[GROOVY-5491] HashMap subclass property access inconsistent

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 1.8.6
• Fix Version/s: 5.0.0-alpha-1
• Component/s: groovy-runtime
• Assignee: Eric Milles
• Reporter: Michael Collado
• Created: 04/Jun/12 13:16
• Updated: 23/Aug/23 09:11
• Resolved: 25/Jan/22 18:21

Description

Creating a subclass of java.util.HashMap with properties defined results in inconsistent and confusing behavior.

class MapSub extends HashMap {
  def foo
  
}

// Using dot notation to get the property results in a map key lookup.
def m = new MapSub(foo: 'bar')
println m.foo    // prints 'null'
println m.getFoo() // prints 'bar'

// Using same dot notation to set the property results in property setter.
m.foo = 'baz'
println m.foo // still prints 'null'
println m.getFoo() // prints 'baz'

This behavior results in subtle and hard to track bugs. I saw another Jira that stated that a design decision was made that Map dot property getters always resulted in a map key lookup. For consistency, I think either the same policy should be applied to Map property setters or the original design decision should be reconsidered to first check for property existence.

Issue Links

• is related to: GROOVY-5001 Map access is given higher precedence when trying to access fields/properties in classes which implement java.util.Map or extend java.util.HashMap or java.util.Properties