Find/REplace works with Umlauts

Author: linkdotnet

CHANGELOG.md

@@ -6,6 +6,9 @@ All notable changes to **ValueStringBuilder** will be documented in this file. T
 
 ## [Unreleased]
 
+### Fixed
+- Fixed a bug where find and friends and Replace did not work properly with umlauts or 2 (or more) byte character 
+
 ## [0.5.0] - 2022-12-26
 
 ### Added

search.go

@@ -3,19 +3,22 @@ package Text
 // Returns all occurences of needle in haystack
 func findAll(haystack []rune, needle string) []int {
 
-	if len(haystack) == 0 || len(needle) == 0 || len(needle) > len(haystack) {
+	needleAsRunes := []rune(needle)
+	lenNeedle := len(needleAsRunes)
+
+	if len(haystack) == 0 || len(needle) == 0 || lenNeedle > len(haystack) {
 		return []int{}
 	}
 
 	items := make([]int, 0, 8)
 
 	for i := 0; i < len(haystack); i++ {
-		for j := 0; j < len(needle); j++ {
-			if haystack[i+j] != rune(needle[j]) {
+		for j := 0; j < lenNeedle; j++ {
+			if haystack[i+j] != needleAsRunes[j] {
 				break
 			}
 
-			if j == len(needle)-1 {
+			if j == lenNeedle-1 {
 				items = append(items, i)
 			}
 		}
@@ -26,16 +29,18 @@ func findAll(haystack []rune, needle string) []int {
 
 // Returns the first occurrence of haystack in needle or -1 if not found.
 func findFirst(haystack []rune, needle string) int {
-	if len(haystack) == 0 || len(needle) == 0 || len(needle) > len(haystack) {
+	needleAsRunes := []rune(needle)
+	lenNeedle := len(needleAsRunes)
+
+	if len(haystack) == 0 || len(needle) == 0 || lenNeedle > len(haystack) {
 		return -1
 	}
 
 	lenHaystack := len(haystack)
-	lenNeedle := len(needle)
 
 	for i := 0; i <= lenHaystack-lenNeedle; i++ {
 		for j := 0; j < lenNeedle; j++ {
-			if haystack[i+j] != rune(needle[j]) {
+			if haystack[i+j] != needleAsRunes[j] {
 				break
 			}
 
@@ -50,16 +55,18 @@ func findFirst(haystack []rune, needle string) int {
 
 // Returns the last occurrence of haystack in needle or -1 if not found.
 func findLast(haystack []rune, needle string) int {
-	if len(haystack) == 0 || len(needle) == 0 || len(needle) > len(haystack) {
+	needleAsRunes := []rune(needle)
+	lenNeedle := len(needleAsRunes)
+
+	if len(haystack) == 0 || len(needle) == 0 || lenNeedle > len(haystack) {
 		return -1
 	}
 
 	lenHaystack := len(haystack)
-	lenNeedle := len(needle)
 
-	for i := lenHaystack - lenNeedle + 1; i >= 0; i-- {
+	for i := lenHaystack - lenNeedle; i >= 0; i-- {
 		for j := 0; j < lenNeedle; j++ {
-			if haystack[i+j] != rune(needle[j]) {
+			if haystack[i+j] != needleAsRunes[j] {
 				break
 			}
 

stringbuilder.go

@@ -16,7 +16,7 @@ func NewStringBuilder(initialCapacity int) *StringBuilder {
 func NewStringBuilderFromString(text string) *StringBuilder {
 	return &StringBuilder{
 		data:     []rune(text),
-		position: len(text),
+		position: len([]rune(text)),
 	}
 }
 
@@ -155,26 +155,29 @@ func (s *StringBuilder) Replace(oldValue string, newValue string) {
 
 	occurrences := s.FindAll(oldValue)
 
-	delta := len(newValue) - len(oldValue)
+	oldValueRunes := []rune(oldValue)
+	newValueRunes := []rune(newValue)
+
+	delta := len(newValueRunes) - len(oldValueRunes)
 
 	for i := 0; i < len(occurrences); i++ {
 		index := occurrences[i] + delta*i
 
 		// newValue is smaller than old value
 		// We can insert the slice and remove the overhead
 		if delta < 0 {
-			copy(s.data[index:], []rune(newValue))
-			s.Remove(index+len(newValue), -delta)
+			copy(s.data[index:], newValueRunes)
+			s.Remove(index+len(newValueRunes), -delta)
 		} else if delta == 0 {
 			// Same length -> We can just replace the memory slice
-			copy(s.data[index:], []rune(newValue))
+			copy(s.data[index:], newValueRunes)
 		} else {
 			// newValue is larger than the old value
 			// First add until the old memory region
 			// and insert afterwards the rest
-			x := len(oldValue)
-			copy(s.data[index:], []rune(newValue[:x]))
-			s.Insert(index+len(oldValue), newValue[len(oldValue):])
+			oldLen := len(oldValueRunes)
+			copy(s.data[index:], []rune(newValueRunes[:oldLen]))
+			s.Insert(index+oldLen, string(newValueRunes[len(oldValueRunes):]))
 		}
 	}
 }

stringbuilder_test.go

@@ -217,6 +217,7 @@ func TestFindFirst(t *testing.T) {
 		{"Empty needle", "n", "", -1},
 		{"Needle longer than haystack", "a", "ab", -1},
 		{"Hello in Hello World", "Hello World", "Hello", 0},
+		{"ö in Helöö", "Hellöö", "ö", 4},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
@@ -239,6 +240,7 @@ func TestFindLast(t *testing.T) {
 		{"Empty needle", "n", "", -1},
 		{"Needle longer than haystack", "a", "ab", -1},
 		{"Hello in Hello World", "Hello World", "Hello", 0},
+		{"ö in Helöö", "Hellöö", "ö", 5},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
@@ -261,6 +263,7 @@ func TestFindAll(t *testing.T) {
 		{"Empty needle", "n", "", []int{}},
 		{"Needle longer than haystack", "a", "ab", []int{}},
 		{"Hello in Hello World", "Hello World", "Hello", []int{0}},
+		{"ö in Helöö", "Hellöö", "ö", []int{4, 5}},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
@@ -292,7 +295,9 @@ func TestReplace(t *testing.T) {
 	}{
 		{"Replace Hello with Hallo", "Hello World", "Hello", "Hallo", "Hallo World"},
 		{"Replace Hello with Ha", "Hello World", "Hello", "Ha", "Ha World"},
-		{"Replace Hello with Hallöchen", "Hello World", "Hello", "Hallochen", "Hallochen World"},
+		{"Replace Hello with Hallochen", "Hello World", "Hello", "Hallochen", "Hallochen World"},
+		{"Replace Hello with Hallöchen", "Hello World", "Hello", "Hallöchen", "Hallöchen World"},
+		{"Replace ö with ä", "äö", "ö", "ä", "ää"},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {