Fix GlobPrefix. (#220)

ncruces · Jan 28, 2025 · b2f7ab8 · b2f7ab8
1 parent c9135b9
commit b2f7ab8
Show file tree

Hide file tree

Showing 2 changed files with 103 additions and 17 deletions.
diff --git a/ext/regexp/regexp.go b/ext/regexp/regexp.go
@@ -16,7 +16,9 @@ package regexp
 import (
 	"errors"
 	"regexp"
+	"regexp/syntax"
 	"strings"
+	"unicode/utf8"
 
 	"github.com/ncruces/go-sqlite3"
 )
@@ -50,16 +52,63 @@ func Register(db *sqlite3.Conn) error {
 //	SELECT column WHERE column GLOB :glob_prefix AND column REGEXP :regexp
 //
 // [LIKE optimization]: https://sqlite.org/optoverview.html#the_like_optimization
-func GlobPrefix(re *regexp.Regexp) string {
-	prefix, complete := re.LiteralPrefix()
-	i := strings.IndexAny(prefix, "*?[")
-	if i < 0 {
-		if complete {
-			return prefix
+func GlobPrefix(expr string) string {
+	re, err := syntax.Parse(expr, syntax.Perl)
+	if err != nil {
+		return "" // no match possible
+	}
+	prog, err := syntax.Compile(re.Simplify())
+	if err != nil {
+		return "" // notest
+	}
+
+	i := &prog.Inst[prog.Start]
+
+	var empty syntax.EmptyOp
+loop1:
+	for {
+		switch i.Op {
+		case syntax.InstFail:
+			return "" // notest
+		case syntax.InstCapture, syntax.InstNop:
+			// skip
+		case syntax.InstEmptyWidth:
+			empty |= syntax.EmptyOp(i.Arg)
+		default:
+			break loop1
 		}
-		i = len(prefix)
+		i = &prog.Inst[i.Out]
+	}
+	if empty&syntax.EmptyBeginText == 0 {
+		return "*" // not anchored
 	}
-	return prefix[:i] + "*"
+
+	var glob strings.Builder
+loop2:
+	for {
+		switch i.Op {
+		case syntax.InstFail:
+			return "" // notest
+		case syntax.InstCapture, syntax.InstEmptyWidth, syntax.InstNop:
+			// skip
+		case syntax.InstRune, syntax.InstRune1:
+			if len(i.Rune) != 1 || syntax.Flags(i.Arg)&syntax.FoldCase != 0 {
+				break loop2
+			}
+			switch r := i.Rune[0]; r {
+			case '*', '?', '[', utf8.RuneError:
+				break loop2
+			default:
+				glob.WriteRune(r)
+			}
+		default:
+			break loop2
+		}
+		i = &prog.Inst[i.Out]
+	}
+
+	glob.WriteByte('*')
+	return glob.String()
 }
 
 func load(ctx sqlite3.Context, i int, expr string) (*regexp.Regexp, error) {

diff --git a/ext/regexp/regexp_test.go b/ext/regexp/regexp_test.go
@@ -3,6 +3,7 @@ package regexp
 import (
 	"database/sql"
 	"regexp"
+	"strings"
 	"testing"
 
 	"github.com/ncruces/go-sqlite3/driver"
@@ -108,19 +109,55 @@ func TestGlobPrefix(t *testing.T) {
 		re   string
 		want string
 	}{
-		{``, ""},
-		{`a`, "a"},
-		{`a*`, "*"},
-		{`a+`, "a*"},
-		{`ab*`, "a*"},
-		{`ab+`, "ab*"},
-		{`a\?b`, "a*"},
+		{`[`, ""},
+		{``, "*"},
+		{`^`, "*"},
+		{`a`, "*"},
+		{`ab`, "*"},
+		{`^a`, "a*"},
+		{`^a*`, "*"},
+		{`^a+`, "a*"},
+		{`^ab*`, "a*"},
+		{`^ab+`, "ab*"},
+		{`^a\?b`, "a*"},
+		{`^[a-z]`, "*"},
 	}
 	for _, tt := range tests {
 		t.Run(tt.re, func(t *testing.T) {
-			if got := GlobPrefix(regexp.MustCompile(tt.re)); got != tt.want {
-				t.Errorf("GlobPrefix() = %v, want %v", got, tt.want)
+			if got := GlobPrefix(tt.re); got != tt.want {
+				t.Errorf("GlobPrefix(%v) = %v, want %v", tt.re, got, tt.want)
 			}
 		})
 	}
 }
+
+func FuzzGlobPrefix(f *testing.F) {
+	f.Add(``, ``)
+	f.Add(`[`, ``)
+	f.Add(`^`, ``)
+	f.Add(`a`, `a`)
+	f.Add(`ab`, `b`)
+	f.Add(`^a`, `a`)
+	f.Add(`^a*`, `ab`)
+	f.Add(`^a+`, `ab`)
+	f.Add(`^ab*`, `ab`)
+	f.Add(`^ab+`, `ab`)
+	f.Add(`^a\?b`, `ab`)
+	f.Add(`^[a-z]`, `ab`)
+
+	f.Fuzz(func(t *testing.T, lit, str string) {
+		re, err := regexp.Compile(lit)
+		if err != nil {
+			t.SkipNow()
+		}
+		if re.MatchString(str) {
+			prefix, ok := strings.CutSuffix(GlobPrefix(lit), "*")
+			if !ok {
+				t.Fatalf("missing * after %q for %q with %q", prefix, lit, str)
+			}
+			if !strings.HasPrefix(str, prefix) {
+				t.Fatalf("missing prefix %q for %q with %q", prefix, lit, str)
+			}
+		}
+	})
+}