Tap fluent chainable

[shaedrich]

Refactor chainable/fluent class methods to use tap() where possible

This doesn't require new tests as the existing tests not failing should prove that there are no side-effects outside of the changed methods

Rules

• Only methods that return $this and one additional line will be reduced to a one-liner (since additional lines could not be transformed into an arrow function. Using normal functions would nullify the line reduction benefit. Only using the last line before the return $this would look inconsistent)
  • These are mostly setters

  public function setA($a)
  {
  - $this->a = $a;
  - 
  - return $this;
  + return tap($this, fn () => $this->a = $a);
  }

• When the last operation before the return $this in a method is a method call on $this, the higher-order tap is used instead
• In two cases, this could be further optimizes by passing a first-class callable as the second argument

[shaedrich]

fyi: The two remaining failed tests failed without a conclusive error message

[crynobone]

The tests are failing due to a segmentation fault. This is not easy to debug but could be due to performance or failure but PHPUnit are no longer able to handle printing the correct output due to the complexity of new code.

Converted to draft since tests are failing.

[devfrey]

Reducing the number of lines does not make code better. I find this more complex to read than before. Changes like these also make it harder for newcomers to contribute. Not to mention the (albeit minor) impact on performance.

[timacdonald]

I love tap. Big tap fan. Huge, even. Maybe the biggest. tap all day, every day.

But tap in a pure setter always feels off to me.

[Jacobs63]

This really adds no value.