Added utf-8 BOM removal (#630)

* Added utf-8 BOM removal * Restore UTF8 BOM after formatting if it was present * Fixed code style
pinterest · Nov 12, 2019 · 6864374 · 6864374
1 parent baac14c
commit 6864374
Showing 1 changed file with 18 additions and 3 deletions.
diff --git a/ktlint-core/src/main/kotlin/com/pinterest/ktlint/core/KtLint.kt b/ktlint-core/src/main/kotlin/com/pinterest/ktlint/core/KtLint.kt
@@ -48,6 +48,7 @@ object KtLint {
     val ANDROID_USER_DATA_KEY = Key<Boolean>("ANDROID")
     val FILE_PATH_USER_DATA_KEY = Key<String>("FILE_PATH")
     val DISABLED_RULES = Key<Set<String>>("DISABLED_RULES")
+    private const val UTF8_BOM = "\uFEFF"
     const val STDIN_FILE = "<stdin>"
 
     private val psiFileFactory: PsiFileFactory
@@ -130,7 +131,7 @@ object KtLint {
      * @throws RuleExecutionException in case of internal failure caused by a bug in rule implementation
      */
     fun lint(params: Params) {
-        val normalizedText = params.text.replace("\r\n", "\n").replace("\r", "\n")
+        val normalizedText = normalizeText(params.text)
         val positionByOffset = calculateLineColByOffset(normalizedText)
         val psiFileName = if (params.script) "file.kts" else "file.kt"
         val psiFile = psiFileFactory.createFileFromText(psiFileName, KotlinLanguage.INSTANCE, normalizedText) as KtFile
@@ -169,6 +170,13 @@ object KtLint {
             .forEach { e -> params.cb(e, false) }
     }
 
+    private fun normalizeText(text: String): String {
+        return text
+            .replace("\r\n", "\n")
+            .replace("\r", "\n")
+            .replaceFirst(UTF8_BOM, "")
+    }
+
     private fun userDataResolver(editorConfigPath: String?, debug: Boolean): (String?) -> Map<String, String> {
         if (editorConfigPath != null) {
             val userData = (
@@ -351,7 +359,8 @@ object KtLint {
      * @throws RuleExecutionException in case of internal failure caused by a bug in rule implementation
      */
     fun format(params: Params): String {
-        val normalizedText = params.text.replace("\r\n", "\n").replace("\r", "\n")
+        val hasUTF8BOM = params.text.startsWith(UTF8_BOM)
+        val normalizedText = normalizeText(params.text)
         val positionByOffset = calculateLineColByOffset(normalizedText)
         val psiFileName = if (params.script) "file.kts" else "file.kt"
         val psiFile = psiFileFactory.createFileFromText(psiFileName, KotlinLanguage.INSTANCE, normalizedText) as KtFile
@@ -414,7 +423,13 @@ object KtLint {
                 .sortedWith(Comparator { (l), (r) -> if (l.line != r.line) l.line - r.line else l.col - r.col })
                 .forEach { (e, corrected) -> params.cb(e, corrected) }
         }
-        return if (mutated) rootNode.text.replace("\n", determineLineSeparator(params.text, params.userData)) else params.text
+
+        if (!mutated) {
+            return params.text
+        }
+
+        return if (hasUTF8BOM) UTF8_BOM else "" + // Restore UTF8 BOM if it was present
+            rootNode.text.replace("\n", determineLineSeparator(params.text, params.userData))
     }
 
     private fun determineLineSeparator(fileContent: String, userData: Map<String, String>): String {