Merge pull request #5 from gobwas/feature/quote

Feature/quote

compiler.go

@@ -1,6 +1,7 @@
 package glob
 
 // TODO use constructor with all matchers, and to their structs private
+// TODO glue multiple Text nodes (like after QuoteMeta)
 
 import (
 	"fmt"

glob.go

@@ -54,3 +54,22 @@ func MustCompile(pattern string, separators ...rune) Glob {
 
 	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 special(s[i]) {
+			b[j] = '\\'
+			j++
+		}
+		b[j] = s[i]
+		j++
+	}
+
+	return string(b[0:j])
+}

glob_test.go

@@ -150,6 +150,39 @@ func TestGlob(t *testing.T) {
 	}
 }
 
+func TestQuoteMeta(t *testing.T) {
+	specialsQuoted := make([]byte, len(specials)*2)
+	for i, j := 0, 0; i < len(specials); i, j = i+1, j+2 {
+		specialsQuoted[j] = '\\'
+		specialsQuoted[j+1] = specials[i]
+	}
+
+	for id, test := range []struct {
+		in, out string
+	}{
+		{
+			in:  `[foo*]`,
+			out: `\[foo\*\]`,
+		},
+		{
+			in:  string(specials),
+			out: string(specialsQuoted),
+		},
+		{
+			in:  string(append([]byte("some text and"), specials...)),
+			out: string(append([]byte("some text and"), specialsQuoted...)),
+		},
+	} {
+		act := QuoteMeta(test.in)
+		if act != test.out {
+			t.Errorf("#%d QuoteMeta(%q) = %q; want %q", id, test.in, act, test.out)
+		}
+		if _, err := Compile(act); err != nil {
+			t.Errorf("#%d _, err := Compile(QuoteMeta(%q) = %q); err = %q", id, test.in, act, err)
+		}
+	}
+}
+
 func BenchmarkParseGlob(b *testing.B) {
 	for i := 0; i < b.N; i++ {
 		Compile(pattern_all)
@@ -193,12 +226,12 @@ func BenchmarkAllGlobMismatch(b *testing.B) {
 		_ = m.Match(fixture_all_mismatch)
 	}
 }
-func BenchmarkAllGlobMatchParallel(b *testing.B) {
+func BenchmarkAllGlobMismatchParallel(b *testing.B) {
 	m, _ := Compile(pattern_all)
 
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			_ = m.Match(fixture_all_match)
+			_ = m.Match(fixture_all_mismatch)
 		}
 	})
 }

lexer.go

@@ -1,6 +1,7 @@
 package glob
 
 import (
+	"bytes"
 	"fmt"
 	"strings"
 	"unicode/utf8"
@@ -19,6 +20,20 @@ const (
 	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
+}
+
 var eof rune = 0
 
 type stateFn func(*lexer) stateFn
@@ -124,7 +139,7 @@ func newLexer(source string) *lexer {
 	l := &lexer{
 		input:       source,
 		state:       lexText,
-		items:       make(chan item, 5),
+		items:       make(chan item, len(source)),
 		termPhrases: make(map[int]int),
 	}
 	return l
@@ -281,9 +296,7 @@ func lexText(l *lexer) stateFn {
 			l.unread()
 			l.emitMaybe(item_text)
 			return lexSeparator
-
 		}
-
 	}
 
 	if l.pos > l.start {

match/text.go

@@ -41,5 +41,5 @@ func (self Text) Index(s string) (int, []int) {
 }
 
 func (self Text) String() string {
-	return fmt.Sprintf("<text:%s>", self.Str)
+	return fmt.Sprintf("<text:`%v`>", self.Str)
 }

readme.md

@@ -25,6 +25,10 @@ func main() {
     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
@@ -76,7 +80,6 @@ func main() {
     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

todo.txt

@@ -1,26 +0,0 @@
-benchmark                                        old ns/op     new ns/op     delta
-
-BenchmarkAllGlobMatch-4                          519           1024          +97.30%
-BenchmarkMultipleGlobMatch-4                     123           218           +77.24%
-BenchmarkAlternativesGlobMatch-4                 164           283           +72.56%
-BenchmarkAlternativesSuffixFirstGlobMatch-4      23.6          23.5          -0.42%
-BenchmarkAlternativesSuffixSecondGlobMatch-4     29.7          30.1          +1.35%
-BenchmarkAlternativesCombineLiteGlobMatch-4      161           352           +118.63%
-BenchmarkAlternativesCombineHardGlobMatch-4      321           649           +102.18%
-BenchmarkPlainGlobMatch-4                        7.17          7.09          -1.12%
-BenchmarkPrefixGlobMatch-4                       8.74          8.64          -1.14%
-BenchmarkSuffixGlobMatch-4                       10.3          9.06          -12.04%
-BenchmarkPrefixSuffixGlobMatch-4                 31.0          15.1          -51.29%
-BenchmarkIndexAny-4                              1414          232           -83.59%
-BenchmarkIndexContains-4                         557           250           -55.12%
-BenchmarkIndexList-4                             207           42.6          -79.42%
-BenchmarkIndexMax-4                              630           111           -82.38%
-BenchmarkIndexMin-4                              515           328           -36.31%
-BenchmarkIndexPrefixSuffix-4                     97.9          86.2          -11.95%
-BenchmarkIndexPrefix-4                           86.1          84.0          -2.44%
-BenchmarkIndexRange-4                            181           144           -20.44%
-BenchmarkRowIndex-4                              185           127           -31.35%
-BenchmarkIndexSingle-4                           82.6          16.0          -80.63%
-BenchmarkIndexSuffix-4                           85.5          84.9          -0.70%
-BenchmarkIndexSuper-4                            450           196           -56.44%
-BenchmarkIndexText-4                             85.3          85.9          +0.70%