0.5.1: more utils and optimizations

Author: magicprinc

gradle.properties

@@ -1,6 +1,6 @@
 group = com.trivago
 
-version = 0.5
+version = 0.5.1
 
 org.gradle.jvmargs = -ea -Dfile.encoding=UTF-8 -XX:+UseG1GC -Xms512m -Xmx1G -Djava.net.preferIPv4Stack=true
 systemProp.file.encoding = UTF-8

src/main/java/com/trivago/fastutilconcurrentwrapper/io/BAIS.java

@@ -8,6 +8,7 @@
 import org.jspecify.annotations.Nullable;
 
 import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
 import java.io.DataInputStream;
 import java.io.IOException;
 import java.io.ObjectInput;
@@ -16,6 +17,8 @@
 import java.util.PrimitiveIterator;
 import java.util.UUID;
 
+import static java.nio.charset.StandardCharsets.*;
+
 /**
  @see java.io.ByteArrayInputStream
 
@@ -39,8 +42,8 @@ public BAIS (byte[] array, @PositiveOrZero int offset, @PositiveOrZero int maxLe
 
 	public BAIS (byte[] array){ super(array); }//new
 
-	public BAIS (BAOS baos) {
-		super(baos.array(), 0, baos.size());
+	public BAIS (ByteArrayOutputStream baos) {
+		super(baos instanceof BAOS us ? us.array() : baos.toByteArray(), 0, baos.size());
 	}//new
 
 	/// The array backing the input stream. capacity = array().length
@@ -280,15 +283,23 @@ public Object readObject () throws ClassNotFoundException, IOException {
 	/// @see BAOS#writeBytes(String)
 	/// @see java.nio.charset.StandardCharsets#ISO_8859_1
 	public String readLatin1String (@PositiveOrZero int strLen) {
+		strLen = Math.min(strLen, available());
 		String s = new String(buf, 0,  pos, strLen);
 		pos += strLen;
 		return s;
 	}
 
+	/// @see java.nio.charset.StandardCharsets#UTF_16BE
+	public String readUTF16String (@PositiveOrZero int strLen) {
+		int byteLen = Math.min(strLen << 1, available());//*2
+		String s = new String(buf, pos, byteLen, UTF_16BE);
+		pos += byteLen;
+		return s;
+	}
+
 	@Override
 	public boolean hasNext () {
-		// available() > 0 == (length - position) > 0
-		return pos < count;
+		return pos < count;// position < length: available() > 0 == (length - position) > 0
 	}
 
 	@Override

src/main/java/com/trivago/fastutilconcurrentwrapper/io/BAOS.java

@@ -19,6 +19,8 @@
 import java.util.Objects;
 import java.util.UUID;
 
+import static java.nio.charset.StandardCharsets.*;
+
 /**
  Simple, fast byte-array output stream that exposes the backing array.
 
@@ -80,6 +82,24 @@ public BAOS (byte[] a) {
 		buf = a;
 	}//new
 
+	public BAOS (byte[] a, @PositiveOrZero int length) {
+		super(0);
+		buf = a;
+		count = Math.min(length, a.length);
+	}//new
+
+	public BAOS (ByteArrayOutputStream baos) {
+		super(0);
+		count = baos.size();
+		if (baos instanceof BAOS us){
+			buf = us.array();
+			position = us.writerIndex();
+		} else {
+			buf = baos.toByteArray();
+			position = count;
+		}
+	}//new ~ clone
+
 	/// Marks this array output stream as empty.
 	@Override public void reset (){ count = 0;  position = 0; }
 
@@ -318,23 +338,23 @@ public void writeObject(Object obj) throws IOException {
 
 	@Override
 	public Appendable append (CharSequence csq) {
-		int len;
-		if (csq != null && (len = csq.length()) > 0)
-				append(csq, 0, len);
+		append(csq, 0, csq.length());
 		return this;
 	}
 
 	/// @see #write(byte[], int, int)
 	/// @see #writeBytes(byte[])
 	@Override
 	public Appendable append (CharSequence csq, int start, int end) {
-		int len = end - start;
+		final int len = end - start;
 		Objects.checkFromIndexSize(start, len, csq.length());
-		if (position + len > buf.length)
-				buf = ByteArrays.grow(buf, position + len, position);
-		for (int i = 0; i < len; i++)
-				buf[position + i] = (byte) csq.charAt(start + i);
-		position += len;
+		final int len2 = len << 1;
+		if (position + len2 > buf.length)
+				buf = ByteArrays.grow(buf, position + len2, position);
+
+		JBytes.DirectByteArrayAccess.copyCharsToByteArray(csq, start,  buf, position,  len);
+
+		position += len2;
 		if (count < position) count = position;
 		return this;
 	}
@@ -348,4 +368,34 @@ public Appendable append (char c) {
 		if (count < position) count = position;
 		return this;
 	}
+
+	@Override
+	public boolean equals (Object obj) {
+		if (this == obj){
+			return true;
+		} else if (obj instanceof BAOS baos){
+			return java.util.Arrays.equals(buf,0,count, baos.array(),0,baos.size());
+		} else if (obj instanceof ByteArrayOutputStream baos){
+			return java.util.Arrays.equals(buf,0,count, baos.toByteArray(),0,baos.size());
+		} else if (obj instanceof byte[] a){
+			return java.util.Arrays.equals(buf,0,count, a,0,a.length);
+		} else if (obj instanceof CharSequence cs){
+			byte[] a = cs.toString().getBytes(UTF_16BE);
+			return java.util.Arrays.equals(buf,0,count, a,0,a.length);
+		} else {
+			return false;
+		}
+	}
+
+	/// @see java.util.Arrays#hashCode(byte[])
+	/// @see jdk.internal.util.ArraysSupport#hashCode(int, byte[], int, int)
+	/// @see java.lang.String#hashCode
+	@Override
+	public int hashCode () {
+		int result = 1;
+		for (int i = 0; i < count; i++){
+			result = 31 * result + buf[i];
+		}
+		return result;
+	}
 }

src/main/java/com/trivago/fastutilconcurrentwrapper/util/JBytes.java

@@ -225,6 +225,7 @@ public static long bytesToLong (byte[] bytes) {
 	 * @see java.io.DataInputStream
 	 * @see jdk.internal.util.ByteArray
 	 * @see jdk.internal.util.ByteArrayLittleEndian
+	 * @see com.dynatrace.hash4j.internal.ByteArrayUtil
 	 */
 	@NoArgsConstructor(access = AccessLevel.PRIVATE)
 	public static final class DirectByteArrayAccess {
@@ -601,5 +602,68 @@ public static void setDoubleRaw(byte[] array, int offset, double value) {
 		private static VarHandle create(Class<?> viewArrayClass) {
 			return MethodHandles.byteArrayViewVarHandle(viewArrayClass, ByteOrder.BIG_ENDIAN);
 		}
+
+		/**
+		 * Reads a {@code long} value from a {@link CharSequence} with given offset.
+		 *
+		 * @param cs a char sequence
+		 * @param off an offset
+		 * @return the value
+		 */
+		public static long getLong (CharSequence cs, int off) {
+			return ((long) cs.charAt(off) << 48)
+					| ((long) cs.charAt(off + 1) << 32)
+					| ((long) cs.charAt(off + 2) << 16)
+					| cs.charAt(off + 3);
+		}
+
+		/**
+		 * Reads an {@code int} value from a {@link CharSequence} with given offset.
+		 *
+		 * @param cs a char sequence
+		 * @param off an offset
+		 * @return the value
+		 */
+		public static int getInt (CharSequence cs, int off) {
+			return ((int) cs.charAt(off) << 16)
+					| (int) cs.charAt(off + 1);
+		}
+
+		/**
+		 * Copies a given number of characters from a {@link CharSequence} into a byte array.
+		 *
+		 * @param cs a char sequence
+		 * @param offsetCharSequence an offset for the char sequence
+		 * @param toByteArray a byte array
+		 * @param offsetByteArray an offset for the byte array
+		 * @param numChars the number of characters to copy
+		 */
+		public static void copyCharsToByteArray (
+			CharSequence cs,
+			int offsetCharSequence,
+			byte[] toByteArray,
+			int offsetByteArray,
+			int numChars
+		){
+			for (int charIdx = 0; charIdx <= numChars - 4; charIdx += 4){
+				setLong(toByteArray,
+					offsetByteArray + (charIdx << 1),
+					getLong(cs, offsetCharSequence + charIdx));
+			}
+
+			if ((numChars & 2) != 0){
+				int charIdx = numChars & 0xFFFF_FFFC;
+				setInt(toByteArray,
+					offsetByteArray + (charIdx << 1),
+					getInt(cs, offsetCharSequence + charIdx));
+			}
+
+			if ((numChars & 1) != 0){
+				int charIdx = numChars & 0xFFFF_FFFE;
+				setChar(toByteArray,
+					offsetByteArray + (charIdx << 1),
+					cs.charAt(offsetCharSequence + charIdx));
+			}
+		}
 	}
 }

src/test/java/com/trivago/fastutilconcurrentwrapper/io/BAISTest.java

@@ -5,10 +5,12 @@
 import org.junit.jupiter.api.DisplayName;
 import org.junit.jupiter.api.Test;
 
+import java.io.IOException;
 import java.io.UTFDataFormatException;
 import java.util.Arrays;
 import java.util.UUID;
 
+import static java.nio.charset.StandardCharsets.*;
 import static org.junit.jupiter.api.Assertions.*;
 
 /// @see BAIS
@@ -326,4 +328,35 @@ void testSkipBytes() {
 		assertEquals(3, skipped);
 		assertEquals(3, stream.position());
 	}
+
+	@Test
+	void _transferTo () throws IOException {
+		String s = "Hello World!";
+		var is = new BAIS(s.getBytes(ISO_8859_1));
+		var w = new BAOS();
+		is.transferTo(w);
+		assertEquals(s, w.toString());
+		assertEquals(0, is.available());
+		is.reset();
+		assertEquals(is.available(), w.length());
+	}
+
+	@Test
+	void _iterator () throws IOException {
+		String s = "Hello World!";
+		var is = new BAIS(s.getBytes(ISO_8859_1));
+		var w = new BAOS(is.array(), is.limit());
+		assertEquals(s, w.toString());
+		assertEquals(is.available(), w.length());
+
+		assertTrue(is.hasNext());
+		assertEquals('H', is.nextInt());
+		assertTrue(is.hasNext());
+		assertEquals('e', is.nextInt());
+
+		is.reset();
+		val sb = new StringBuilder();
+		is.forEachRemaining((int ch)->sb.append((char) ch));
+		assertEquals(s, sb.toString());
+	}
 }

src/test/java/com/trivago/fastutilconcurrentwrapper/io/BAOSTest.java

@@ -1,6 +1,7 @@
 package com.trivago.fastutilconcurrentwrapper.io;
 
 import com.trivago.fastutilconcurrentwrapper.util.JBytes;
+import it.unimi.dsi.fastutil.bytes.ByteArrays;
 import lombok.val;
 import org.junit.jupiter.api.BeforeEach;
 import org.junit.jupiter.api.DisplayName;
@@ -22,6 +23,7 @@
 import java.util.UUID;
 import java.util.stream.IntStream;
 
+import static java.nio.charset.StandardCharsets.*;
 import static org.junit.jupiter.api.Assertions.*;
 
 ///@see BAOS
@@ -607,7 +609,7 @@ void asStr () {
 
 		x.reset();
 		x.writeChars("Я$€");
-		assertEquals("Я$€", x.toString(StandardCharsets.UTF_16BE));
+		assertEquals("Я$€", x.toString(UTF_16BE));
 	}
 
 	@Test
@@ -1077,4 +1079,51 @@ void _ascii () {
 		assertEquals(sss, baos.toString());
 		assertEquals(sss, baos.toString(StandardCharsets.ISO_8859_1));
 	}
+
+	@Test
+	@SuppressWarnings({"ConstantValue", "EqualsBetweenInconvertibleTypes", "EqualsWithItself"})
+	void _appendable () {
+		var w = new BAOS();
+		assertEquals(1, w.hashCode());
+		assertEquals(1, Arrays.hashCode(ByteArrays.EMPTY_ARRAY));
+
+		var s1 = "Проверка связи!";
+		w.append(s1);
+		w.append("");// nop
+		w.append("012 test 012", 3, 9);
+		w.append("012 test 012", 3, 3);// nop
+		w.append('!');
+		assertEquals(s1+" test !", w.toString(UTF_16BE));
+		assertEquals(44, w.length());
+
+		val s3 = "🚀".repeat(50_000);
+		w.append(s3);
+		String total = s1 + " test !" + s3;
+		assertEquals(total, w.toString(UTF_16BE));
+		assertEquals(200_044, w.length());// 44+
+		assertEquals(2, "🚀".length());// 2 chars == 4 bytes
+
+		assertEquals(542738558, w.hashCode());
+		assertEquals(w.hashCode(), Arrays.hashCode(w.toByteArray()));
+		assertEquals(w.hashCode(), Arrays.hashCode(total.getBytes(UTF_16BE)));
+
+		boolean eq = w.equals(total);
+		assertTrue(eq);
+		eq = w.equals(total.getBytes(UTF_16BE));
+		assertTrue(eq);
+		eq = w.equals(w);
+		assertTrue(eq);
+		eq = w.equals(null);
+		assertFalse(eq);
+
+		var clone = new BAOS(w);
+		assertEquals(clone, w);
+		assertEquals(w.hashCode(), clone.hashCode());
+		// verify with BAIS
+		var is = new BAIS(w);
+
+		assertEquals(s1, is.readUTF16String(s1.length()));
+		assertEquals(" test !", is.readUTF16String(7));
+		assertEquals(s3, is.readUTF16String(s3.length()));
+	}
 }