RunesLength: len,
+		Segments:    []int{len},
+	}
+}
+
+func (self Row) matchAll(s string) bool {
+	var idx int
+	for _, m := range self.Matchers {
+		length := m.Len()
+
+		var next, i int
+		for next = range s[idx:] {
+			i++
+			if i == length {
+				break
+			}
+		}
+
+		if i < length || !m.Match(s[idx:idx+next+1]) {
+			return false
+		}
+
+		idx += next + 1
+	}
+
+	return true
+}
+
+func (self Row) lenOk(s string) bool {
+	var i int
+	for range s {
+		i++
+		if i > self.RunesLength {
+			return false
+		}
+	}
+	return self.RunesLength == i
+}
+
+func (self Row) Match(s string) bool {
+	return self.lenOk(s) && self.matchAll(s)
+}
+
+func (self Row) Len() (l int) {
+	return self.RunesLength
+}
+
+func (self Row) Index(s string) (int, []int) {
+	for i := range s {
+		if len(s[i:]) < self.RunesLength {
+			break
+		}
+		if self.matchAll(s[i:]) {
+			return i, self.Segments
+		}
+	}
+	return -1, nil
+}
+
+func (self Row) String() string {
+	return fmt.Sprintf("<row_%d:[%s]>", self.RunesLength, self.Matchers)
+}
diff --git a/vendor/github.com/gobwas/glob/match/segments.go b/vendor/github.com/gobwas/glob/match/segments.go
new file mode 100644
index 000000000..9ea6f3094
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/segments.go
@@ -0,0 +1,91 @@
+package match
+
+import (
+	"sync"
+)
+
+type SomePool interface {
+	Get() []int
+	Put([]int)
+}
+
+var segmentsPools [1024]sync.Pool
+
+func toPowerOfTwo(v int) int {
+	v--
+	v |= v >> 1
+	v |= v >> 2
+	v |= v >> 4
+	v |= v >> 8
+	v |= v >> 16
+	v++
+
+	return v
+}
+
+const (
+	cacheFrom             = 16
+	cacheToAndHigher      = 1024
+	cacheFromIndex        = 15
+	cacheToAndHigherIndex = 1023
+)
+
+var (
+	segments0 = []int{0}
+	segments1 = []int{1}
+	segments2 = []int{2}
+	segments3 = []int{3}
+	segments4 = []int{4}
+)
+
+var segmentsByRuneLength [5][]int = [5][]int{
+	0: segments0,
+	1: segments1,
+	2: segments2,
+	3: segments3,
+	4: segments4,
+}
+
+func init() {
+	for i := cacheToAndHigher; i >= cacheFrom; i >>= 1 {
+		func(i int) {
+			segmentsPools[i-1] = sync.Pool{New: func() interface{} {
+				return make([]int, 0, i)
+			}}
+		}(i)
+	}
+}
+
+func getTableIndex(c int) int {
+	p := toPowerOfTwo(c)
+	switch {
+	case p >= cacheToAndHigher:
+		return cacheToAndHigherIndex
+	case p <= cacheFrom:
+		return cacheFromIndex
+	default:
+		return p - 1
+	}
+}
+
+func acquireSegments(c int) []int {
+	// make []int with less capacity than cacheFrom
+	// is faster than acquiring it from pool
+	if c < cacheFrom {
+		return make([]int, 0, c)
+	}
+
+	return segmentsPools[getTableIndex(c)].Get().([]int)[:0]
+}
+
+func releaseSegments(s []int) {
+	c := cap(s)
+
+	// make []int with less capacity than cacheFrom
+	// is faster than acquiring it from pool
+	if c < cacheFrom {
+		return
+	}
+
+	segmentsPools[getTableIndex(c)].Put(s)
+}
diff --git a/vendor/github.com/gobwas/glob/match/single.go b/vendor/github.com/gobwas/glob/match/single.go
new file mode 100644
index 000000000..ee6e3954c
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/single.go
@@ -0,0 +1,43 @@
+package match
+
+import (
+	"fmt"
+	"github.com/gobwas/glob/util/runes"
+	"unicode/utf8"
+)
+
+// single represents ?
+type Single struct {
+	Separators []rune
+}
+
+func NewSingle(s []rune) Single {
+	return Single{s}
+}
+
+func (self Single) Match(s string) bool {
+	r, w := utf8.DecodeRuneInString(s)
+	if len(s) > w {
+		return false
+	}
+
+	return runes.IndexRune(self.Separators, r) == -1
+}
+
+func (self Single) Len() int {
+	return lenOne
+}
+
+func (self Single) Index(s string) (int, []int) {
+	for i, r := range s {
+		if runes.IndexRune(self.Separators, r) == -1 {
+			return i, segmentsByRuneLength[utf8.RuneLen(r)]
+		}
+	}
+
+	return -1, nil
+}
+
+func (self Single) String() string {
+	return fmt.Sprintf("<single:![%s]>", string(self.Separators))
+}
diff --git a/vendor/github.com/gobwas/glob/match/suffix.go b/vendor/github.com/gobwas/glob/match/suffix.go
new file mode 100644
index 000000000..85bea8c68
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/suffix.go
@@ -0,0 +1,35 @@
+package match
+
+import (
+	"fmt"
+	"strings"
+)
+
+type Suffix struct {
+	Suffix string
+}
+
+func NewSuffix(s string) Suffix {
+	return Suffix{s}
+}
+
+func (self Suffix) Len() int {
+	return lenNo
+}
+
+func (self Suffix) Match(s string) bool {
+	return strings.HasSuffix(s, self.Suffix)
+}
+
+func (self Suffix) Index(s string) (int, []int) {
+	idx := strings.Index(s, self.Suffix)
+	if idx == -1 {
+		return -1, nil
+	}
+
+	return 0, []int{idx + len(self.Suffix)}
+}
+
+func (self Suffix) String() string {
+	return fmt.Sprintf("<suffix:%s>", self.Suffix)
+}
diff --git a/vendor/github.com/gobwas/glob/match/suffix_any.go b/vendor/github.com/gobwas/glob/match/suffix_any.go
new file mode 100644
index 000000000..c5106f819
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/suffix_any.go
@@ -0,0 +1,43 @@
+package match
+
+import (
+	"fmt"
+	"strings"
+
+	sutil "github.com/gobwas/glob/util/strings"
+)
+
+type SuffixAny struct {
+	Suffix     string
+	Separators []rune
+}
+
+func NewSuffixAny(s string, sep []rune) SuffixAny {
+	return SuffixAny{s, sep}
+}
+
+func (self SuffixAny) Index(s string) (int, []int) {
+	idx := strings.Index(s, self.Suffix)
+	if idx == -1 {
+		return -1, nil
+	}
+
+	i := sutil.LastIndexAnyRunes(s[:idx], self.Separators) + 1
+
+	return i, []int{idx + len(self.Suffix) - i}
+}
+
+func (self SuffixAny) Len() int {
+	return lenNo
+}
+
+func (self SuffixAny) Match(s string) bool {
+	if !strings.HasSuffix(s, self.Suffix) {
+		return false
+	}
+	return sutil.IndexAnyRunes(s[:len(s)-len(self.Suffix)], self.Separators) == -1
+}
+
+func (self SuffixAny) String() string {
+	return fmt.Sprintf("<suffix_any:![%s]%s>", string(self.Separators), self.Suffix)
+}
diff --git a/vendor/github.com/gobwas/glob/match/super.go b/vendor/github.com/gobwas/glob/match/super.go
new file mode 100644
index 000000000..3875950bb
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/super.go
@@ -0,0 +1,33 @@
+package match
+
+import (
+	"fmt"
+)
+
+type Super struct{}
+
+func NewSuper() Super {
+	return Super{}
+}
+
+func (self Super) Match(s string) bool {
+	return true
+}
+
+func (self Super) Len() int {
+	return lenNo
+}
+
+func (self Super) Index(s string) (int, []int) {
+	segments := acquireSegments(len(s) + 1)
+	for i := range s {
+		segments = append(segments, i)
+	}
+	segments = append(segments, len(s))
+
+	return 0, segments
+}
+
+func (self Super) String() string {
+	return fmt.Sprintf("<super>")
+}
diff --git a/vendor/github.com/gobwas/glob/match/text.go b/vendor/github.com/gobwas/glob/match/text.go
new file mode 100644
index 000000000..0a17616d3
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/match/text.go
@@ -0,0 +1,45 @@
+package match
+
+import (
+	"fmt"
+	"strings"
+	"unicode/utf8"
+)
+
+// raw represents raw string to match
+type Text struct {
+	Str         string
+	RunesLength int
+	BytesLength int
+	Segments    []int
+}
+
+func NewText(s string) Text {
+	return Text{
+		Str:         s,
+		RunesLength: utf8.RuneCountInString(s),
+		BytesLength: len(s),
+		Segments:    []int{len(s)},
+	}
+}
+
+func (self Text) Match(s string) bool {
+	return self.Str == s
+}
+
+func (self Text) Len() int {
+	return self.RunesLength
+}
+
+func (self Text) Index(s string) (int, []int) {
+	index := strings.Index(s, self.Str)
+	if index == -1 {
+		return -1, nil
+	}
+
+	return index, self.Segments
+}
+
+func (self Text) String() string {
+	return fmt.Sprintf("<text:`%v`>", self.Str)
+}
diff --git a/vendor/github.com/gobwas/glob/readme.md b/vendor/github.com/gobwas/glob/readme.md
new file mode 100644
index 000000000..f58144e73
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/readme.md
@@ -0,0 +1,148 @@
+# glob.[go](https://golang.org)
+
+[![GoDoc][godoc-image]][godoc-url] [![Build Status][travis-image]][travis-url]
+
+> Go Globbing Library.
+
+## Install
+
+```shell
+    go get github.com/gobwas/glob
+```
+
+## Example
+
+```go
+
+package main
+
+import "github.com/gobwas/glob"
+
+func main() {
+    var g glob.Glob
+    
+    // create simple glob
+    g = glob.MustCompile("*.github.com")
+    g.Match("api.github.com") // true
+    
+    // quote meta characters and then create simple glob 
+    g = glob.MustCompile(glob.QuoteMeta("*.github.com"))
+    g.Match("*.github.com") // true
+    
+    // create new glob with set of delimiters as ["."]
+    g = glob.MustCompile("api.*.com", '.')
+    g.Match("api.github.com") // true
+    g.Match("api.gi.hub.com") // false
+    
+    // create new glob with set of delimiters as ["."]
+    // but now with super wildcard
+    g = glob.MustCompile("api.**.com", '.')
+    g.Match("api.github.com") // true
+    g.Match("api.gi.hub.com") // true
+        
+    // create glob with single symbol wildcard
+    g = glob.MustCompile("?at")
+    g.Match("cat") // true
+    g.Match("fat") // true
+    g.Match("at") // false
+    
+    // create glob with single symbol wildcard and delimiters ['f']
+    g = glob.MustCompile("?at", 'f')
+    g.Match("cat") // true
+    g.Match("fat") // false
+    g.Match("at") // false 
+    
+    // create glob with character-list matchers 
+    g = glob.MustCompile("[abc]at")
+    g.Match("cat") // true
+    g.Match("bat") // true
+    g.Match("fat") // false
+    g.Match("at") // false
+    
+    // create glob with character-list matchers 
+    g = glob.MustCompile("[!abc]at")
+    g.Match("cat") // false
+    g.Match("bat") // false
+    g.Match("fat") // true
+    g.Match("at") // false 
+    
+    // create glob with character-range matchers 
+    g = glob.MustCompile("[a-c]at")
+    g.Match("cat") // true
+    g.Match("bat") // true
+    g.Match("fat") // false
+    g.Match("at") // false
+    
+    // create glob with character-range matchers 
+    g = glob.MustCompile("[!a-c]at")
+    g.Match("cat") // false
+    g.Match("bat") // false
+    g.Match("fat") // true
+    g.Match("at") // false 
+    
+    // create glob with pattern-alternatives list 
+    g = glob.MustCompile("{cat,bat,[fr]at}")
+    g.Match("cat") // true
+    g.Match("bat") // true
+    g.Match("fat") // true
+    g.Match("rat") // true
+    g.Match("at") // false 
+    g.Match("zat") // false 
+}
+
+```
+
+## Performance
+
+This library is created for compile-once patterns. This means, that compilation could take time, but 
+strings matching is done faster, than in case when always parsing template.
+
+If you will not use compiled `glob.Glob` object, and do `g := glob.MustCompile(pattern); g.Match(...)` every time, then your code will be much more slower.
+
+Run `go test -bench=.` from source root to see the benchmarks:
+
+Pattern | Fixture | Match | Speed (ns/op)
+--------|---------|-------|--------------
+`[a-z][!a-x]*cat*[h][!b]*eyes*` | `my cat has very bright eyes` | `true` | 432
+`[a-z][!a-x]*cat*[h][!b]*eyes*` | `my dog has very bright eyes` | `false` | 199
+`https://*.google.*` | `https://account.google.com` | `true` | 96
+`https://*.google.*` | `https://google.com` | `false` | 66
+`{https://*.google.*,*yandex.*,*yahoo.*,*mail.ru}` | `http://yahoo.com` | `true` | 163
+`{https://*.google.*,*yandex.*,*yahoo.*,*mail.ru}` | `http://google.com` | `false` | 197
+`{https://*gobwas.com,http://exclude.gobwas.com}` | `https://safe.gobwas.com` | `true` | 22
+`{https://*gobwas.com,http://exclude.gobwas.com}` | `http://safe.gobwas.com` | `false` | 24
+`abc*` | `abcdef` | `true` | 8.15
+`abc*` | `af` | `false` | 5.68
+`*def` | `abcdef` | `true` | 8.84
+`*def` | `af` | `false` | 5.74
+`ab*ef` | `abcdef` | `true` | 15.2
+`ab*ef` | `af` | `false` | 10.4
+
+The same things with `regexp` package:
+
+Pattern | Fixture | Match | Speed (ns/op)
+--------|---------|-------|--------------
+`^[a-z][^a-x].*cat.*[h][^b].*eyes.*$` | `my cat has very bright eyes` | `true` | 2553
+`^[a-z][^a-x].*cat.*[h][^b].*eyes.*$` | `my dog has very bright eyes` | `false` | 1383
+`^https:\/\/.*\.google\..*$` | `https://account.google.com` | `true` | 1205
+`^https:\/\/.*\.google\..*$` | `https://google.com` | `false` | 767
+`^(https:\/\/.*\.google\..*|.*yandex\..*|.*yahoo\..*|.*mail\.ru)$` | `http://yahoo.com` | `true` | 1435
+`^(https:\/\/.*\.google\..*|.*yandex\..*|.*yahoo\..*|.*mail\.ru)$` | `http://google.com` | `false` | 1674
+`^(https:\/\/.*gobwas\.com|http://exclude.gobwas.com)$` | `https://safe.gobwas.com` | `true` | 1039
+`^(https:\/\/.*gobwas\.com|http://exclude.gobwas.com)$` | `http://safe.gobwas.com` | `false` | 272
+`^abc.*$` | `abcdef` | `true` | 237
+`^abc.*$` | `af` | `false` | 100
+`^.*def$` | `abcdef` | `true` | 464
+`^.*def$` | `af` | `false` | 265
+`^ab.*ef$` | `abcdef` | `true` | 375
+`^ab.*ef$` | `af` | `false` | 145
+
+[godoc-image]: https://godoc.org/github.com/gobwas/glob?status.svg
+[godoc-url]: https://godoc.org/github.com/gobwas/glob
+[travis-image]: https://travis-ci.org/gobwas/glob.svg?branch=master
+[travis-url]: https://travis-ci.org/gobwas/glob
+
+## Syntax
+
+Syntax is inspired by [standard wildcards](http://tldp.org/LDP/GNU-Linux-Tools-Summary/html/x11655.htm),
+except that `**` is aka super-asterisk, that do not sensitive for separators.
\ No newline at end of file
diff --git a/vendor/github.com/gobwas/glob/syntax/ast/ast.go b/vendor/github.com/gobwas/glob/syntax/ast/ast.go
new file mode 100644
index 000000000..3220a694a
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/syntax/ast/ast.go
@@ -0,0 +1,122 @@
+package ast
+
+import (
+	"bytes"
+	"fmt"
+)
+
+type Node struct {
+	Parent   *Node
+	Children []*Node
+	Value    interface{}
+	Kind     Kind
+}
+
+func NewNode(k Kind, v interface{}, ch ...*Node) *Node {
+	n := &Node{
+		Kind:  k,
+		Value: v,
+	}
+	for _, c := range ch {
+		Insert(n, c)
+	}
+	return n
+}
+
+func (a *Node) Equal(b *Node) bool {
+	if a.Kind != b.Kind {
+		return false
+	}
+	if a.Value != b.Value {
+		return false
+	}
+	if len(a.Children) != len(b.Children) {
+		return false
+	}
+	for i, c := range a.Children {
+		if !c.Equal(b.Children[i]) {
+			return false
+		}
+	}
+	return true
+}
+
+func (a *Node) String() string {
+	var buf bytes.Buffer
+	buf.WriteString(a.Kind.String())
+	if a.Value != nil {
+		buf.WriteString(" =")
+		buf.WriteString(fmt.Sprintf("%v", a.Value))
+	}
+	if len(a.Children) > 0 {
+		buf.WriteString(" [")
+		for i, c := range a.Children {
+			if i > 0 {
+				buf.WriteString(", ")
+			}
+			buf.WriteString(c.String())
+		}
+		buf.WriteString("]")
+	}
+	return buf.String()
+}
+
+func Insert(parent *Node, children ...*Node) {
+	parent.Children = append(parent.Children, children...)
+	for _, ch := range children {
+		ch.Parent = parent
+	}
+}
+
+type List struct {
+	Not   bool
+	Chars string
+}
+
+type Range struct {
+	Not    bool
+	Lo, Hi rune
+}
+
+type Text struct {
+	Text string
+}
+
+type Kind int
+
+const (
+	KindNothing Kind = iota
+	KindPattern
+	KindList
+	KindRange
+	KindText
+	KindAny
+	KindSuper
+	KindSingle
+	KindAnyOf
+)
+
+func (k Kind) String() string {
+	switch k {
+	case KindNothing:
+		return "Nothing"
+	case KindPattern:
+		return "Pattern"
+	case KindList:
+		return "List"
+	case KindRange:
+		return "Range"
+	case KindText:
+		return "Text"
+	case KindAny:
+		return "Any"
+	case KindSuper:
+		return "Super"
+	case KindSingle:
+		return "Single"
+	case KindAnyOf:
+		return "AnyOf"
+	default:
+		return ""
+	}
+}
diff --git a/vendor/github.com/gobwas/glob/syntax/ast/parser.go b/vendor/github.com/gobwas/glob/syntax/ast/parser.go
new file mode 100644
index 000000000..429b40943
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/syntax/ast/parser.go
@@ -0,0 +1,157 @@
+package ast
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gobwas/glob/syntax/lexer"
+	"unicode/utf8"
+)
+
+type Lexer interface {
+	Next() lexer.Token
+}
+
+type parseFn func(*Node, Lexer) (parseFn, *Node, error)
+
+func Parse(lexer Lexer) (*Node, error) {
+	var parser parseFn
+
+	root := NewNode(KindPattern, nil)
+
+	var (
+		tree *Node
+		err  error
+	)
+	for parser, tree = parserMain, root; parser != nil; {
+		parser, tree, err = parser(tree, lexer)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	return root, nil
+}
+
+func parserMain(tree *Node, lex Lexer) (parseFn, *Node, error) {
+	for {
+		token := lex.Next()
+		switch token.Type {
+		case lexer.EOF:
+			return nil, tree, nil
+
+		case lexer.Error:
+			return nil, tree, errors.New(token.Raw)
+
+		case lexer.Text:
+			Insert(tree, NewNode(KindText, Text{token.Raw}))
+			return parserMain, tree, nil
+
+		case lexer.Any:
+			Insert(tree, NewNode(KindAny, nil))
+			return parserMain, tree, nil
+
+		case lexer.Super:
+			Insert(tree, NewNode(KindSuper, nil))
+			return parserMain, tree, nil
+
+		case lexer.Single:
+			Insert(tree, NewNode(KindSingle, nil))
+			return parserMain, tree, nil
+
+		case lexer.RangeOpen:
+			return parserRange, tree, nil
+
+		case lexer.TermsOpen:
+			a := NewNode(KindAnyOf, nil)
+			Insert(tree, a)
+
+			p := NewNode(KindPattern, nil)
+			Insert(a, p)
+
+			return parserMain, p, nil
+
+		case lexer.Separator:
+			p := NewNode(KindPattern, nil)
+			Insert(tree.Parent, p)
+
+			return parserMain, p, nil
+
+		case lexer.TermsClose:
+			return parserMain, tree.Parent.Parent, nil
+
+		default:
+			return nil, tree, fmt.Errorf("unexpected token: %s", token)
+		}
+	}
+	return nil, tree, fmt.Errorf("unknown error")
+}
+
+func parserRange(tree *Node, lex Lexer) (parseFn, *Node, error) {
+	var (
+		not   bool
+		lo    rune
+		hi    rune
+		chars string
+	)
+	for {
+		token := lex.Next()
+		switch token.Type {
+		case lexer.EOF:
+			return nil, tree, errors.New("unexpected end")
+
+		case lexer.Error:
+			return nil, tree, errors.New(token.Raw)
+
+		case lexer.Not:
+			not = true
+
+		case lexer.RangeLo:
+			r, w := utf8.DecodeRuneInString(token.Raw)
+			if len(token.Raw) > w {
+				return nil, tree, fmt.Errorf("unexpected length of lo character")
+			}
+			lo = r
+
+		case lexer.RangeBetween:
+			//
+
+		case lexer.RangeHi:
+			r, w := utf8.DecodeRuneInString(token.Raw)
+			if len(token.Raw) > w {
+				return nil, tree, fmt.Errorf("unexpected length of lo character")
+			}
+
+			hi = r
+
+			if hi < lo {
+				return nil, tree, fmt.Errorf("hi character '%s' should be greater than lo '%s'", string(hi), string(lo))
+			}
+
+		case lexer.Text:
+			chars = token.Raw
+
+		case lexer.RangeClose:
+			isRange := lo != 0 && hi != 0
+			isChars := chars != ""
+
+			if isChars == isRange {
+				return nil, tree, fmt.Errorf("could not parse range")
+			}
+
+			if isRange {
+				Insert(tree, NewNode(KindRange, Range{
+					Lo:  lo,
+					Hi:  hi,
+					Not: not,
+				}))
+			} else {
+				Insert(tree, NewNode(KindList, List{
+					Chars: chars,
+					Not:   not,
+				}))
+			}
+
+			return parserMain, tree, nil
+		}
+	}
+}
diff --git a/vendor/github.com/gobwas/glob/syntax/lexer/lexer.go b/vendor/github.com/gobwas/glob/syntax/lexer/lexer.go
new file mode 100644
index 000000000..a1c8d1962
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/syntax/lexer/lexer.go
@@ -0,0 +1,273 @@
+package lexer
+
+import (
+	"bytes"
+	"fmt"
+	"github.com/gobwas/glob/util/runes"
+	"unicode/utf8"
+)
+
+const (
+	char_any           = '*'
+	char_comma         = ','
+	char_single        = '?'
+	char_escape        = '\\'
+	char_range_open    = '['
+	char_range_close   = ']'
+	char_terms_open    = '{'
+	char_terms_close   = '}'
+	char_range_not     = '!'
+	char_range_between = '-'
+)
+
+var specials = []byte{
+	char_any,
+	char_single,
+	char_escape,
+	char_range_open,
+	char_range_close,
+	char_terms_open,
+	char_terms_close,
+}
+
+func Special(c byte) bool {
+	return bytes.IndexByte(specials, c) != -1
+}
+
+type tokens []Token
+
+func (i *tokens) shift() (ret Token) {
+	ret = (*i)[0]
+	copy(*i, (*i)[1:])
+	*i = (*i)[:len(*i)-1]
+	return
+}
+
+func (i *tokens) push(v Token) {
+	*i = append(*i, v)
+}
+
+func (i *tokens) empty() bool {
+	return len(*i) == 0
+}
+
+var eof rune = 0
+
+type lexer struct {
+	data string
+	pos  int
+	err  error
+
+	tokens     tokens
+	termsLevel int
+
+	lastRune     rune
+	lastRuneSize int
+	hasRune      bool
+}
+
+func NewLexer(source string) *lexer {
+	l := &lexer{
+		data:   source,
+		tokens: tokens(make([]Token, 0, 4)),
+	}
+	return l
+}
+
+func (l *lexer) Next() Token {
+	if l.err != nil {
+		return Token{Error, l.err.Error()}
+	}
+	if !l.tokens.empty() {
+		return l.tokens.shift()
+	}
+
+	l.fetchItem()
+	return l.Next()
+}
+
+func (l *lexer) peek() (r rune, w int) {
+	if l.pos == len(l.data) {
+		return eof, 0
+	}
+
+	r, w = utf8.DecodeRuneInString(l.data[l.pos:])
+	if r == utf8.RuneError {
+		l.errorf("could not read rune")
+		r = eof
+		w = 0
+	}
+
+	return
+}
+
+func (l *lexer) read() rune {
+	if l.hasRune {
+		l.hasRune = false
+		l.seek(l.lastRuneSize)
+		return l.lastRune
+	}
+
+	r, s := l.peek()
+	l.seek(s)
+
+	l.lastRune = r
+	l.lastRuneSize = s
+
+	return r
+}
+
+func (l *lexer) seek(w int) {
+	l.pos += w
+}
+
+func (l *lexer) unread() {
+	if l.hasRune {
+		l.errorf("could not unread rune")
+		return
+	}
+	l.seek(-l.lastRuneSize)
+	l.hasRune = true
+}
+
+func (l *lexer) errorf(f string, v ...interface{}) {
+	l.err = fmt.Errorf(f, v...)
+}
+
+func (l *lexer) inTerms() bool {
+	return l.termsLevel > 0
+}
+
+func (l *lexer) termsEnter() {
+	l.termsLevel++
+}
+
+func (l *lexer) termsLeave() {
+	l.termsLevel--
+}
+
+var inTextBreakers = []rune{char_single, char_any, char_range_open, char_terms_open}
+var inTermsBreakers = append(inTextBreakers, char_terms_close, char_comma)
+
+func (l *lexer) fetchItem() {
+	r := l.read()
+	switch {
+	case r == eof:
+		l.tokens.push(Token{EOF, ""})
+
+	case r == char_terms_open:
+		l.termsEnter()
+		l.tokens.push(Token{TermsOpen, string(r)})
+
+	case r == char_comma && l.inTerms():
+		l.tokens.push(Token{Separator, string(r)})
+
+	case r == char_terms_close && l.inTerms():
+		l.tokens.push(Token{TermsClose, string(r)})
+		l.termsLeave()
+
+	case r == char_range_open:
+		l.tokens.push(Token{RangeOpen, string(r)})
+		l.fetchRange()
+
+	case r == char_single:
+		l.tokens.push(Token{Single, string(r)})
+
+	case r == char_any:
+		if l.read() == char_any {
+			l.tokens.push(Token{Super, string(r) + string(r)})
+		} else {
+			l.unread()
+			l.tokens.push(Token{Any, string(r)})
+		}
+
+	default:
+		l.unread()
+
+		var breakers []rune
+		if l.inTerms() {
+			breakers = inTermsBreakers
+		} else {
+			breakers = inTextBreakers
+		}
+		l.fetchText(breakers)
+	}
+}
+
+func (l *lexer) fetchRange() {
+	var wantHi bool
+	var wantClose bool
+	var seenNot bool
+	for {
+		r := l.read()
+		if r == eof {
+			l.errorf("unexpected end of input")
+			return
+		}
+
+		if wantClose {
+			if r != char_range_close {
+				l.errorf("expected close range character")
+			} else {
+				l.tokens.push(Token{RangeClose, string(r)})
+			}
+			return
+		}
+
+		if wantHi {
+			l.tokens.push(Token{RangeHi, string(r)})
+			wantClose = true
+			continue
+		}
+
+		if !seenNot && r == char_range_not {
+			l.tokens.push(Token{Not, string(r)})
+			seenNot = true
+			continue
+		}
+
+		if n, w := l.peek(); n == char_range_between {
+			l.seek(w)
+			l.tokens.push(Token{RangeLo, string(r)})
+			l.tokens.push(Token{RangeBetween, string(n)})
+			wantHi = true
+			continue
+		}
+
+		l.unread() // unread first peek and fetch as text
+		l.fetchText([]rune{char_range_close})
+		wantClose = true
+	}
+}
+
+func (l *lexer) fetchText(breakers []rune) {
+	var data []rune
+	var escaped bool
+
+reading:
+	for {
+		r := l.read()
+		if r == eof {
+			break
+		}
+
+		if !escaped {
+			if r == char_escape {
+				escaped = true
+				continue
+			}
+
+			if runes.IndexRune(breakers, r) != -1 {
+				l.unread()
+				break reading
+			}
+		}
+
+		escaped = false
+		data = append(data, r)
+	}
+
+	if len(data) > 0 {
+		l.tokens.push(Token{Text, string(data)})
+	}
+}
diff --git a/vendor/github.com/gobwas/glob/syntax/lexer/token.go b/vendor/github.com/gobwas/glob/syntax/lexer/token.go
new file mode 100644
index 000000000..2797c4e83
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/syntax/lexer/token.go
@@ -0,0 +1,88 @@
+package lexer
+
+import "fmt"
+
+type TokenType int
+
+const (
+	EOF TokenType = iota
+	Error
+	Text
+	Char
+	Any
+	Super
+	Single
+	Not
+	Separator
+	RangeOpen
+	RangeClose
+	RangeLo
+	RangeHi
+	RangeBetween
+	TermsOpen
+	TermsClose
+)
+
+func (tt TokenType) String() string {
+	switch tt {
+	case EOF:
+		return "eof"
+
+	case Error:
+		return "error"
+
+	case Text:
+		return "text"
+
+	case Char:
+		return "char"
+
+	case Any:
+		return "any"
+
+	case Super:
+		return "super"
+
+	case Single:
+		return "single"
+
+	case Not:
+		return "not"
+
+	case Separator:
+		return "separator"
+
+	case RangeOpen:
+		return "range_open"
+
+	case RangeClose:
+		return "range_close"
+
+	case RangeLo:
+		return "range_lo"
+
+	case RangeHi:
+		return "range_hi"
+
+	case RangeBetween:
+		return "range_between"
+
+	case TermsOpen:
+		return "terms_open"
+
+	case TermsClose:
+		return "terms_close"
+
+	default:
+		return "undef"
+	}
+}
+
+type Token struct {
+	Type TokenType
+	Raw  string
+}
+
+func (t Token) String() string {
+	return fmt.Sprintf("%v<%q>", t.Type, t.Raw)
+}
diff --git a/vendor/github.com/gobwas/glob/syntax/syntax.go b/vendor/github.com/gobwas/glob/syntax/syntax.go
new file mode 100644
index 000000000..1d168b148
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/syntax/syntax.go
@@ -0,0 +1,14 @@
+package syntax
+
+import (
+	"github.com/gobwas/glob/syntax/ast"
+	"github.com/gobwas/glob/syntax/lexer"
+)
+
+func Parse(s string) (*ast.Node, error) {
+	return ast.Parse(lexer.NewLexer(s))
+}
+
+func Special(b byte) bool {
+	return lexer.Special(b)
+}
diff --git a/vendor/github.com/gobwas/glob/util/runes/runes.go b/vendor/github.com/gobwas/glob/util/runes/runes.go
new file mode 100644
index 000000000..a72355641
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/util/runes/runes.go
@@ -0,0 +1,154 @@
+package runes
+
+func Index(s, needle []rune) int {
+	ls, ln := len(s), len(needle)
+
+	switch {
+	case ln == 0:
+		return 0
+	case ln == 1:
+		return IndexRune(s, needle[0])
+	case ln == ls:
+		if Equal(s, needle) {
+			return 0
+		}
+		return -1
+	case ln > ls:
+		return -1
+	}
+
+head:
+	for i := 0; i < ls && ls-i >= ln; i++ {
+		for y := 0; y < ln; y++ {
+			if s[i+y] != needle[y] {
+				continue head
+			}
+		}
+
+		return i
+	}
+
+	return -1
+}
+
+func LastIndex(s, needle []rune) int {
+	ls, ln := len(s), len(needle)
+
+	switch {
+	case ln == 0:
+		if ls == 0 {
+			return 0
+		}
+		return ls
+	case ln == 1:
+		return IndexLastRune(s, needle[0])
+	case ln == ls:
+		if Equal(s, needle) {
+			return 0
+		}
+		return -1
+	case ln > ls:
+		return -1
+	}
+
+head:
+	for i := ls - 1; i >= 0 && i >= ln; i-- {
+		for y := ln - 1; y >= 0; y-- {
+			if s[i-(ln-y-1)] != needle[y] {
+				continue head
+			}
+		}
+
+		return i - ln + 1
+	}
+
+	return -1
+}
+
+// IndexAny returns the index of the first instance of any Unicode code point
+// from chars in s, or -1 if no Unicode code point from chars is present in s.
+func IndexAny(s, chars []rune) int {
+	if len(chars) > 0 {
+		for i, c := range s {
+			for _, m := range chars {
+				if c == m {
+					return i
+				}
+			}
+		}
+	}
+	return -1
+}
+
+func Contains(s, needle []rune) bool {
+	return Index(s, needle) >= 0
+}
+
+func Max(s []rune) (max rune) {
+	for _, r := range s {
+		if r > max {
+			max = r
+		}
+	}
+
+	return
+}
+
+func Min(s []rune) rune {
+	min := rune(-1)
+	for _, r := range s {
+		if min == -1 {
+			min = r
+			continue
+		}
+
+		if r < min {
+			min = r
+		}
+	}
+
+	return min
+}
+
+func IndexRune(s []rune, r rune) int {
+	for i, c := range s {
+		if c == r {
+			return i
+		}
+	}
+	return -1
+}
+
+func IndexLastRune(s []rune, r rune) int {
+	for i := len(s) - 1; i >= 0; i-- {
+		if s[i] == r {
+			return i
+		}
+	}
+
+	return -1
+}
+
+func Equal(a, b []rune) bool {
+	if len(a) == len(b) {
+		for i := 0; i < len(a); i++ {
+			if a[i] != b[i] {
+				return false
+			}
+		}
+
+		return true
+	}
+
+	return false
+}
+
+// HasPrefix tests whether the string s begins with prefix.
+func HasPrefix(s, prefix []rune) bool {
+	return len(s) >= len(prefix) && Equal(s[0:len(prefix)], prefix)
+}
+
+// HasSuffix tests whether the string s ends with suffix.
+func HasSuffix(s, suffix []rune) bool {
+	return len(s) >= len(suffix) && Equal(s[len(s)-len(suffix):], suffix)
+}
diff --git a/vendor/github.com/gobwas/glob/util/strings/strings.go b/vendor/github.com/gobwas/glob/util/strings/strings.go
new file mode 100644
index 000000000..e8ee1920b
--- /dev/null
+++ b/vendor/github.com/gobwas/glob/util/strings/strings.go
@@ -0,0 +1,39 @@
+package strings
+
+import (
+	"strings"
+	"unicode/utf8"
+)
+
+func IndexAnyRunes(s string, rs []rune) int {
+	for _, r := range rs {
+		if i := strings.IndexRune(s, r); i != -1 {
+			return i
+		}
+	}
+
+	return -1
+}
+
+func LastIndexAnyRunes(s string, rs []rune) int {
+	for _, r := range rs {
+		i := -1
+		if 0 <= r && r < utf8.RuneSelf {
+			i = strings.LastIndexByte(s, byte(r))
+		} else {
+			sub := s
+			for len(sub) > 0 {
+				j := strings.IndexRune(s, r)
+				if j == -1 {
+					break
+				}
+				i = j
+				sub = sub[i+1:]
+			}
+		}
+		if i != -1 {
+			return i
+		}
+	}
+	return -1
+}
diff --git a/vendor/github.com/ryanuber/go-glob/README.md b/vendor/github.com/ryanuber/go-glob/README.md
deleted file mode 100644
index 48f7fcb05..000000000
--- a/vendor/github.com/ryanuber/go-glob/README.md
+++ /dev/null
@@ -1,29 +0,0 @@
-# String globbing in golang [![Build Status](https://travis-ci.org/ryanuber/go-glob.svg)](https://travis-ci.org/ryanuber/go-glob)
-
-`go-glob` is a single-function library implementing basic string glob support.
-
-Globs are an extremely user-friendly way of supporting string matching without
-requiring knowledge of regular expressions or Go's particular regex engine. Most
-people understand that if you put a `*` character somewhere in a string, it is
-treated as a wildcard. Surprisingly, this functionality isn't found in Go's
-standard library, except for `path.Match`, which is intended to be used while
-comparing paths (not arbitrary strings), and contains specialized logic for this
-use case. A better solution might be a POSIX basic (non-ERE) regular expression
-engine for Go, which doesn't exist currently.
-
-Example
-=======
-
-```
-package main
-
-import "github.com/ryanuber/go-glob"
-
-func main() {
-    glob.Glob("*World!", "Hello, World!") // true
-    glob.Glob("Hello,*", "Hello, World!") // true
-    glob.Glob("*ello,*", "Hello, World!") // true
-    glob.Glob("World!", "Hello, World!")  // false
-    glob.Glob("/home/*", "/home/ryanuber/.bashrc") // true
-}
-```
diff --git a/vendor/github.com/ryanuber/go-glob/glob.go b/vendor/github.com/ryanuber/go-glob/glob.go
deleted file mode 100644
index d9d46379a..000000000
--- a/vendor/github.com/ryanuber/go-glob/glob.go
+++ /dev/null
@@ -1,51 +0,0 @@
-package glob
-
-import "strings"
-
-// The character which is treated like a glob
-const GLOB = "*"
-
-// Glob will test a string pattern, potentially containing globs, against a
-// subject string. The result is a simple true/false, determining whether or
-// not the glob pattern matched the subject text.
-func Glob(pattern, subj string) bool {
-	// Empty pattern can only match empty subject
-	if pattern == "" {
-		return subj == pattern
-	}
-
-	// If the pattern _is_ a glob, it matches everything
-	if pattern == GLOB {
-		return true
-	}
-
-	parts := strings.Split(pattern, GLOB)
-
-	if len(parts) == 1 {
-		// No globs in pattern, so test for equality
-		return subj == pattern
-	}
-
-	leadingGlob := strings.HasPrefix(pattern, GLOB)
-	trailingGlob := strings.HasSuffix(pattern, GLOB)
-	end := len(parts) - 1
-
-	// Check the first section. Requires special handling.
-	if !leadingGlob && !strings.HasPrefix(subj, parts[0]) {
-		return false
-	}
-
-	// Go over the middle parts and ensure they match.
-	for i := 1; i < end; i++ {
-		if !strings.Contains(subj, parts[i]) {
-			return false
-		}
-
-		// Trim evaluated text from subj as we loop over the pattern.
-		idx := strings.Index(subj, parts[i]) + len(parts[i])
-		subj = subj[idx:]
-	}
-
-	// Reached the last section. Requires special handling.
-	return trailingGlob || strings.HasSuffix(subj, parts[end])
-}
diff --git a/vendor/vendor.json b/vendor/vendor.json
index f31daa925..0625b8844 100644
--- a/vendor/vendor.json
+++ b/vendor/vendor.json
@@ -9,6 +9,14 @@
 		{"path":"github.com/circonus-labs/circonusllhist","checksumSHA1":"C4Z7+l5GOpOCW5DcvNYzheGvQRE=","revision":"d724266ae5270ae8b87a5d2e8081f04e307c3c18","revisionTime":"2016-05-26T04:38:13Z"},
 		{"path":"github.com/cyberdelia/go-metrics-graphite","checksumSHA1":"8/Q1JbAHUmL4sDURLq6yron4K/I=","revision":"b8345b7f01d571b05366d5791a034d872f1bb36f","revisionTime":"2015-08-25T20:22:00-07:00"},
 		{"path":"github.com/fatih/structs","checksumSHA1":"zKLQNNEpgfrJOwVrDDyBMYEIx/w=","revision":"5ada2f449b108d87dbd8c1e60c32cdd065c27886","revisionTime":"2016-06-01T09:31:17Z"},
+		{"path":"github.com/gobwas/glob","checksumSHA1":"/RE0VzeaXdaBUNeNMUXWqzly7qY=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
+		{"path":"github.com/gobwas/glob/compiler","checksumSHA1":"UIxNdlqa2kcYi5EEe79/WzyPiz4=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
+		{"path":"github.com/gobwas/glob/match","checksumSHA1":"jNdbCxBabqjv9tREXA4Nx45Y4wg=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
+		{"path":"github.com/gobwas/glob/syntax","checksumSHA1":"2KhDNUE98XhHnIgg2S9E4gcWoig=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
+		{"path":"github.com/gobwas/glob/syntax/ast","checksumSHA1":"FPIMdPSazNZGNU+ZmPtIeZpVHFU=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
+		{"path":"github.com/gobwas/glob/syntax/lexer","checksumSHA1":"umyztSrQrQIl9JgwhDDb6wrSLeI=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
+		{"path":"github.com/gobwas/glob/util/runes","checksumSHA1":"vux29fN5O22F3QPLOnjQ37278vg=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
+		{"path":"github.com/gobwas/glob/util/strings","checksumSHA1":"Sxp2AHWkkLkIq6w9aMcLmzcwD6w=","revision":"19c076cdf202b3d1c0489bdfa2f2f289f634474b","revisionTime":"2018-02-08T21:18:42Z"},
 		{"path":"github.com/hashicorp/consul/api","checksumSHA1":"GsJ84gKbQno8KbojhVTgSVWNues=","revision":"402636ff2db998edef392ac6d59210d2170b3ebf","revisionTime":"2017-04-05T04:22:14Z","version":"v0.8.0","versionExact":"v0.8.0"},
 		{"path":"github.com/hashicorp/errwrap","checksumSHA1":"cdOCt0Yb+hdErz8NAQqayxPmRsY=","revision":"7554cd9344cec97297fa6649b055a8c98c2a1e55","revisionTime":"2014-10-27T22:47:10-07:00"},
 		{"path":"github.com/hashicorp/go-cleanhttp","checksumSHA1":"Uzyon2091lmwacNsl1hCytjhHtg=","revision":"ad28ea4487f05916463e2423a55166280e8254b5","revisionTime":"2016-04-07T17:41:26Z"},
@@ -34,7 +42,6 @@
 		{"path":"github.com/pkg/profile","checksumSHA1":"C3yiSMdTQxSY3xqKJzMV9T+KnIc=","revision":"5b67d428864e92711fcbd2f8629456121a56d91f","revisionTime":"2017-05-09T09:25:25Z"},
 		{"path":"github.com/rcrowley/go-metrics","checksumSHA1":"ODTWX4h8f+DW3oWZFT0yTmfHzdg=","revision":"3e5e593311103d49927c8d2b0fd93ccdfe4a525c","revisionTime":"2015-07-19T09:56:14-07:00"},
 		{"path":"github.com/rogpeppe/fastuuid","checksumSHA1":"ehRkDJisGCCSYdNgyvs1gSywSPE=","revision":"6724a57986aff9bff1a1770e9347036def7c89f6","revisionTime":"2015-01-06T09:31:45Z"},
-		{"path":"github.com/ryanuber/go-glob","checksumSHA1":"EYkx4sXDLSEd1xUtGoXRsfd5cpw=","revision":"572520ed46dbddaed19ea3d9541bdd0494163693","revisionTime":"2016-02-26T08:37:05Z"},
 		{"path":"github.com/sergi/go-diff/diffmatchpatch","checksumSHA1":"iWCtyR1TkJ22Bi/ygzfKDvOQdQY=","revision":"24e2351369ec4949b2ed0dc5c477afdd4c4034e8","revisionTime":"2017-01-18T13:12:30Z"},
 		{"path":"golang.org/x/net/http2","checksumSHA1":"N1akwAdrHVfPPrsFOhG2ouP21VA=","revision":"f2499483f923065a842d38eb4c7f1927e6fc6e6d","revisionTime":"2017-01-14T04:22:49Z"},
 		{"path":"golang.org/x/net/http2/hpack","checksumSHA1":"HzuGD7AwgC0p1az1WAQnEFnEk98=","revision":"f2499483f923065a842d38eb4c7f1927e6fc6e6d","revisionTime":"2017-01-14T04:22:49Z"},