fixes #65 add support for \pX syntax for single-char code classes

Author: dlclark

regexp_test.go

@@ -1296,3 +1296,36 @@ func TestGoodReverseOrderMessage(t *testing.T) {
 		t.Fatalf("expected %q got %q", expected, err.Error())
 	}
 }
+
+func TestParseShortSlashP(t *testing.T) {
+	re := MustCompile(`[!\pL\pN]{1,}`, 0)
+	m, err := re.FindStringMatch("this23! is a! test 1a 2b")
+	if err != nil {
+		t.Fatalf("Unexpected error: %v", err)
+	}
+	if m.String() != "this23!" {
+		t.Fatalf("Expected match")
+	}
+}
+
+func TestParseShortSlashNegateP(t *testing.T) {
+	re := MustCompile(`\PNa`, 0)
+	m, err := re.FindStringMatch("this is a test 1a 2b")
+	if err != nil {
+		t.Fatalf("Unexpected error: %v", err)
+	}
+	if m.String() != " a" {
+		t.Fatalf("Expected match")
+	}
+}
+
+func TestParseShortSlashPEnd(t *testing.T) {
+	re := MustCompile(`\pN`, 0)
+	m, err := re.FindStringMatch("this is a test 1a 2b")
+	if err != nil {
+		t.Fatalf("Unexpected error: %v", err)
+	}
+	if m.String() != "1" {
+		t.Fatalf("Expected match")
+	}
+}

syntax/parser.go

@@ -1311,6 +1311,17 @@ func (p *parser) scanBasicBackslash(scanOnly bool) (*regexNode, error) {
 
 // Scans X for \p{X} or \P{X}
 func (p *parser) parseProperty() (string, error) {
+	// RE2 and PCRE supports \pX syntax (no {} and only 1 letter unicode cats supported)
+	// since this is purely additive syntax it's not behind a flag
+	if p.charsRight() >= 1 && p.rightChar(0) != '{' {
+		ch := string(p.moveRightGetChar())
+		// check if it's a valid cat
+		if !isValidUnicodeCat(ch) {
+			return "", p.getErr(ErrUnknownSlashP, ch)
+		}
+		return ch, nil
+	}
+
 	if p.charsRight() < 3 {
 		return "", p.getErr(ErrIncompleteSlashP)
 	}
@@ -1427,7 +1438,7 @@ func (p *parser) scanCapname() string {
 	return string(p.pattern[startpos:p.textpos()])
 }
 
-//Scans contents of [] (not including []'s), and converts to a set.
+// Scans contents of [] (not including []'s), and converts to a set.
 func (p *parser) scanCharSet(caseInsensitive, scanOnly bool) (*CharSet, error) {
 	ch := '\x00'
 	chPrev := '\x00'