/*
* Copyright (c) 2014, 2018, Oracle and/or its affiliates. All rights reserved.
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
* This code is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License version 2 only, as
* published by the Free Software Foundation. Oracle designates this
* particular file as subject to the "Classpath" exception as provided
* by Oracle in the LICENSE file that accompanied this code.
*
* This code is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* version 2 for more details (a copy is included in the LICENSE file that
* accompanied this code).
*
* You should have received a copy of the GNU General Public License version
* 2 along with this work; if not, write to the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
*
* Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
* or visit www.oracle.com if you need additional information or have any
* questions.
*/
package jdk.internal.net.http.hpack;
import jdk.internal.net.http.hpack.HPACK.Logger;
import java.nio.ByteBuffer;
import java.nio.ReadOnlyBufferException;
import java.util.LinkedList;
import java.util.List;
import java.util.concurrent.atomic.AtomicLong;
import static java.lang.String.format;
import static java.util.Objects.requireNonNull;
import static jdk.internal.net.http.hpack.HPACK.Logger.Level.EXTRA;
import static jdk.internal.net.http.hpack.HPACK.Logger.Level.NORMAL;
/**
* Encodes headers to their binary representation.
*
* <p> Typical lifecycle looks like this:
*
* <p> {@link #Encoder(int) new Encoder}
* ({@link #setMaxCapacity(int) setMaxCapacity}?
* {@link #encode(ByteBuffer) encode})*
*
* <p> Suppose headers are represented by {@code Map<String, List<String>>}.
* A supplier and a consumer of {@link ByteBuffer}s in forms of
* {@code Supplier<ByteBuffer>} and {@code Consumer<ByteBuffer>} respectively.
* Then to encode headers, the following approach might be used:
*
* <pre>{@code
* for (Map.Entry<String, List<String>> h : headers.entrySet()) {
* String name = h.getKey();
* for (String value : h.getValue()) {
* encoder.header(name, value); // Set up header
* boolean encoded;
* do {
* ByteBuffer b = buffersSupplier.get();
* encoded = encoder.encode(b); // Encode the header
* buffersConsumer.accept(b);
* } while (!encoded);
* }
* }
* }</pre>
*
* <p> Though the specification <a href="https://tools.ietf.org/html/rfc7541#section-2">does not define</a>
* how an encoder is to be implemented, a default implementation is provided by
* the method {@link #header(CharSequence, CharSequence, boolean)}.
*
* <p> To provide a custom encoding implementation, {@code Encoder} has to be
* extended. A subclass then can access methods for encoding using specific
* representations (e.g. {@link #literal(int, CharSequence, boolean) literal},
* {@link #indexed(int) indexed}, etc.)
*
* @apiNote
*
* <p> An Encoder provides an incremental way of encoding headers.
* {@link #encode(ByteBuffer)} takes a buffer a returns a boolean indicating
* whether, or not, the buffer was sufficiently sized to hold the
* remaining of the encoded representation.
*
* <p> This way, there's no need to provide a buffer of a specific size, or to
* resize (and copy) the buffer on demand, when the remaining encoded
* representation will not fit in the buffer's remaining space. Instead, an
* array of existing buffers can be used, prepended with a frame that encloses
* the resulting header block afterwards.
*
* <p> Splitting the encoding operation into header set up and header encoding,
* separates long lived arguments ({@code name}, {@code value},
* {@code sensitivity}, etc.) from the short lived ones (e.g. {@code buffer}),
* simplifying each operation itself.
*
* @implNote
*
* <p> The default implementation does not use dynamic table. It reports to a
* coupled Decoder a size update with the value of {@code 0}, and never changes
* it afterwards.
*
* @since 9
*/
public class Encoder {
private static final AtomicLong ENCODERS_IDS = new AtomicLong();
// TODO: enum: no huffman/smart huffman/always huffman
private static final boolean DEFAULT_HUFFMAN = true;
private final Logger logger;
private final long id;
private final IndexedWriter indexedWriter = new IndexedWriter();
private final LiteralWriter literalWriter = new LiteralWriter();
private final LiteralNeverIndexedWriter literalNeverIndexedWriter
= new LiteralNeverIndexedWriter();
private final LiteralWithIndexingWriter literalWithIndexingWriter
= new LiteralWithIndexingWriter();
private final SizeUpdateWriter sizeUpdateWriter = new SizeUpdateWriter();
private final BulkSizeUpdateWriter bulkSizeUpdateWriter
= new BulkSizeUpdateWriter();
private BinaryRepresentationWriter writer;
private final HeaderTable headerTable;
private boolean encoding;
private int maxCapacity;
private int currCapacity;
private int lastCapacity;
private long minCapacity;
private boolean capacityUpdate;
private boolean configuredCapacityUpdate;
/**
* Constructs an {@code Encoder} with the specified maximum capacity of the
* header table.
*
* <p> The value has to be agreed between decoder and encoder out-of-band,
* e.g. by a protocol that uses HPACK
* (see <a href="https://tools.ietf.org/html/rfc7541#section-4.2">4.2. Maximum Table Size</a>).
*
* @param maxCapacity
* a non-negative integer
*
* @throws IllegalArgumentException
* if maxCapacity is negative
*/
public Encoder(int maxCapacity) {
id = ENCODERS_IDS.incrementAndGet();
this.logger = HPACK.getLogger().subLogger("Encoder#" + id);
if (logger.isLoggable(NORMAL)) {
logger.log(NORMAL, () -> format("new encoder with maximum table size %s",
maxCapacity));
}
if (logger.isLoggable(EXTRA)) {
/* To correlate with logging outside HPACK, knowing
hashCode/toString is important */
logger.log(EXTRA, () -> {
String hashCode = Integer.toHexString(
System.identityHashCode(this));
/* Since Encoder can be subclassed hashCode AND identity
hashCode might be different. So let's print both. */
return format("toString='%s', hashCode=%s, identityHashCode=%s",
toString(), hashCode(), hashCode);
});
}
if (maxCapacity < 0) {
throw new IllegalArgumentException(
"maxCapacity >= 0: " + maxCapacity);
}
// Initial maximum capacity update mechanics
minCapacity = Long.MAX_VALUE;
currCapacity = -1;
setMaxCapacity0(maxCapacity);
headerTable = new HeaderTable(lastCapacity, logger.subLogger("HeaderTable"));
}
/**
* Sets up the given header {@code (name, value)}.
*
* <p> Fixates {@code name} and {@code value} for the duration of encoding.
*
* @param name
* the name
* @param value
* the value
*
* @throws NullPointerException
* if any of the arguments are {@code null}
* @throws IllegalStateException
* if the encoder hasn't fully encoded the previous header, or
* hasn't yet started to encode it
* @see #header(CharSequence, CharSequence, boolean)
*/
public void header(CharSequence name, CharSequence value)
throws IllegalStateException {
header(name, value, false);
}
/**
* Sets up the given header {@code (name, value)} with possibly sensitive
* value.
*
* <p> If the {@code value} is sensitive (think security, secrecy, etc.)
* this encoder will compress it using a special representation
* (see <a href="https://tools.ietf.org/html/rfc7541#section-6.2.3">6.2.3. Literal Header Field Never Indexed</a>).
*
* <p> Fixates {@code name} and {@code value} for the duration of encoding.
*
* @param name
* the name
* @param value
* the value
* @param sensitive
* whether or not the value is sensitive
*
* @throws NullPointerException
* if any of the arguments are {@code null}
* @throws IllegalStateException
* if the encoder hasn't fully encoded the previous header, or
* hasn't yet started to encode it
* @see #header(CharSequence, CharSequence)
* @see DecodingCallback#onDecoded(CharSequence, CharSequence, boolean)
*/
public void header(CharSequence name,
CharSequence value,
boolean sensitive) throws IllegalStateException {
if (logger.isLoggable(NORMAL)) {
logger.log(NORMAL, () -> format("encoding ('%s', '%s'), sensitive: %s",
name, value, sensitive));
}
// Arguably a good balance between complexity of implementation and
// efficiency of encoding
requireNonNull(name, "name");
requireNonNull(value, "value");
HeaderTable t = getHeaderTable();
int index = t.indexOf(name, value);
if (index > 0) {
indexed(index);
} else if (index < 0) {
if (sensitive) {
literalNeverIndexed(-index, value, DEFAULT_HUFFMAN);
} else {
literal(-index, value, DEFAULT_HUFFMAN);
}
} else {
if (sensitive) {
literalNeverIndexed(name, DEFAULT_HUFFMAN, value, DEFAULT_HUFFMAN);
} else {
literal(name, DEFAULT_HUFFMAN, value, DEFAULT_HUFFMAN);
}
}
}
/**
* Sets a maximum capacity of the header table.
*
* <p> The value has to be agreed between decoder and encoder out-of-band,
* e.g. by a protocol that uses HPACK
* (see <a href="https://tools.ietf.org/html/rfc7541#section-4.2">4.2. Maximum Table Size</a>).
*
* <p> May be called any number of times after or before a complete header
* has been encoded.
*
* <p> If the encoder decides to change the actual capacity, an update will
* be encoded before a new encoding operation starts.
*
* @param capacity
* a non-negative integer
*
* @throws IllegalArgumentException
* if capacity is negative
* @throws IllegalStateException
* if the encoder hasn't fully encoded the previous header, or
* hasn't yet started to encode it
*/
public void setMaxCapacity(int capacity) {
if (logger.isLoggable(NORMAL)) {
logger.log(NORMAL, () -> format("setting maximum table size to %s",
capacity));
}
setMaxCapacity0(capacity);
}
private void setMaxCapacity0(int capacity) {
checkEncoding();
if (capacity < 0) {
throw new IllegalArgumentException("capacity >= 0: " + capacity);
}
int calculated = calculateCapacity(capacity);
if (logger.isLoggable(NORMAL)) {
logger.log(NORMAL, () -> format("actual maximum table size will be %s",
calculated));
}
if (calculated < 0 || calculated > capacity) {
throw new IllegalArgumentException(
format("0 <= calculated <= capacity: calculated=%s, capacity=%s",
calculated, capacity));
}
capacityUpdate = true;
// maxCapacity needs to be updated unconditionally, so the encoder
// always has the newest one (in case it decides to update it later
// unsolicitedly)
// Suppose maxCapacity = 4096, and the encoder has decided to use only
// 2048. It later can choose anything else from the region [0, 4096].
maxCapacity = capacity;
lastCapacity = calculated;
minCapacity = Math.min(minCapacity, lastCapacity);
}
/**
* Calculates actual capacity to be used by this encoder in response to
* a request to update maximum table size.
*
* <p> Default implementation does not add anything to the headers table,
* hence this method returns {@code 0}.
*
* <p> It is an error to return a value {@code c}, where {@code c < 0} or
* {@code c > maxCapacity}.
*
* @param maxCapacity
* upper bound
*
* @return actual capacity
*/
protected int calculateCapacity(int maxCapacity) {
return 0;
}
/**
* Encodes the {@linkplain #header(CharSequence, CharSequence) set up}
* header into the given buffer.
*
* <p> The encoder writes as much as possible of the header's binary
* representation into the given buffer, starting at the buffer's position,
* and increments its position to reflect the bytes written. The buffer's
* mark and limit will not be modified.
*
* <p> Once the method has returned {@code true}, the current header is
* deemed encoded. A new header may be set up.
*
* @param headerBlock
* the buffer to encode the header into, may be empty
*
* @return {@code true} if the current header has been fully encoded,
* {@code false} otherwise
*
* @throws NullPointerException
* if the buffer is {@code null}
* @throws ReadOnlyBufferException
* if this buffer is read-only
* @throws IllegalStateException
* if there is no set up header
*/
public final boolean encode(ByteBuffer headerBlock) {
if (!encoding) {
throw new IllegalStateException("A header hasn't been set up");
}
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("writing to %s", headerBlock));
}
if (!prependWithCapacityUpdate(headerBlock)) { // TODO: log
return false;
}
boolean done = writer.write(headerTable, headerBlock);
if (done) {
writer.reset(); // FIXME: WHY?
encoding = false;
}
return done;
}
private boolean prependWithCapacityUpdate(ByteBuffer headerBlock) {
if (capacityUpdate) {
if (!configuredCapacityUpdate) {
List<Integer> sizes = new LinkedList<>();
if (minCapacity < currCapacity) {
sizes.add((int) minCapacity);
if (minCapacity != lastCapacity) {
sizes.add(lastCapacity);
}
} else if (lastCapacity != currCapacity) {
sizes.add(lastCapacity);
}
bulkSizeUpdateWriter.maxHeaderTableSizes(sizes);
configuredCapacityUpdate = true;
}
boolean done = bulkSizeUpdateWriter.write(headerTable, headerBlock);
if (done) {
minCapacity = lastCapacity;
currCapacity = lastCapacity;
bulkSizeUpdateWriter.reset();
capacityUpdate = false;
configuredCapacityUpdate = false;
}
return done;
}
return true;
}
protected final void indexed(int index) throws IndexOutOfBoundsException {
checkEncoding();
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("indexed %s", index));
}
encoding = true;
writer = indexedWriter.index(index);
}
protected final void literal(int index,
CharSequence value,
boolean useHuffman)
throws IndexOutOfBoundsException {
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("literal without indexing ('%s', '%s')",
index, value));
}
checkEncoding();
encoding = true;
writer = literalWriter
.index(index).value(value, useHuffman);
}
protected final void literal(CharSequence name,
boolean nameHuffman,
CharSequence value,
boolean valueHuffman) {
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("literal without indexing ('%s', '%s')",
name, value));
}
checkEncoding();
encoding = true;
writer = literalWriter
.name(name, nameHuffman).value(value, valueHuffman);
}
protected final void literalNeverIndexed(int index,
CharSequence value,
boolean valueHuffman)
throws IndexOutOfBoundsException {
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("literal never indexed ('%s', '%s')",
index, value));
}
checkEncoding();
encoding = true;
writer = literalNeverIndexedWriter
.index(index).value(value, valueHuffman);
}
protected final void literalNeverIndexed(CharSequence name,
boolean nameHuffman,
CharSequence value,
boolean valueHuffman) {
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("literal never indexed ('%s', '%s')",
name, value));
}
checkEncoding();
encoding = true;
writer = literalNeverIndexedWriter
.name(name, nameHuffman).value(value, valueHuffman);
}
protected final void literalWithIndexing(int index,
CharSequence value,
boolean valueHuffman)
throws IndexOutOfBoundsException {
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("literal with incremental indexing ('%s', '%s')",
index, value));
}
checkEncoding();
encoding = true;
writer = literalWithIndexingWriter
.index(index).value(value, valueHuffman);
}
protected final void literalWithIndexing(CharSequence name,
boolean nameHuffman,
CharSequence value,
boolean valueHuffman) {
if (logger.isLoggable(EXTRA)) { // TODO: include huffman info?
logger.log(EXTRA, () -> format("literal with incremental indexing ('%s', '%s')",
name, value));
}
checkEncoding();
encoding = true;
writer = literalWithIndexingWriter
.name(name, nameHuffman).value(value, valueHuffman);
}
protected final void sizeUpdate(int capacity)
throws IllegalArgumentException {
if (logger.isLoggable(EXTRA)) {
logger.log(EXTRA, () -> format("dynamic table size update %s",
capacity));
}
checkEncoding();
// Ensure subclass follows the contract
if (capacity > this.maxCapacity) {
throw new IllegalArgumentException(
format("capacity <= maxCapacity: capacity=%s, maxCapacity=%s",
capacity, maxCapacity));
}
writer = sizeUpdateWriter.maxHeaderTableSize(capacity);
}
protected final int getMaxCapacity() {
return maxCapacity;
}
protected final HeaderTable getHeaderTable() {
return headerTable;
}
protected final void checkEncoding() { // TODO: better name e.g. checkIfEncodingInProgress()
if (encoding) {
throw new IllegalStateException(
"Previous encoding operation hasn't finished yet");
}
}
}