[#IDEADEV-7527] Good code is red: Ambiguous method call

IDEA: Development

Good code is red: Ambiguous method call

Created: 24 Jul 06 15:36 Updated: 25 Jul 06 21:45

Component/s:

Editor. Error Highlighting

Fix Version/s:

Demetra 5561

Original Estimate:

Unknown

Remaining Estimate:

Unknown

Time Spent:

Unknown

File Attachments:

None

Image Attachments:

1. ScreenHunter_016.jpg
(9 kb)

Build:	5,557
Fixed in build:	5,562
Severity:	Extreme

Description

« Hide

This is new to build 5557.
Hundreds of our files become red...

class Key<T> {
    Object add(T v) {
        return v;
    }
}

class WKey<W, T> extends Key<T> {

    @Override
    W add(T v) {
        return null;
    }
}

class IBug {

    public <W, T> void addItem(WKey<W, T> key, T v) {
        key.add(v); // --> demetra draw this in red, see attachment
    }
}

All

Comments

Work Log

Change History

Sort Order:

[ Permlink | « Hide ]

Michael Abato - 24 Jul 06 17:58

I was about to submit an item that may be another case of this one. The following code is marked red in 5557, but compiles:

package com.example;

import java.util.*;

public class A {

  public static class Bx<T> {
    public Collection<? extends String> foo(T t) {
      return null;
    }
  }
  public static interface B<T> {
    public Collection<? extends String> foo(T t);
  }
  public static class C implements B<String> {
    public StringList foo(String s) {
      return new StringList();
    }
  }

  public static class StringList extends ArrayList<String> { }

  public static void main(String[] args) {
    new C().foo("sdf");
  }
}

The critical point seems to be that the analyzer doesn't realize StringList satisfies Collection<? extends String> - variations not naming the ArrayList<String> as a separate class are fine.

Not as critical for me - it doesn't penetrate our code. I'll obviously check in the next release, but I'm adding it here so that you could check it yourself.

[ Permlink | « Hide ]

Eugene Vigdorchik - 24 Jul 06 18:00

Yeah, this is another incarnation of the same bug, and it has been fixed.

[ Permlink | « Hide ]

Boaz Nahum - 25 Jul 06 11:13

Sorry, it still unresoved in 5561.

[ Permlink | « Hide ]

Eugene Vigdorchik - 25 Jul 06 12:33

It must be the case that my fix did not get it into 5561. Last minute fix, you know...

[ Permlink | « Hide ]

Boaz Nahum - 25 Jul 06 12:38

We are all developers and we are all human.

[ Permlink | « Hide ]

Sebastian Graca - 25 Jul 06 20:46

There's a typo in the error message: "ambigous"

[ Permlink | « Hide ]

Michael Abato - 25 Jul 06 21:45

I've confirmed my variant persists in 5561, now in several places in my projects. I'll wait until the fix makes it into a release to either confirm or break it out to a separate issue...