Add proper support for unicode characters

src/commonMain/kotlin/io/github/optimumcode/json/schema/internal/factories/string/LengthAssertion.kt

@@ -5,6 +5,7 @@ import io.github.optimumcode.json.schema.ErrorCollector
 import io.github.optimumcode.json.schema.ValidationError
 import io.github.optimumcode.json.schema.internal.AssertionContext
 import io.github.optimumcode.json.schema.internal.JsonSchemaAssertion
+import io.github.optimumcode.json.schema.internal.factories.string.util.codePointCount
 import kotlinx.serialization.json.JsonElement
 import kotlinx.serialization.json.JsonPrimitive
 import kotlinx.serialization.json.contentOrNull
@@ -20,14 +21,15 @@ internal class LengthAssertion(
  return true
  }
  val content = element.contentOrNull ?: return true
- if (check(content.length, lengthValue)) {
+ val codePointCount = content.codePointCount()
+ if (check(codePointCount, lengthValue)) {
  return true
  }
  errorCollector.onError(
  ValidationError(
  schemaPath = path,
  objectPath = context.objectPath,
- message = "string length (${content.length}) $errorMessage $lengthValue",
+ message = "string length ($codePointCount) $errorMessage $lengthValue",
  ),
  )
  return false

src/commonMain/kotlin/io/github/optimumcode/json/schema/internal/factories/string/util/UnicodeUtil.kt

@@ -0,0 +1,21 @@
+package io.github.optimumcode.json.schema.internal.factories.string.util
+
+internal fun CharSequence.codePointCount(): Int {
+ val endIndex = length
+ var index = 0
+ var count = 0
+ while (index < endIndex) {
+ val firstChar = this[index]
+ index++
+ if (firstChar.isHighSurrogate() && index < endIndex) {
+ val nextChar = this[index]
+ if (nextChar.isLowSurrogate()) {
+ index++
+ }
+ }
+
+ count++
+ }
+
+ return count
+}

src/commonTest/kotlin/io/github/optimumcode/json/schema/assertions/string/JsonSchemaMaxLengthValidationTest.kt

@@ -4,6 +4,7 @@ import io.github.optimumcode.json.pointer.JsonPointer
 import io.github.optimumcode.json.schema.JsonSchema
 import io.github.optimumcode.json.schema.ValidationError
 import io.github.optimumcode.json.schema.base.KEY
+import io.github.optimumcode.json.schema.internal.factories.string.util.codePointCount
 import io.kotest.core.spec.style.FunSpec
 import io.kotest.matchers.collections.shouldContainExactly
 import io.kotest.matchers.collections.shouldHaveSize
@@ -32,6 +33,8 @@ class JsonSchemaMaxLengthValidationTest : FunSpec() {
  "╒",
  "V",
  "",
+ "💩".repeat(20),
+ "💩",
  )
  for (str in validStrings) {
  test("'$str' passes validation") {
@@ -46,6 +49,7 @@ class JsonSchemaMaxLengthValidationTest : FunSpec() {
  "EFDzZMRawYGD9eNfknAUB",
  "⌻ⲝ⣞ℤ⸍⠗⠜ↈ✋☧⾛✩ⓥ⇩⡽⚘\u20FC◭┐⥸⒗",
  "⠺⪒⑸⋶⥠⇀⨑⨋ⅸ⥼\u245F⏇Ⓙⴷ⻘⢢≧\u20C8⬫⡜⸁",
+ "💩".repeat(21),
  )
  for (str in invalidStrings) {
  test("'$str' does not pass validation") {
@@ -56,7 +60,7 @@ class JsonSchemaMaxLengthValidationTest : FunSpec() {
  ValidationError(
  schemaPath = JsonPointer("/maxLength"),
  objectPath = JsonPointer.ROOT,
- message = "string length (${str.length}) must be less or equal to 20",
+ message = "string length (${str.codePointCount()}) must be less or equal to 20",
  ),
  )
  }

src/commonTest/kotlin/io/github/optimumcode/json/schema/assertions/string/JsonSchemaMinLengthValidationTest.kt

@@ -4,6 +4,7 @@ import io.github.optimumcode.json.pointer.JsonPointer
 import io.github.optimumcode.json.schema.JsonSchema
 import io.github.optimumcode.json.schema.ValidationError
 import io.github.optimumcode.json.schema.base.KEY
+import io.github.optimumcode.json.schema.internal.factories.string.util.codePointCount
 import io.kotest.core.spec.style.FunSpec
 import io.kotest.matchers.collections.shouldContainExactly
 import io.kotest.matchers.collections.shouldHaveSize
@@ -31,6 +32,8 @@ class JsonSchemaMinLengthValidationTest : FunSpec() {
  "JpEblYiJE57H70qGNXs",
  "ⅵ┡\u243A⢻␀⁾⡪∛⫑⏽",
  "Si1kaAhdpS",
+ "💩".repeat(11),
+ "💩".repeat(10),
  )
  for (str in validStrings) {
  test("'$str' passes validation") {
@@ -47,6 +50,7 @@ class JsonSchemaMinLengthValidationTest : FunSpec() {
  " ⍘↽♔⚪➕ⷰ➖",
  "⧦",
  "",
+ "💩".repeat(9),
  )
  for (str in invalidStrings) {
  test("'$str' does not pass validation") {
@@ -57,7 +61,7 @@ class JsonSchemaMinLengthValidationTest : FunSpec() {
  ValidationError(
  schemaPath = JsonPointer("/minLength"),
  objectPath = JsonPointer.ROOT,
- message = "string length (${str.length}) must be greater or equal to 10",
+ message = "string length (${str.codePointCount()}) must be greater or equal to 10",
  ),
  )
  }