Add a documentation for simplifiable if expression

[Pierre-Sassoulas]

Type of Changes

	Type
✓	🔨 Refactoring
✓	📜 Docs

Description

I realized during implementation that simplifiable if expression became worst and that I disagree with #1984 basically. If there are two values but one if the opposite of the other we can pretty much simplify the if expression. It look like the example from the issue come directly from this stackoverflow question but the top answer make a compelling point that this is simplifiable.

Refs #5953
Closes #5882

[coveralls]

Pull Request Test Coverage Report for Build 2403335689

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 95.51%

---

Totals
Change from base Build 2403218426:	0.0%
Covered Lines:	16230
Relevant Lines:	16993

---

💛 - Coveralls

[DanielNoord]

We need to expect a message here.

[Pierre-Sassoulas]

Yes, it's not raising anything right now, so I did not expect it to work I think it's a problem and that's why this issue is labelled discussion:

I realized during implementation that simplifiable if expression became worst and that I disagree with #1984 basically. If there are two values but one if the opposite of the other we can pretty much simplify the if expression. It look like the example from the issue come directly from this stackoverflow question but the top answer make a compelling point that this is simplifiable.

[DanielNoord]

Ah okay.

I'm not sure whether I think we should raise here. I'll let others give their opinion as well.

[jacobtylerwalls]

I think #1984 was correct, because even if the two values are given opposing boolean values, we can't simplify, because they might have had state before.

E.g. this raises TypeError because b had state:

def x():
  a = []
  b = []
  if bool('special test'):
    a = True
  else:
    b = False
  return a + b

But if we refactor to return bool('special test'), then the behavior is different.

[Pierre-Sassoulas]

You're right, thank you for your input 👍 I think this is still simplifiable because clearly the code style is horrible (with type redefinition of a and b and the conditional can still be simplified : clearly we don't have this type of construct in the primer 😄) but it's another MR's scope.

[github-actions]

🤖 According to the primer, this change has no effect on the checked open source code. 🤖🎉