cpovirk

This commit **refactors** the internal logic of the **`NonSerializableMemoizingSupplier`** within the **Guava `Suppliers` utility class**. It **simplifies** the implementation by reverting to using `null` as a sentinel value to indicate a successfully computed result, rather than a dedicated `SUCCESSFULLY_COMPUTED` object. This **maintenance refactoring** removes complex checks and potential reliance on specific lambda-translation behaviors, making the `get()` and `toString()` methods cleaner and potentially more performant. The change improves code clarity and robustness without altering external API behavior for non-serializable memoized suppliers.

This commit **refactors** the **testing infrastructure** for **Guava's `ClassToInstanceMap` implementations** by extracting shared test generators and helper classes into a new dedicated utility class, `ClassToInstanceMapTesting`. This new class centralizes common test logic, improving reusability and maintainability across tests for `ImmutableClassToInstanceMap` and `MutableClassToInstanceMap`. The existing test classes, `ImmutableClassToInstanceMapTest` and `MutableClassToInstanceMapTest`, have been updated to leverage these new utilities. This **maintenance** effort also addresses minor code warnings, enhancing overall test code quality without impacting production code.

This commit performs a **refactoring** effort to standardize argument and null validation across various **Guava core and test libraries**. It **replaces manual validation checks** with calls to `Preconditions.checkArgument` and `Preconditions.checkNotNull` in numerous methods and constructors. Specifically, this affects components within `com.google.common.collect`, `com.google.common.io`, `com.google.common.math`, `com.google.common.reflect`, and `com.google.common.util.concurrent`, as well as several test utilities in `guava-testlib`. This change improves code consistency and readability by leveraging Guava's own utility for precondition checks, with no functional impact on the library's behavior.

This commit performs a **style refactoring** by replacing fully qualified calls to primitive parsing methods (e.g., `Integer.parseInt`, `Double.parseDouble`) with **static imports**. This change primarily affects classes within **Guava's `primitives` package**, such as `Doubles`, `Floats`, `UnsignedBytes`, and `UnsignedInts`, along with several **test files** across `cache`, `collect`, and `net` modules. The goal is to **improve code readability and conciseness** by allowing direct calls to `parseInt`, `parseDouble`, and similar methods. This is a **purely cosmetic change** that has no impact on the functionality or behavior of the affected code.

This commit **refactors** the Guava codebase to enhance readability and conciseness by introducing **static imports** for methods within the **`Funnels`** and **`Hashing`** utility classes. This **maintenance** effort simplifies method calls across various modules, primarily affecting the **`com.google.common.hash`** package, including `BloomFilterStrategies` and hash function implementations, as well as `com.google.common.net.InetAddresses`. The change is extensively applied to **test files** in both the main Guava and Android Guava modules, with the exception of `Funnels.asOutputStream`. As a secondary update within tests, some hash functions were upgraded from MD5 to SHA512, further modernizing the test suite.

This commit primarily **refactors** the **Guava test suite** and related test utilities to leverage Java 8 lambdas and method references. Numerous test methods across the **`com.google.common.base` and `com.google.common.cache` modules** have been updated to use these modern functional constructs for interfaces like `Runnable`, `Callable`, `Predicate`, `Weigher`, and `RemovalListener`. Additionally, utility classes such as `SafeTreeMap` and `SafeTreeSet` within **`guava-testlib`** were refactored to use lambdas for `NATURAL_ORDER` comparators. This **refactoring** effort significantly **improves the conciseness and readability** of the test code by replacing verbose anonymous inner classes with more compact lambda expressions.

This commit **updates the `CONTRIBUTING.md` file** to provide clear instructions on how to run a single test within the **multi-module project**. It adds an example and explains the necessary Maven flags, such as `-DfailIfNoTests=false`, to ensure proper execution without issues like `No tests were executed`. This **documentation enhancement** improves the developer experience by simplifying the process of targeted testing, which is crucial for efficient debugging and contribution.

This commit performs a significant **refactoring** to modernize the codebase by replacing numerous anonymous inner class implementations with **lambda expressions and method references**. Primarily, this impacts a wide range of **Guava test suites** within `guava-testlib` and `guava-tests`, covering modules such as `collect`, `eventbus`, `graph`, `hash`, and `io`, making test setup, assertions, and concurrent operations more concise. Additionally, the `Combiner.run` method in the **`com.google.common.util.concurrent.Futures`** utility class is refactored to use `Executors.callable`. This change is purely a **code style and readability improvement**, with no functional impact on the library's behavior or its public API.

This commit performs a **refactoring** to leverage Java's **autounboxing** feature, replacing explicit calls to `intValue()`, `longValue()`, and `charValue()` with more concise code. This **style improvement** enhances readability and maintainability across various **Guava modules**, including `com.google.common.collect`, `com.google.common.io`, `com.google.common.primitives`, and `com.google.common.util.concurrent`. Specifically, methods like `ImmutableRangeSet.size()`, `BaseEncoding`'s encoding/decoding logic, `Ints.tryParse()`, and `AtomicLongMap`'s update methods are affected. Additionally, the `FreshValueGenerator` in `guava-testlib` and a test in `LongsTest` are updated. The change is purely cosmetic and introduces no functional alterations to the library.

This commit performs a significant **refactoring** across numerous **test classes** within the `guava-testlib` and `guava-tests` modules. It modernizes the test codebase by replacing anonymous inner classes and traditional functional interface implementations with more concise **lambda expressions and method references**. This change primarily affects test setup, assertion logic, and utility methods in packages like `com.google.common.testing`, `com.google.common.math`, `com.google.common.net`, and `com.google.common.reflect`. The goal is to improve the **readability and maintainability of the test suite** by leveraging modern Java features, with no impact on production code functionality.

This commit performs **maintenance** and **refactoring** within the **test utilities** by replacing a custom `joinUninterruptibly` implementation in `InterruptionUtil.java` with the robust production version from `Uninterruptibles`. This ensures consistency and leverages the established utility for handling thread interruptions in tests. Additionally, a comment regarding `TearDown` behavior, specifically `fail()` and `suppressThrows`, is updated to reflect the switch from `TearDownTestCase` to direct `TearDownStack` usage. This change improves the clarity and correctness of the test suite's documentation and utility usage, with no impact on production code.

Lambdafy tests for `util.concurrent`.

This commit **re-implements a workaround** for a specific **Android Marshmallow ART interrupt bug** that could cause class loading to swallow interrupts, leading to flaky test failures in the **`UninterruptiblesTest`** suite. It **refactors the test setup** in `UninterruptiblesTest.java` by adding a new retry mechanism to `runBare`, replacing a previous workaround to ensure test stability on older Android versions. This **maintenance fix** addresses the platform bug and also **suppresses `WaitNotInLoop` warnings** in several test methods. This temporary solution ensures the reliability of concurrency tests until the `minSdkVersion` is updated to Nougat.

This commit **refactors** the `com.google.common.base.FinalizableReferenceQueue` to **eliminate reflection** from its "direct" fallback mechanism for loading the finalizer. It introduces a `FinalizerStarter` interface, enabling direct invocation of the finalizer without reflection in certain scenarios, which significantly improves compatibility with **GraalVM, Proguard, and R8**. While the `Finalizer` class still utilizes reflection for `finalizeReferent`, this change addresses an initial step towards reducing reflection reliance for AOT compilation. Additionally, the **Proguard configuration** (`proguard/base.pro`) is updated with a comment clarifying this non-reflective approach for Android environments.

This commit performs extensive **maintenance** and **refactoring** across various **Guava test suites** to address nullness-checking errors. It involves updating nullability annotations, adjusting assertion styles, modifying array types, and refactoring test methods (e.g., using lambdas) within `com.google.common.base`, `com.google.common.net`, and `com.google.common.util.concurrent` packages for both Android and core builds. Specifically, tests for `Functions`, `Joiner`, `Predicates`, `Splitter`, `Strings`, `ToStringHelper`, `Utf8`, `MediaType`, `PercentEscaper`, `UrlEscaperTesting`, and several `Future` related tests are updated. This work improves the **null-safety** and robustness of the test infrastructure, ensuring better adherence to nullness contracts and compiler checks without altering the library's public API.

This commit performs a **refactoring** across several **Guava collection classes** to improve encapsulation and align with Kotlin's stricter visibility rules. It changes the visibility of internal constructors in `HashBasedTable` and `TreeMultiset`, as well as the `TreeMultiset.distinctElements` method, from package-private to **private**. Additionally, the `MinMaxPriorityQueue.removeAt` method is made private, with a new `removeAtForTesting` method introduced to preserve testability, necessitating updates in `MinMaxPriorityQueueTest`. This **maintenance** work enhances the internal consistency and prevents unintended external access to implementation details within these core data structures.

This commit performs **maintenance** by addressing **`DuplicateAssertion` warnings** reported by Error Prone in the test suite. It adds a `@SuppressWarnings("DuplicateAssertion")` annotation to the `testPresentInstances_callingIteratorTwice` method within `OptionalTest.java` in both the `android/guava-tests` and `guava-tests` modules. This change specifically targets the **`com.google.common.base.Optional`** component's tests, suppressing a linter warning that dates back to an old commit. The fix improves code quality by cleaning up build output without altering any test logic or production code.

Migrate off `assertSame` and `assertNotSame` and onto Truth.

This commit **refactors** the **test suite** for the `html` package by standardizing string assertions across its test files. Specifically, it updates `HtmlEscapersTest.java` in both the core Guava and Android test modules, replacing `assertEquals(expected, actual)` with the more idiomatic `assertThat(actual).isEqualTo(expected)` for `String` comparisons within the `testHtmlEscaper` methods. This **maintenance** change improves the readability and expressiveness of test failures, aligning with preferred assertion styles without altering any production code or test logic. The scope is limited to enhancing the consistency and clarity of existing unit tests.

This commit performs a **test refactoring** within the **Guava `CharsetsTest` suite** for both standard and Android builds. It updates existing `Charset` assertions from the less expressive `assertEquals(expected, actual)` to the more readable and consistent `assertThat(actual).isEqualTo(expected)` style. This change improves the **readability and maintainability of the test code** without altering any production functionality or introducing new features. The scope is limited to internal test consistency, with no impact on downstream consumers.