update: add more tests

Signed-off-by: Milos Gajdos <[email protected]>

document/text/character_test.go

@@ -60,13 +60,28 @@ func TestCharSplitter(t *testing.T) {
 			input:   "foo.bar.baz.123",
 			exp:     []string{"foo", ".bar", ".baz", ".123"},
 		},
+		{
+			size:    1,
+			overlap: 0,
+			keepSep: true,
+			sep:     Sep{Value: `\.`, IsRegexp: true},
+			input:   "foo.bar.baz.123",
+			exp:     []string{"foo", ".bar", ".baz", ".123"},
+		},
 		{
 			size:    1,
 			overlap: 0,
 			sep:     Sep{Value: ".", IsRegexp: false},
 			input:   "foo.bar.baz.123",
 			exp:     []string{"foo", "bar", "baz", "123"},
 		},
+		{
+			size:    1,
+			overlap: 0,
+			sep:     Sep{Value: `\.`, IsRegexp: true},
+			input:   "foo.bar.baz.123",
+			exp:     []string{"foo", "bar", "baz", "123"},
+		},
 	}
 
 	for _, tc := range testCases {

document/text/splitter.go

@@ -157,11 +157,10 @@ func (s *Splitter) splitText(text string, sep Sep) []string {
 	if sep.Value != "" {
 		if s.keepSep {
 			sepVal := sep.Value
-			if !sep.IsRegexp {
-				// NOTE: we must do this to unescape
-				// the escaped separator
-				sepVal, _ = unquoteMeta(sep.Value)
-			}
+			// NOTE: we must do this to unescape
+			// the escaped separator so we keep the raw separator.
+			sepVal, _ = unquoteMeta(sep.Value)
+
 			var results []string
 			splits := regexp.MustCompile("("+sep.Value+")").Split(text, -1)
 			// NOTE: we start iterating from 1, not 0!