…tyle/RedundantParentheses

…e-negative-method-block

Register offense for parentheses around method calls with blocks in `Style/RedundantParentheses`

…nclusion is called with modifier

Fix false-positive for `Layout/EmptyLinesAfterModuleInclusion` when inclusion is called with modifier

This PR fixes false positives for `Lint/UselessAssignment`
when duplicate assignments appear in `if` branch inside a loop
and the variable is used outside `while` loop.

Fixes #14402.

…ess_assignment

[Fix #14402] Fix false positives for `Lint/UselessAssignment`

This PR fixes false negatives for `Layout/EmptyLinesAroundClassBody`
when a class body starts with a blank line and defines a multiline superclass.

Fixes #14411.

…pty_lines_around_class_body

[Fix #14411] Fix false negatives for `Layout/EmptyLinesAroundClassBody`

If you had a formatter that is autoloaded, the LOADED_FEATURES
at the beggining of the execution does not have it and this is used for
the saving of the cache

This causes problems because after the first call when running the
inspection of the files, it gets loaded and the main process reads
the cache that the parallel process created with this new cache key,
causing a invalidation.

We make a memoization to fix the issue so the cache key remains
 consistent throughout the run.

…load-formatters

[Fix #11664] Cache wasn't getting used when using parallelization with custom formatters in certain cases

This PR fixes false positives for `Style/MapToHash`
when using `to_h` with block argument.

Fixes #14416.

This PR fixes false positives for `Style/MapToSet` when using `to_set` with block argument.
This is an issue similar to #14417.

…_to_hash

[Fix #14416] Fix false positives for `Style/MapToHash`

…_to_set

Fix false positives for `Style/MapToSet`

This PR fixes false positives for `Style/SafeNavigation`
when ternary expression with operator method call with method chain.

Fixes #14420.

…ts` with multiline strings that contain only whitespace

The previous implementation was doing a bunch of work to decide if a line was
supposed to be considered or not.
For example, Module.new with method definitions may contain newlines that
should be ignored.

This new implementation only considers the ranges between arguments. The old implementationseemed
to have issues with array-style method calls. To be honest, I did not try to understand the original implementation.

It's not safe to assume that it is always nil when the test starts
since other tests could already have accessed it.

I believe it's also not testing much, except that this method is indeed using a instance variable,
so I opted to remove the test.

Writing something that really tests this seems very difficult.

Followup to #14414