Upgrade errorprone to 2.36.0

[risdenk]

Description

Builds on top of #14201 which started with 2.18.0 and helps with the error described here #14200. This adds all the rules to error-prone.gradle and upgrades to the latest error prone plugin version (https://github.com/tbroyer/gradle-errorprone-plugin/releases) and error prone version.

This was checked with

./gradlew check -x test -Pvalidation.errorprone=true -Pjavac.failOnWarnings=false -Pvalidation.git.failOnModified=false

These are ones that potentially have problems but are commented out since the fixes were not trivial

• ArrayRecordComponent
• ClassInitializationDeadlock
• LockOnNonEnclosingClassLiteral

All the rules from the different versions are linked below:

https://github.com/google/error-prone/releases/tag/v2.19.0

• NotJavadoc
• StringCaseLocaleUsage
• UnnecessaryTestMethodPrefix

https://github.com/google/error-prone/releases/tag/v2.20.0

• InlineTrivialConstant
• UnnecessaryStringBuilder
• BanClassLoader
• DereferenceWithNullBranch
• DoNotUseRuleChain
• LockOnNonEnclosingClassLiteral
• MissingRefasterAnnotation
• NamedLikeContextualKeyword
• NonApiType

https://github.com/google/error-prone/releases/tag/v2.21.0

• AttemptedNegativeZero: Prevents accidental use of -0, which is the same as 0. The floating-point negative zero is -0.0.
• ICCProfileGetInstance: Warns on uses of ICC_Profile.getInstance(String), due to JDK-8191622.
• MutableGuiceModule: Fields in Guice modules should be final.
• NullableOptional: Discourages @Nullable-annotated Optionals.
• OverridingMethodInconsistentArgumentNamesChecker: Arguments of overriding method are inconsistent with overridden method.

https://github.com/google/error-prone/releases/tag/v2.22.0

• ClosingStandardOutputStreams: Prevents accidentally closing System.{out,err} with try-with-resources
• TruthContainsExactlyElementsInUsage: containsExactly is preferred over containsExactlyElementsIn when creating new iterables
• UnnecessaryAsync: detects unnecessary use of async primitives in local (and hence single-threaded) scopes
• ReturnAtTheEndOfVoidFunction: detects unnecessary return statements at the end of void functions
• MultimapKeys: Suggests using keySet() instead of iterating over Multimap.keys(), which does not collapse duplicates

https://github.com/google/error-prone/releases/tag/v2.23.0

• DuplicateDateFormatField
• NonFinalStaticField
• StringCharset
• StringFormatWithLiteral
• SuperEqualsIsObjectEquals

https://github.com/google/error-prone/releases/tag/v2.24.0

• MultipleNullnessAnnotations: Discourage multiple nullness annotations
• NullableTypeParameter: Discourage nullness annotations on type parameters
• NullableWildcard: Discourage nullness annotations on wildcards
• SuperCallToObjectMethod: Generalization of SuperEqualsIsObjectEquals, now covers hashCode

https://github.com/google/error-prone/releases/tag/v2.25.0

• JUnitIncompatibleType: Detects incompatible types passed to an assertion, similar to TruthIncompatibleType
• RedundantSetterCall: Detects fields set twice in the same chained expression. Generalization of previous ProtoRedundantSet check to also handle AutoValue.

https://github.com/google/error-prone/releases/tag/v2.26.0

• SystemConsoleNull: Null-checking System.console() is not a reliable way to detect if the console is connected to a terminal.
• EnumOrdinal: Discourage uses of Enum.ordinal()

https://github.com/google/error-prone/releases/tag/v2.27.0

• ClassInitializationDeadlock detects class initializers that reference subtypes of the current class, which can result in deadlocks.
• MockitoDoSetup suggests using when/thenReturn over doReturn/when for additional type safety.
• VoidUsed suggests using a literal null instead of referring to a Void-typed variable.
• TruthSelfEquals has been renamed and generalized as SelfAssertion

https://github.com/google/error-prone/releases/tag/v2.28.0

• DeeplyNested discourages very deeply nested code that can cause StackOverflowErrors during compilation.
• GuiceNestedCombine discourages nesting of Modules.combine(), which is unnecessary.
• PatternMatchingInstanceof migrates code to use pattern matching for instanceof
• SunApi discourages use of internal proprietary JDK APIs which may be removed from future releases.

https://github.com/google/error-prone/releases/tag/v2.29.0

• MissingRuntimeRetention
• SetUnrecognized
• StatementSwitchToExpressionSwitch

https://github.com/google/error-prone/releases/tag/v2.30.0

• AutoValueBoxedValues: Detects unnecessary boxing in AutoValue classes.
• DefaultLocale: Detects implicit use of the JVM default locale, which can result in differing behaviour between JVM executions.
• UnnecessaryBreakInSwitch: Remove unnecessary break statements in -> switches.

https://github.com/google/error-prone/releases/tag/v2.31.0

• AutoValueBoxedValues: AutoValue instances should not usually contain boxed types that are not Nullable. We recommend removing the unnecessary boxing.

https://github.com/google/error-prone/releases/tag/v2.32.0

• no rule changes

https://github.com/google/error-prone/releases/tag/v2.33.0

• ArrayRecordComponent: Record components should not be arrays.

https://github.com/google/error-prone/releases/tag/v2.34.0

• FutureTransformAsync: The usage of transformAsync is not necessary when all the return values of the transformation function are immediate futures.
• MisformattedTestData: Reformats Java source code inside Error Prone test cases.
• MisleadingEscapedSpace: Using \s anywhere except at the end of a line in a text block is potentially misleading.
• ThrowIfUncheckedKnownUnchecked: throwIfUnchecked(knownUnchecked) is equivalent to throw knownUnchecked.

https://github.com/google/error-prone/releases/tag/v2.35.0

• JavaDurationGetSecondsToToSeconds: Prefer duration.toSeconds() over duration.getSeconds()

https://github.com/google/error-prone/releases/tag/v2.36.0

• DuplicateBranches: Discourage conditional expressions and if statements where both branches are the same
• RedundantControlFlow: Reports redundant continue statements.

[risdenk]

I don't understand why this was the same for both...

[risdenk]

This tries to avoid checking all the generated code. The list was taken from the Spotless module and modified into regex form.

[risdenk]

This is where all the /** to /* comes from

[risdenk]

This is where a majority of the changes come from. I do think it makes things a lot clearer, but am happy to revert this change too.

[rmuir]

you've got my curiosity :)

[rmuir]

serious chunk of work, thank you for working through it all!

[dweiss]

Nice. Thank you, @risdenk