Avoid allocating large buffers in File.writeText #KT-51058

Also fixes File.appendText and Path.write/appendText. Merge-request: KT-MR-12804 Merged-by: Abduqodiri Qurbonzoda <abduqodiri.qurbonzoda@jetbrains.com>
2023-11-08 11:18:56 +00:00
parent 54ab3d39db
commit 999b0f7099
3 changed files with 191 additions and 5 deletions
@@ -5,12 +5,130 @@

 package kotlin.jdk7.test

+import java.io.ByteArrayOutputStream
+import java.nio.CharBuffer
+import java.nio.charset.Charset
+import java.nio.file.Path
 import java.nio.file.StandardOpenOption
 import kotlin.io.path.*
 import kotlin.random.Random
 import kotlin.test.*

 class PathReadWriteTest : AbstractPathTest() {
+
+    private fun String.encodeToByteArray(charset: Charset): ByteArray {
+        val out = ByteArrayOutputStream()
+        out.writer(charset).use { it.append(this) }
+        return out.toByteArray()
+    }
+
+    private val hexFormat = HexFormat {
+        bytes.bytesPerLine = 32
+        bytes.bytesPerGroup = 8
+    }
+
+    private fun Path.testContentEquals(expectedContent: ByteArray, charset: Charset) {
+        val expected = expectedContent.toHexString(hexFormat)
+        val actualContent = readBytes()
+        val actual = actualContent.toHexString(hexFormat)
+        assertEquals(expected, actual, "$charset. Expected size is ${expectedContent.size}, actual size is ${actualContent.size}")
+    }
+
+    private fun Path.testWriteText(text: String, charset: Charset) {
+        // Path.writeText
+        val encodedText = text.encodeToByteArray(charset)
+
+        writeText(text, charset)
+        testContentEquals(encodedText, charset)
+
+        writeText(StringBuilder(text), charset)
+        testContentEquals(encodedText, charset)
+
+        val position = 1.coerceAtMost(text.length)
+        val limit = (text.length - 1).coerceAtLeast(position)
+        val charBuffer = CharBuffer.wrap(text, position, limit)
+        val encodedCharBuffer = text.substring(position, limit).encodeToByteArray(charset)
+
+        writeText(charBuffer, charset)
+        testContentEquals(encodedCharBuffer, charset)
+        assertEquals(position, charBuffer.position())
+        assertEquals(limit, charBuffer.limit())
+        assertEquals(text.length, charBuffer.capacity())
+
+        // Path.appendText
+        val prefix = "_"
+        val encodedPrefix = prefix.encodeToByteArray(charset)
+
+        writeText(prefix, charset)
+        appendText(text, charset)
+        testContentEquals(encodedPrefix + encodedText, charset)
+
+        writeText(prefix, charset)
+        appendText(StringBuilder(text), charset)
+        testContentEquals(encodedPrefix + encodedText, charset)
+
+        writeText(prefix, charset)
+        appendText(charBuffer, charset)
+        testContentEquals(encodedPrefix + encodedCharBuffer, charset)
+        assertEquals(position, charBuffer.position())
+        assertEquals(limit, charBuffer.limit())
+        assertEquals(text.length, charBuffer.capacity())
+
+        // File.writeText
+        toFile().writeText(text, charset)
+        testContentEquals(encodedText, charset)
+
+        // File.appendText
+        toFile().writeText(prefix, charset)
+        toFile().appendText(text, charset)
+        testContentEquals(encodedPrefix + encodedText, charset)
+    }
+
+    @Test fun writeText() {
+        val charsets = listOf(
+            Charsets.UTF_8,
+            Charsets.UTF_16,
+            Charsets.UTF_32,
+            Charsets.ISO_8859_1,
+            Charsets.US_ASCII,
+        )
+
+        val highSurrogate = Char.MIN_HIGH_SURROGATE
+        val lowSurrogate = Char.MIN_LOW_SURROGATE
+
+        val smallString = "Hello"
+
+        val chunkSize = DEFAULT_BUFFER_SIZE
+        val string = "k".repeat(chunkSize - 1)
+
+        val path = createTempFile().cleanup()
+
+        for (charset in charsets) {
+            path.testWriteText("$highSurrogate", charset)
+
+            path.testWriteText("$lowSurrogate", charset)
+
+            path.testWriteText("$smallString$highSurrogate", charset)
+
+            path.testWriteText("$smallString$lowSurrogate", charset)
+
+            path.testWriteText("$string$highSurrogate", charset)
+
+            path.testWriteText("$string$lowSurrogate", charset)
+
+            path.testWriteText("$string$highSurrogate$lowSurrogate$string", charset)
+
+            path.testWriteText("$string$lowSurrogate$highSurrogate$string", charset)
+
+            path.testWriteText(
+                "$string$highSurrogate$lowSurrogate${string.substring(2)}$highSurrogate$lowSurrogate",
+                charset
+            )
+
+            path.testWriteText("$string$lowSurrogate$highSurrogate$lowSurrogate$string", charset)
+        }
+    }
+
    @Test
    fun appendText() {
        val file = createTempFile().cleanup()