Rule request: identical operands in comparison operator rule

[marcelofabri]

// not ok
rhs.applyToDictionaries == rhs.applyToDictionaries

// ok
lhs.applyToDictionaries == rhs.applyToDictionaries

See #1368 for more history.

[masters3d]

What do you think of this regex? (\b\S*\b)\s*(==|!=|<|>|<=|>=)\s*\1
escaped: "(\\b\\S*\\b)\\s*(==|!=|<|>|<=|>=)\\s*\\1"

[masters3d]

I think instead of matching anything \S we should probably match only valid operands [A-Za-z0-9_\.]+
That would give us this:
(\b[A-Za-z0-9_\.]+\b)\s*(==|!=|<|>|<=|>=)\s*\1

[masters3d]

@marcelofabri Do you think this rule should be on by default?

[jpsim]

Seems like a reasonable place to start to me! OSSCheck would undoubtedly catch if this produces common false positives.

[masters3d]

@jpsim You had an idea of more rules around this sort of line of thinking? (I think I remember you posted something on twitter with a document that pointed out some of these common issues.)

[masters3d]

https://www.viva64.com/en/b/0509/

[ZevEisenberg]

Just lost a half hour to this exact issue. Came here to post my +1.

[masters3d]

Equal to (a == a)
Not equal to (a != a)
Greater than (a > a)
Less than (a < a)
Greater than or equal to (a >= a)
Less than or equal to (a <= a)

In addition to the basic operators above

we should also include:
Logical AND (a && b) // avoid a && a
Logical OR (a || b) // avoid a || a

surprisingly Logical NOT (!a) can be essentially ignored which is good since this could have been tripped the unwrap operator.

ternary conditional operator:
bool ? a : a // Avoid same value for true and false case.

[realm-probot]

Hey - looks like you forgot to add a T:* label - could you please add one?