Fix bug for text starting with ScriptInherited, fixes #336

Author: tdewolff

text/text.go

@@ -13,6 +13,10 @@ type ScriptItem struct {
 	Text  string
 }
 
+func (item *ScriptItem) String() string {
+	return fmt.Sprintf("{%v %v %v}", item.Script, item.Level, item.Text)
+}
+
 // ScriptItemizer divides the string in parts for each different script. Also separates on different embedding levels and unicode.ReplacementChar (replaced by object).
 func ScriptItemizer(runes []rune, embeddingLevels []int) []ScriptItem {
 	if len(runes) == 0 {
@@ -27,8 +31,10 @@ func ScriptItemizer(runes []rune, embeddingLevels []int) []ScriptItem {
 		if script == ScriptInherited {
 			if r == '\u200C' || r == '\u200D' {
 				script = ScriptCommon
+			} else if level < len(scripts) {
+				script = scripts[level] // take level from preceding base character
 			} else {
-				script = scripts[level] // take leven from preceding base character
+				script = ScriptUnknown
 			}
 		}
 		prevScript := scripts[len(scripts)-1]

text/text_test.go

@@ -0,0 +1,26 @@
+package text
+
+import (
+	"testing"
+
+	"github.com/tdewolff/test"
+)
+
+func TestScriptItemizer(t *testing.T) {
+	var tests = []struct {
+		str   string
+		items []ScriptItem
+	}{
+		{"abc", []ScriptItem{{Latin, 0, "abc"}}},
+		{"\u064bياعادلا", []ScriptItem{{Arabic, 1, "\u064bياعادلا"}}},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.str, func(t *testing.T) {
+			runes := []rune(tt.str)
+			embeddingLevels := EmbeddingLevels(runes)
+			items := ScriptItemizer(runes, embeddingLevels)
+			test.T(t, items, tt.items)
+		})
+	}
+}