RELEASE NOTES: JDK 17

JEP 356: Enhanced Pseudo-Random Number Generators

Provide new interface types and implementations for pseudorandom number generators (PRNGs), including jumpable PRNGs and an additional class of splittable PRNG algorithms (LXM).

For further details, see JEP 356.

Hex Formatting and Parsing Utility

java.util.HexFormat provides conversions to and from hexadecimal for primitive types and byte arrays. The options for delimiter, prefix, suffix, and uppercase or lowercase are provided by factory methods returning HexFormat instances.

URLClassLoader No Longer Throws Undocumented IllegalArgumentException From getResource[s]() and findResource[s]()

In the event that there is a problem getting a resource, URLClassLoader.getResource() and findResource() now return null instead of throwing an undocumented IllegalArgumentException. The same is true of Enumerations obtained from URLClassLoader.getResources() and URLClassLoader.findResources().

This behavior conforms with the long-standing specification. The situation would typically occur on Windows, due to the use of a Windows-style path ("c:/windows").

DatagramSocket Can Be Used Directly to Join Multicast Groups

java.net.DatagramSocket has been updated in this release to add support for joining multicast groups. It now defines joinGroup and leaveGroup methods to join and leave multicast groups. The class level API documentation of java.net.DatagramSocket has been updated to explain how a plain DatagramSocket can be configured and used to join and leave multicast groups.

This change means that the DatagramSocket API can be used for multicast applications without needing to use the legacy java.net.MulticastSocket API. The MulticastSocket API works as before, although most of its methods are deprecated.

More information on the rationale of this change can be seen in the CSR JDK-8260667.

Deprecate the Socket Implementation Factory Mechanism

The following static methods used to set the system-wide socket implementation factories have been deprecated:

• static void ServerSocket.setSocketFactory​(SocketImplFactory fac)
• static void Socket.setSocketImplFactory​(SocketImplFactory fac)
• static void DatagramSocket.setDatagramSocketImplFactory​(DatagramSocketImplFactory fac)

These API points were used to statically configure a system-wide factory for the corresponding socket types in the java.net package. These methods have mostly been obsolete since Java 1.4.

Deprecate JVM TI Heap functions 1.0

The following JVM TI functions have been deprecated in this release:

• IterateOverObjectsReachableFromObject
• IterateOverReachableObjects
• IterateOverHeap
• IterateOverInstancesOfClass

These functions were superseded in JVM TI version 1.2 (Java SE 6) by more powerful and flexible versions. These functions will be changed to return an error in a future release to indicate that they are no longer implemented/supported. The VM flags -Xlog:jvmti=trace and -XX:TraceJVMTI=<function_name> can be used to identify any residual usages of these functions. For example, -Xlog:jvmti=trace -XX:TraceJVMTI=IterateOverHeap is one way to get trace output when IterateOverHeap is used.

Requirements of an Agent's premain Method Changed to Conform to the Specification

The java.lang.instrument implementation has been changed in this release to require that agent premain and agentmain methods are public. The specification has always required this, but it was not enforced. Attempting to run with an agent where these methods are not public will fail with an exception such as:

java.lang.IllegalAccessException: method <fully-qualified-class-name>.premain must be declared public.

A related change in this release is that the premain and agentmain methods must be defined in the agent class. The implementation no longer searches for these methods in superclasses.

Support for CLDR Version 39

Locale data based on Unicode Consortium's CLDR has been upgraded to version 39. For the detailed locale data changes, please refer to the Unicode Consortium's CLDR release notes: - http://cldr.unicode.org/index/downloads/cldr-39

ISO 639 Language Codes for Hebrew/Indonesian/Yiddish

Historically, Java has used old/obsolete ISO 639 language codes for Hebrew/Indonesian/Yiddish languages to maintain compatibility. From Java 17, the default codes are the current codes. For example, "he" is now the language code for "Hebrew" instead of "iw". A new system property has also been introduced to revert to the legacy behavior. If -Djava.locale.useOldISOCodes=true is specified on the command line, it behaves the same way as the prior releases.

Extra '0' in java.util.Formatter for '%012a' Conversion With a Sign Character

In previous releases, formatter conversions with a %a conversion that used the 0 padding flag and a width specifier would produce paddings containing too many zeros if a leading sign or space character was also specified by their respective flags. This has been fixed so that paddings no longer include too many leading zeros.

System Property for Native Character Encoding Name

A new system property native.encoding has been introduced. This system property provides the underlying host environment's character encoding name. For example, typically it has UTF-8 in Linux and macOS platforms, and Cp1252 in Windows (en-US). Refer to the CSR for more detail.

JEP 382: New macOS Rendering Pipeline

The Java 2D API used by the Swing APIs for rendering, can now use the new Apple Metal accelerated rendering API for macOS.

This is currently disabled by default, so rendering still uses OpenGL APIs, which are deprecated by Apple but still available and supported.

To enable Metal, an application should specify its use by setting the system property:

-Dsun.java2d.metal=true

Use of Metal or OpenGL is transparent to applications since this is a difference of internal implementation and has no effect on Java APIs. The metal pipeline requires macOS 10.14.x or later. Attempts to set it on earlier releases will be ignored.

For further details, see JEP 382.

SunPKCS11 Provider Supports ChaCha20-Poly1305 Cipher and ChaCha20 KeyGenerator if Supported by PKCS11 Library

SunPKCS11 provider is enhanced to support the following crypto services and algorithms when the underlying PKCS11 library supports the corresponding PKCS#11 mechanisms:

ChaCha20 KeyGenerator <=> CKM_CHACHA20_KEY_GEN mechanism  
CHACHA20-POLY1305 Cipher <=> CKM_CHACHA20_POLY1305 mechanism  
CHACHA20-POLY1305 AlgorithmParameters <=> CKM_CHACHA20_POLY1305 mechanism  
CHACHA20 SecretKeyFactory <=> CKM_CHACHA20_POLY1305 mechanism

New SunPKCS11 Configuration Properties

SunPKCS11 provider adds new provider configuration attributes to better control native resources usage. The SunPKCS11 provider consumes native resources in order to work with native PKCS11 libraries. To manage and better control the native resources, additional configuration attributes are added to control the frequency of clearing native references as well as whether to destroy the underlying PKCS11 Token after logout.

The 3 new attributes for SunPKCS11 provider configuration file are: 1) destroyTokenAfterLogout (boolean, defaults to false) If set to true, when java.security.AuthProvider.logout() is called upon the SunPKCS11 provider instance, the underlying Token object will be destroyed and resources will be freed. This essentially renders the SunPKCS11 provider instance unusable after logout() calls. Note that a PKCS11 provider with this attribute set to true should not be added to the system provider list since the provider object is not usable after a logout() method call.

2) cleaner.shortInterval (integer, defaults to 2000, in milliseconds) This defines the frequency for clearing native references during busy period (such as, how often should the cleaner thread processes the no-longer-needed native references in the queue to free up native memory). Note that the cleaner thread will switch to the 'longInterval' frequency after 200 failed tries (such as, when no references are found in the queue).

3) cleaner.longInterval (integer, defaults to 60000, in milliseconds) This defines the frequency for checking native reference during non-busy period (such as, how often should the cleaner thread check the queue for native references). Note that the cleaner thread will switch back to the 'shortInterval' value if native PKCS11 references for cleaning are detected.

macOS on ARM Early Access Available

A new macOS is now available for ARM systems. The ARM port should behave similarly to the Intel port. There are no known feature differences. When reporting issues on macOS, please specify if using ARM or x64.

JEP 409: Sealed Classes

Sealed Classes have been added to the Java Language. Sealed classes and interfaces restrict which other classes or interfaces may extend or implement them.

Sealed Classes were proposed by JEP 360 and delivered in JDK 15 as a preview feature. They were proposed again, with refinements, by JEP 397 and delivered in JDK 16 as a preview feature. Now in JDK 17, Sealed Classes are being finalized with no changes from JDK 16.

For further details, see JEP 409.

JEP 406: Pattern Matching for switch (Preview)

Enhance the Java programming language with pattern matching for switch expressions and statements, along with extensions to the language of patterns. Extending pattern matching to switch allows an expression to be tested against a number of patterns, each with a specific action, so that complex data-oriented queries can be expressed concisely and safely.

For further details, see JEP 406.

JEP 306: Restore Always-Strict Floating-Point Semantics

Floating-point operations are now consistently strict, rather than having both "strict" floating-point semantics (strictfp) and subtly different "default" floating-point semantics. This restores the original floating-point semantics of the language and VM, matching the semantics before the introduction of "strict" and "default" floating-point modes in Java SE 1.2.

For further details, see JEP 306.

Removal of sun.misc.Unsafe::defineAnonymousClass

sun.misc.Unsafe::defineAnonymousClass API has been removed in JDK 17. The API replacement is java.lang.invoke.MethodHandles.Lookup::defineHiddenClass and java.lang.invoke.MethodHandles.Lookup::defineHiddenClassWithClassData.

Unified Logging Supports Asynchronous Log Flushing

To avoid undesirable delays in a thread using unified logging, the user can now request that the unified logging system operate in asynchronous mode. This is done by passing the command-line option -Xlog:async. In asynchronous logging mode, log sites enqueue all logging messages to a buffer. A standalone thread is responsible for flushing them to the corresponding outputs. The intermediate buffer is bounded. On buffer exhaustion, the enqueuing message is discarded. The user can control the size of the intermediate buffer by using the command-line option -XX:AsyncLogBufferSize=<bytes>.

XML Implementation Specific Features and Properties

Documentation for Implementation Specific Features and Properties has been added to the java.xml module summary. Along with the existing properties, two new properties are introduced in JDK 17. The following section describes the changes in more detail:

1) Added javadoc for the XML processing limits.

XML processing limits were introduced in JDK 7u45 and JDK 8. They were previously documented in the Java Tutorial Processing Limits section.

The definitions for these limits have been added to the java.xml module summary. See JDK-8261670.

2) Moved the javadoc for JAXP Lookup Mechanism to the java.xml module summary.

The javadoc for JAXP Lookup Mechanism has been moved to the module summary. The original javadoc in JAXP factories are replaced with a link to that section in the module summary.

See JDK-8261673.

3) Added a property to control the newline after the XML header for DOM LSSerializer.

The DOM Load and Save LSSerializer did not have an explicit control for whether or not the XML Declaration ends with a newline. In this release, a JDK implementation specific property, jdk.xml.isStandalone, and its corresponding System property, jdk.xml.isStandalone, have been added to control the addition of a newline and acts independently without having to set the pretty-print property. This property can be used to reverse the incompatible change introduced in Java SE 7 Update 4 with an update of Xalan 2.7.1 in which a newline is omitted after the XML header.

Usage:

   // to set the property, get an instance of LSSerializer
   LSSerializer ser = impl.createLSSerializer();
   // the isStandalone property is effective whether or not pretty-print is set
   ser.getDomConfig().setParameter("format-pretty-print", pretty ? true : false);
   ser.getDomConfig().setParameter("jdk.xml.isStandalone", standalone ? true : false);

   // to use the System property, set it before initializing a LSSerializer
   System.setProperty("jdk.xml.isStandalone", standalone ? “true” : "false");

   // to clear the property, place the line anywhere after the LSSerializer is initialized
   System.clearProperty("jdk.xml.isStandalone");

See JDK-8249867.

4) Added a property to control the newline after the XML header for XSLTC Serializerjava.xml.

The XSLTC Serializer supported a property, http://www.oracle.com/xml/is-standalone, introduced through JDK-7150637, to control whether or not the XML Declaration ends with a newline. It is, however, not compliant with the new specification for Implementation Specific Features and Properties. In order to maintain compatibility, the legacy property is preserved, and a new property, jdk.xml.xsltcIsStandalone, along with its corresponding System property, jdk.xml.xsltcIsStandalone, have been created to perform the same function for the XSLTC Serializer as the isStandalone property for DOMLS LSSerializer. Note that the former has an extra prefix xsltc to avoid conflict with the later in case it is set through the System property.

Usage:

   // to set the property, get an instance of the Transformer
   Transformer transformer = getTransformer(…);
   // the isStandalone property is effective whether or not pretty-print is set
   transformer.setOutputProperty(OutputKeys.INDENT, pretty ? "yes" : "no");
   transformer.setOutputProperty("jdk.xml.xsltcIsStandalone", standalone ? "yes" : "no");

   // to use the System property, set it before initializing a Transformer
   System.setProperty("jdk.xml.xsltcIsStandalone", standalone ? "yes" : "no");

   // to clear the property, place the line anywhere after the Transformer is initialized
   System.clearProperty("jdk.xml.xsltcIsStandalone");

See JDK-8260858.

5) Added existing features and properties and standardizing the prefix to jdk.xml.

Existing features and properties have been added to the Implementation Specific Features and Properties tables in the java.xml module summary. All of the features and properties, existing and new, now have a prefix of jdk.xml as redefined in the Naming Convention section. System properties are searchable in the Java API documentation by the full name, such as jdk.xml.entityExpansionLimit.

See JDK-8265252.

Collections.unmodifiable* Methods Are Idempotent for Their Corresponding Collection

The unmodifiable* methods in java.util.Collections will no longer re-wrap a given collection with an unmodifiable view if that collection has already been wrapped by same method.

TreeMap.computeIfAbsent Mishandles Existing Entries Whose Values Are null

A bug has been fixed in enhancement JDK-8176894 that inadvertently introduced erroneous behavior in the TreeMap.computeIfAbsent method. The erroneous behavior was that, if the map contained an existing mapping whose value was null, the computeIfAbsent method would immediately return null. Other TreeMap methods modified by the enhancement were unaffected. To conform with the specification, computeIfAbsent now calls the mapping function and updates the map with the function's result.

Updated Specifications of KeyStoreSpi.engineStore(KeyStore.LoadStoreParameter) and KeyStore.store(KeyStore.LoadStoreParameter) Methods

The specifications of the KeyStoreSpi.engineStore(KeyStore.LoadStoreParameter param) and KeyStore.store(KeyStore.LoadStoreParameter param) methods have been updated to specify that an UnsupportedOperationException is thrown if the implementation does not support the engineStore() operation. This change adjusts the specification to match the existing behavior.

Removed Telia Company's Sonera Class2 CA Certificate

The following root certificate has been removed from the cacerts truststore: ` + Telia Company + soneraclass2ca DN: CN=Sonera Class2 CA, O=Sonera, C=FI `

Updated keytool to Create AKID From SKID of Issuing Certificate as Specified by RFC 5280

The gencert command of the keytool utility has been updated to create AKID from the SKID of the issuing certificate as specified by RFC 5280.

jarsigner Tool Warns if Weak Algorithms Are Used in Signer’s Certificate Chain

The jarsigner tool has been updated to warn users when weak keys or cryptographic algorithms are used in certificates of the signer’s certificate chain.

JEP 411: Deprecate the Security Manager for Removal

The Security Manager and APIs related to it have been deprecated and will be removed in a future release. To ensure that developers and users are aware that the Security Manager is deprecated for removal, the Java runtime issues a warning at startup if the Security Manager is enabled on the command line via java -Djava.security.manager. The Java runtime also issues a warning at run time if the Security Manager is enabled dynamically via the System::setSecurityManager API. These warnings cannot be disabled.

See JEP 411 for more information and a list of APIs that have been deprecated for removal.

Added 2 HARICA Root CA Certificates

The following root certificates have been added to the cacerts truststore: ``` + HARICA + haricarootca2015 DN: CN=Hellenic Academic and Research Institutions RootCA 2015, O=Hellenic Academic and Research Institutions Cert. Authority, L=Athens, C=GR

• haricaeccrootca2015 DN: CN=Hellenic Academic and Research Institutions ECC RootCA 2015, O=Hellenic Academic and Research Institutions Cert. Authority, L=Athens, C=GR ```

Disable SHA-1 JARs

JARs signed with SHA-1 algorithms are now restricted by default and treated as if they were unsigned. This applies to the algorithms used to digest, sign, and optionally timestamp the JAR. It also applies to the signature and digest algorithms of the certificates in the certificate chain of the code signer and the Timestamp Authority, and any CRLs or OCSP responses that are used to verify if those certificates have been revoked.

In order to reduce the compatibility risk for applications that have been previously timestamped or use private CAs, there are two exceptions to this policy:

• Any JAR signed with SHA-1 algorithms and timestamped prior to January 01, 2019 will not be restricted.
• Any JAR signed with a SHA-1 certificate that does not chain back to a Root CA included by default in the JDK cacerts keystore will not be restricted.

These exceptions may be removed in a future JDK release.

Users can, at their own risk, remove these restrictions by modifying the java.security configuration file (or overriding it using the java.security.properties system property) and removing "SHA1 jdkCA & usage SignedJAR & denyAfter 2019-01-01" from the jdk.certpath.disabledAlgorithms security property and "SHA1 jdkCA & denyAfter 2019-01-01" from the jdk.jar.disabledAlgorithms security property.

Provide Support for Specifying a Signer in Keytool -genkeypair Command

The -signer and -signerkeypass options have been added to the -genkeypair command of the keytool utility. The -signer option specifies the keystore alias of a PrivateKeyEntry for the signer and the -signerkeypass option specifies the password used to protect the signer’s private key. These options allow keytool -genkeypair to sign the certificate by using the signer’s private key. This is especially useful for generating a certificate with a key agreement algorithm as its public key algorithm.

New System Property Added to Enable the OCSP Nonce Extension

A new system property, jdk.security.certpath.ocspNonce, has been added to enable the OCSP Nonce Extension. This system property is disabled by default, and can be enabled by setting it to the value true. If set to true, the JDK implementation of PKIXRevocationChecker includes a nonce extension containing a 16 byte nonce with each OCSP request. See RFC 8954 for more details on the OCSP Nonce Extension.

Remove Vestiges of Intermediate JSR 175 Annotation Format

When annotations were added to the platform in Java SE 5.0, early builds used a different representation of annotations in the class file than the final format. Support for this intermediate format has now been removed. Reading an annotation from a class file using the intermediate format which differs from the final format yields an exception similar to:

java.lang.reflect.GenericSignatureFormatError: Signature Parse error: Expected Field Type Signature

Recompiling the sources or otherwise regenerating the class file to conform to the proper format will resolve the issue.

Add java.time.InstantSource

A new interface java.time.InstantSource has been introduced. This interface is an abstraction from java.time.Clock that only focuses on the current instant and does not refer to the time zone.

JEP 415: Implement Context-Specific Deserialization Filters

JEP 415: Context-Specific Deserialization Filters allows applications to configure context-specific and dynamically-selected deserialization filters via a JVM-wide filter factory that is invoked to select a filter for each individual deserialization operation.

The Java Core Libraries Developers Guide for Serialization Filtering describes use cases and provides examples.

JDK Flight Recorder Event for Deserialization

A new JDK Flight Recorder (JFR) event has been added to monitor deserialization of objects. When JFR is enabled and the JFR configuration includes deserialization events, JFR will emit an event whenever the running program attempts to deserialize an object. The deserialization event is named java/deserialization, and it is disabled by default. The deserialization event contains information that is used by the serialization filter mechanism. Additionally, if a filter is enabled, the JFR event indicates whether the filter accepted or rejected deserialization of the object.

The new Deserialization Event captures:

• Whether a serialization filter is configured or not.
• The serialization filter status, if one is configured.
• The class of the object being deserialized.
• The number of array elements when deserializing an array.
• The current graph depth.
• The current number of object references.
• The current number of bytes in the stream that have been consumed.
• The exception type and message, if thrown by the serialization filter.

Refer to Context-Specific Deserialization Filter and Serialization Filtering Guide for details.

JDK Flight Recorder Event for Deserialization

It is now possible to monitor deserialization of objects using JDK Flight Recorder (JFR). When JFR is enabled and the JFR configuration includes deserialization events, JFR will emit an event whenever the running program attempts to deserialize an object. The deserialization event is named jdk.Deserialization, and it is disabled by default. The deserialization event contains information that is used by the serialization filter mechanism; see the ObjectInputFilter specification. Additionally, if a filter is enabled, the JFR event indicates whether the filter accepted or rejected deserialization of the object. For further information about how to use the JFR deserialization event, see the article Monitoring Deserialization to Improve Application Security. For reference information about using and configuring JFR, see the JFR Runtime Guide and JFR Command Reference sections of the JDK Mission Control documentation.

JDK Flight Recorder Event for Deserialization

It is now possible to monitor deserialization of objects using JDK Flight Recorder (JFR). When JFR is enabled and the JFR configuration includes deserialization events, JFR will emit an event whenever the running program attempts to deserialize an object. The deserialization event is named jdk.Deserialization, and it is disabled by default. The deserialization event contains information that is used by the serialization filter mechanism; see the ObjectInputFilter specification. Additionally, if a filter is enabled, the JFR event indicates whether the filter accepted or rejected deserialization of the object. For further information about how to use the JFR deserialization event, see the article Monitoring Deserialization to Improve Application Security. For reference information about using and configuring JFR, see the JFR Runtime Guide and JFR Command Reference sections of the JDK Mission Control documentation.

JEP 391: macOS/AArch64 Port

macOS 11.0 now supports the AArch64 architecture. This JEP implements support for the macos-aarch64 platform in the JDK. One of the features added is support for the W^X (write xor execute) memory. It is enabled only for macos-aarch64 and can be extended to other platforms at some point. The JDK can be either cross-compiled on an Intel machine or compiled on an Apple M1-based machine.

For further details, see JEP 391.

Deprecate 3DES and RC4 in Kerberos

The des3-hmac-sha1 and rc4-hmac Kerberos encryption types (etypes) are now deprecated and disabled by default. Users can set allow_weak_crypto = true in the krb5.conf configuration file to re-enable them (along with other weak etypes including des-cbc-crc and des-cbc-md5) at their own risk. To disable a subset of the weak etypes, users can list preferred etypes explicitly in any of the default_tkt_enctypes, default_tgs_enctypes, or permitted_enctypes settings.

Use permitted_enctypes if default_tkt_enctypes or default_tgs_enctypes is not present

Use permittedenctypes as the default value of defaulttktenctypes or defaulttgs_enctypes if any of the them are not defined in krb5.conf.

Enable XML Signature Secure Validation Mode by Default

The XML Signature secure validation mode has been enabled by default (previously it was not enabled by default unless running with a security manager). When enabled, validation of XML signatures are subject to stricter checking of algorithms and other constraints as specified by the jdk.xml.dsig.secureValidationPolicy security property.

If necessary, and at their own risk, applications can disable the mode by setting the org.jcp.xml.dsig.secureValidation property to Boolean.FALSE with the DOMValidateContext.setProperty() API.

Disable SHA-1 XML Signatures

XML signatures that use SHA-1 based digest or signature algorithms have been disabled by default. SHA-1 is no longer a recommended algorithm for digital signatures. If necessary, and at their own risk, applications can workaround this policy by modifying the jdk.xml.dsig.secureValidationPolicy security property and re-enabling the SHA-1 algorithms.

JEP 412: Foreign Function & Memory API (Incubator)

Introduce an API by which Java programs can interoperate with code and data outside of the Java runtime. By efficiently invoking foreign functions (i.e., code outside the JVM), and by safely accessing foreign memory (i.e., memory not managed by the JVM), the API enables Java programs to call native libraries and process native data without the brittleness and danger of JNI.

For further details, see JEP 412.

JEP 403: Strongly Encapsulate JDK Internals

Strongly encapsulate all internal elements of the JDK, except for critical internal APIs such as sun.misc.Unsafe.

With this change, the java launcher option --illegal-access is obsolete. If used on the command line it causes a warning message to be issued, and otherwise has no effect. Existing code that must use internal classes, methods, or fields of the JDK can still be made to work by using the --add-opens launcher option, or the Add-Opens JAR-file manifest attribute, to open specific packages.

For further details, please see JEP 403.

JEP 407: Remove RMI Activation

The Remote Method Invocation (RMI) Activation mechanism has been removed. RMI Activation was an obsolete part of RMI that has been optional since Java SE 8. RMI Activation was deprecated for removal by JEP 385 in Java SE 15, and it was removed from this release by JEP 407. The rmid tool has also been removed. See JEP 385 for background, rationale, risks, and alternatives. The rest of RMI remains unchanged.

JEP 410: Remove the Experimental AOT and JIT Compiler

AOT Compiler related code in HotSpot VM has been removed. Using HotSpot VM options defined by JEP295 produce "Unrecognized VM option" error on VM initialization.

For further details, see JEP 410.

Experimental Compiler Blackholes Support

The experimental support for Compiler Blackholes is added. These are useful for low-level benchmarking, to avoid dead-code elimination on the critical paths, without affecting the benchmark performance. Current support is implemented as CompileCommand, accessible as -XX:CompileCommand=blackhole,<method>, with the plan to eventually graduate it to a public API.

JMH is already able to auto-detect and use this facility when instructed/available. Please consult JMH documentation for the next steps.

Modernization of Ideal Graph Visualizer

Ideal Graph Visualizer (IGV), a tool to explore visually and interactively the intermediate representation used in the HotSpot VM C2 just-in-time (JIT) compiler, has been modernized. Enhancements include:

• Support for running IGV on up to JDK 15 (the latest version supported by IGV's underlying NetBeans Platform)
• Faster, Maven-based IGV build system
• Stabilization of block formation, group removal, and node tracking
• More intuitive coloring and node categorization in default filters
• Ranked quick node search with more natural default behavior

The modernized IGV is partially compatible with graphs generated from earlier JDK releases. It supports basic functionality such as graph loading and visualization, but auxiliary functionality such as node clustering and coloring might be affected.

Details about building and running IGV are available in the src/utils/IdealGraphVisualizer/README.md file in the tool's source directory.

New Class Hierarchy Analysis Implementation in the HotSpot JVM

A new Class Hierarchy Analysis implementation is introduced in the HotSpot JVM. It features enhanced handling of abstract and default methods which improves inlining decisions made by the JIT-compilers. The new implementation supersedes the original one and is turned on by default.

To help diagnose possible issues related to the new implementation, the original implementation can be turned on by specifying the -XX:+UnlockDiagnosticVMOptions -XX:-UseVtableBasedCHA command-line flags.

The original implementation may be removed in a future release.

Behavior changes for SSLSocket input stream shut down

The SunJSSE close notification checks for SSLSocket have been made less strict to conform to changes in the Transport Layer Security (TLS) RFCs.

If an application tries to close the input stream of an SSLSocket (via shutdownInput() method) without having received a close notification message from its peer, the SSLSocket will no longer:

1. trigger the transmission of a TLS fatal-level alert to the peer, and
2. invalidate the current TLS session.

The new behavior will still consider this condition an error and will throw a local javax.net.ssl.SSLException. A fatal-level alert will no longer be sent to the peer, and the underlying session will remain valid.

In addition, the internal transport context for the SSLSocket will also now be closed. Previously, this step didn't occur if a fatal message was generated.

SocketExceptions Are Not Wrapped Into SSLExceptions in SSLSocketImpl

This release reverts the behavior of SSLSocketImpl and SSLTransport introduced by JDK-8196584. SocketException will now be thrown as is instead of being suppressed into an SSLException.

Configurable Extensions With System Properties

Two new system properties have been added. The system property, jdk.tls.client.disableExtensions, is used to disable TLS extensions used in the client. The system property, jdk.tls.server.disableExtensions, is used to disable TLS extensions used in the server. If an extension is disabled, it will be neither produced nor processed in the handshake messages.

The property string is a list of comma separated standard TLS extension names, as registered in the IANA documentation (for example, servername, statusrequest, and signaturealgorithmscert). Note that the extension names are case sensitive. Unknown, unsupported, misspelled and duplicated TLS extension name tokens will be ignored.

Please note that the impact of blocking TLS extensions is complicated. For example, a TLS connection may not be able to be established if a mandatory extension is disabled. Please do not disable mandatory extensions, and do not use this feature unless you clearly understand the impact.

JDK Now Accepts RSA Keys in PKCS#1 Format

RSA private and public keys in PKCS#1 format can now be accepted by JDK providers, such as the RSA KeyFactory.impl from the SunRsaSign provider. The RSA private or public key object should have the PKCS#1 format and an encoding matching the ASN.1 syntax for a PKCS#1 RSA private key and public key.

Fix Issues With the KW and KWP Modes of SunJCE Provider

Support for AES/KW/NoPadding, AES/KW/PKCS5Padding and AES/KWP/NoPadding ciphers is added to SunJCE provider since jdk 17. The cipher block size for these transformations should be 8 instead of 16. In addition, for KWP mode, only the default IV, i.e. 0xA65959A6, is allowed to ensure maximum interoperability with other implementations. Other IV values will be rejected with exception during Cipher.init(...) calls.

SunJCE Provider Supports KW and KWP Modes With AES Cipher

The SunJCE provider has been enhanced to support the AES Key Wrap Algorithm (RFC 3394) and the AES Key Wrap with Padding Algorithm (RFC 5649). In earlier releases, the SunJCE provider supported RFC 3394 under the "AESWrap" cipher algorithm that could only be used to wrap and unwrap keys. With this enhancement, two block cipher modes, KW and KWP, have been added that support data encryption/decryption and key wrap/unwrap by using AES. Please check the "SunJCE provider" section of the "JDK Providers Documentation" guide for more details.

Add support for UserDefinedFileAttributeView on macOS

The file system provider implementation on macOS has been updated in this release to support extended attributes. The java.nio.file.attribute.UserDefinedFileAttributeView API can now be used to obtain a view of a file's extended attributes. This (optional) view was not supported in previous JDK releases.

New Implementation of java.nio.channels.Selector on Microsoft Windows

The Windows implementation of the java.nio.channels.Selector API has been replaced in this release to use a new more scalable implementation. No behavior or compatibility issues were observed during testing of the new implementation. The old implementation has not been removed and the JDK can be configured to use the old implementation, if needed, by running with -Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.WindowsSelectorProvider on the command line.

New API for Accessing Large Icons

A new method, javax.swing.filechooser.FileSystemView.getSystemIcon(File, int, int), is available in JDK 17 that enables access to higher quality icons when possible. It is fully implemented for the Windows platform; however, results on other platforms might vary and will be enhanced later. For example, by using the following code: ` FileSystemView fsv = FileSystemView.getFileSystemView(); Icon icon = fsv.getSystemIcon(new File("application.exe"), 64, 64); JLabel label = new JLabel(icon); `

The user can obtain a higher quality icon for the "application.exe" file. This icon is suitable for creating a label that can be better scaled in a HighDPI environment.

JEP 398: Deprecate the Applet API for Removal

JEP 398: Deprecate the Applet API for Removal. It is essentially irrelevant since all web-browser vendors have either removed support for Java browser plug-ins or announced plans to do so.

The Applet API was previously deprecated, though not for removal, by JEP 289 in Java 9.

RegEx Pattern Matching Loses Character Class After Intersection (&&) Operator

This release fixes a buggy behavior in regular expression pattern intersection. In prior releases, if a nested character class were included in some intersections after the intersection (&&) operator, it would be ignored and not included in the generated matcher from the pattern. This change brings the behavior in line with the intersection regex patterns seen in Ruby.

Check for Empty Paragraphs

DocLint (invoked from javac and javadoc with the -Xdoclint option) now checks for constructs that lead to empty paragraphs in the generated documentation, which might be flagged by an HTML validator. The most common cause is the redundant use of <p> at the end of a block of text.

"Related Packages" on a Package Summary Page

The summary page for a package now includes a section listing any "related packages". The set of related packages is determined heuristically on common naming conventions, and may include the following:

• The "parent" package (that is, the package for which a package is a subpackage)
• Sibling packages (that is, other packages with the same parent package)
• Any subpackages

The related packages need not all be in the same module.

Source Details in Error Messages

When JavaDoc reports an issue in an input source file, it displays the source line for the issue, and a line containing a caret (^) pointing to the position on the line, in a manner similar to compiler (javac) diagnostic messages.

In addition, logging and other "info" messages are now written to the standard error stream, leaving the standard output stream to be used for output that is specifically requested by command-line options, such as command-line help.

Ids Used by the Standard Doclet

"Multi-word" ids in the HTML generated by the Standard Doclet have been converted to a uniform style of lowercase words separated by hyphens. This primarily affects the ids used to navigate within the generated documentation and does not affect the ids used for field and method declarations, and which may be used in external pages to reference such declarations within the documentation.

Improved "Help" Page

The content of the "Help" page generated by the Standard Doclet has been revised, improved, and new information added.

• There is a new "Navigation" section that provides general information on how to navigate around the documentation.
• Information about the different kinds of pages has been gathered into a new section, along with new information about pages that were not previously documented.
• There is a brief "table of contents" at the top of the page that provides links to all of the sections and subsections on the page.

In addition, the HELP link in the navigation bar for each kind of page now links directly to the section on the Help page for that kind of page.

Improved Nested Class Summary

When a class or interface has nested classes or interfaces, the list is improved to show the kind of class or interface, such as enum class, record class, annotation interface, as appropriate.

New Page for "New API" and Improved "Deprecated" Page

JavaDoc can now generate a page summarizing the recent changes in an API. The list of recent releases to be included is specified with the --since command-line option. These values are used to find the declarations with matching @since tags to be included on the new page. The --since-label command-line option provides text to use in the heading of the "New API" page.

On the page that summarizes deprecated items, you can view items grouped by the release in which they were deprecated.

Improved Package Summary Pages

The summary page for a package has been restructured to display the different kinds of classes and interfaces in a single tabbed table, instead of a series of separate tables. Additional links have been provided in the navigation bar at the top of the page, to aid in faster navigation to different parts of the page.

`javadoc` output streams

As part of an overall cleanup, the javadoc tool now writes all diagnostic output to the standard error stream, reserving use of the standard output stream for output that is specifically requested, such as command-line help. This is the same as the equivalent behavior in javac.

Legal Headers for Generated Files

The set of files generated by the Standard Doclet typically includes some files with associated licensing requirements. The Standard Doclet now provides support for including the associated legal files, with default behavior for the common case and a new command-line option (--legal-notices) to override that behavior when appropriate.

Improved Output for @see Tags

When a declaration has a series of @see tags, the output is generated in the form of an HTML <ul> list, instead of a simple comma-separated list of links. The style of the list depends on the number and kind of the links.

Console Charset API

java.io.Console has been updated to define a new method that returns the Charset for the console. The returned Charset may be different from the one returned from Charset.defaultCharset() method. For example, it returns IBM437 while Charset.defaultCharset() returns windows-1252 on Windows (en-US). Refer to the CSR for more detail.

Parallel GC Enables Adaptive Parallel Reference Processing by Default

Parallel GC now ergonomically determines the optimal number of threads to use for processing java.lang.ref.Reference instances during garbage collection. The option -XX:ParallelRefProcEnabled is now true (enabled) by default.

The change improves this phase of the garbage collection pause significantly on machines with more than one thread available for garbage collection.

If you experience increased garbage collection pauses, you can revert to the original behavior by specifying -XX:-ParallelRefProcEnabled on the command line.

The ergonomics of java.lang.ref.Reference processing can be tuned by using the experimental option -XX:ReferencesPerThread (default value: 1000).