groovy

String(Buffer|Builder).metaClass.toString ignored

Details

  • Type: Bug Bug
  • Status: Closed Closed
  • Priority: Minor Minor
  • Resolution: Duplicate
  • Affects Version/s: 1.7.2
  • Fix Version/s: None
  • Component/s: groovy-jdk
  • Labels:
    None
  • Number of attachments :
    0

Description

I assume this is a hard fix, as it is mentioned in GROOVY-2599, but the following code passes all the asserts. I would have expected them all to fail?

String.metaClass.toString = {-> "silly"}
StringBuffer.metaClass.toString = {-> "silly"}
StringBuilder.metaClass.toString = {-> "silly"}

assert "hello world".toString() == "hello world"
assert new StringBuffer("hello world").toString() == "hello world"
assert new StringBuilder("hello world").toString() == "hello world"

I have added it as a new issue as I don't believe GROOVY-2599 actually covers this problem, it just crops up in the comments

Issue Links

duplicates

Bug - A problem which impairs or prevents the functions of the product. GROOVY-3493 Cannot override methods via metaclass that are part of an interface implementation

  • Critical - Crashes, loss of data, severe memory leak.
  • Open - The issue is open and ready for the assignee to start work on it.

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Roshan Dawrani added a comment - - edited

The issue you are facing here is not related to toString() per se. It is about a more generic scenario where you override an interface method on the implementing class.

So, the following works because foo() is defined on a class and you override on the same class:

class Test {
    def foo() {
        "original foo"
    }
}

Test.metaClass.foo = {"new foo"}

assert new Test().foo() == "new foo"

But the following does not work - because now the method you override using class Foo is defined on the interface ITest.

interface ITest {
    def foo()
}

class Test implements ITest {
    def foo() {
        "original foo"
    }
}

Test.metaClass.foo = {"new foo"}

assert new Test().foo() == "new foo"

The toString() parallel to the above is that toString() in case of String/StringBuffer/StringBuilder is defined by java.lang.CharSequence - and that is why overriding them using String/StringBuffer/StringBuilder metaclasses has no effect (same as example above)

If you change your example to the following, let's say, then all usages of toString() work as expected:

CharSequence.metaClass.toString = {-> "silly"}

println "hello world".toString()
println new StringBuffer("hello world").toString()
println new StringBuilder("hello world").toString()
Show
Roshan Dawrani added a comment - - edited The issue you are facing here is not related to toString() per se. It is about a more generic scenario where you override an interface method on the implementing class. So, the following works because foo() is defined on a class and you override on the same class: class Test { def foo() { "original foo" } } Test.metaClass.foo = { " new foo" } assert new Test().foo() == " new foo" But the following does not work - because now the method you override using class Foo is defined on the interface ITest. interface ITest { def foo() } class Test implements ITest { def foo() { "original foo" } } Test.metaClass.foo = { " new foo" } assert new Test().foo() == " new foo" The toString() parallel to the above is that toString() in case of String/StringBuffer/StringBuilder is defined by java.lang.CharSequence - and that is why overriding them using String/StringBuffer/StringBuilder metaclasses has no effect (same as example above) If you change your example to the following, let's say, then all usages of toString() work as expected: CharSequence.metaClass.toString = {-> "silly" } println "hello world" .toString() println new StringBuffer ( "hello world" ).toString() println new StringBuilder( "hello world" ).toString()
Hide
Permalink
Roshan Dawrani added a comment -

God knows why someone had to define "public String toString()" on java.lang.CharSequence !

I see the name of the JDK implementer of that interface but no mail id

Show
Roshan Dawrani added a comment - God knows why someone had to define "public String toString()" on java.lang.CharSequence ! I see the name of the JDK implementer of that interface but no mail id
Hide
Permalink
blackdrag blackdrag added a comment -

still EMC should handle that case

Show
blackdrag blackdrag added a comment - still EMC should handle that case
Hide
Permalink
Roshan Dawrani added a comment -

Oh, I was just separating the issue from the toString() context. I think I have seen other issues about interface/class/EMC situation and I thought it might be a known issue.

Show
Roshan Dawrani added a comment - Oh, I was just separating the issue from the toString() context. I think I have seen other issues about interface/class/EMC situation and I thought it might be a known issue.
Hide
Permalink
Roshan Dawrani added a comment -

Just checked and stripped of toString() context, it looks to be an exact duplicate of GROOVY-3493. Isn't it?

Show
Roshan Dawrani added a comment - Just checked and stripped of toString() context, it looks to be an exact duplicate of GROOVY-3493 . Isn't it?

People

  • Assignee:
    Roshan Dawrani
    Reporter:
    Tim Yates
Vote (0)
Watch (0)

Dates

  • Created:
    Updated:
    Resolved: