Bugzilla – Full Text Bug Listing
| Summary: | [1.7][compiler] Incorrect name clash error with ecj | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] JDT | Reporter: | Ayushman Jain <amj87.iitr> | ||||
| Component: | Core | Assignee: | Srikanth Sankaran <srikanth_sankaran> | ||||
| Status: | VERIFIED FIXED | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | P3 | CC: | jarthana, satyam.kandula | ||||
| Version: | 3.7 | Flags: | satyam.kandula:
review+
|
||||
| Target Milestone: | 3.7.1 | ||||||
| Hardware: | All | ||||||
| OS: | All | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
This was fixed in javac 7 via oracle bug http://bugs.sun.com/view_bug.do?bug_id=7020657 Another example, may be related :
public abstract class Bug extends SuperBug
{
abstract <T extends Bug> T clone () throws CloneNotSupportedException;
}
class SuperBug
{
@Override
public SuperBug clone () throws CloneNotSupportedException
{
return (SuperBug) super.clone ();
}
}
This also compiles fine with javac7b147
(In reply to comment #2) > Another example, may be related : I think this problem is unrelated - as this is not a regression with respect to 3.6. The comment# 0 test case is a regression introduced by the fix for bug 334306 (In reply to comment #3) > The comment# 0 test case is a regression introduced by the fix for > bug 334306 The fix for bug 334306 brought this out, but looks like the problem got introduced at the time of the fix for bug 175987. Why is the code if (one.declaringClass.isInterface()) { for (int i = 0; i < length; i++) if (!areTypesEqual(oneArgs[i], twoArgs[i])) return false; } else { even there in org.eclipse.jdt.internal.compiler.lookup.MethodVerifier15.areParametersEqual(MethodBinding, MethodBinding) ??? (In reply to comment #2) > Another example, may be related : I think this test case is the same as the one in bug 345947 and so this piece will be worked on as part of that fix. I'll add a comment to that bug carrying over this test case. This bug itself is not a regression with respect to 3.6. Created attachment 201058 [details]
Proposed patch
To make the review easier, I have preserved the original indentation.
Before releasing I'll reindent the code.
Satyam, please review, TIA. Released in HEAD and 3.7 maintenance branch. There is a difference between ecj and javac7 when the line ' int get(List l); ' gets commented out. I have filed bug 345947 to track it. This change in behavior is not caused by this bug fix. Otherwise, the changes look good. Hence +1 for the changes. (In reply to comment #9) > There is a difference between ecj and javac7 when the line ' int get(List l); ' > gets commented out. I have filed bug 345947 to track it. This change in > behavior is not caused by this bug fix. > > Otherwise, the changes look good. Hence +1 for the changes. I actually meant bug 354229 and not bug 345947. Verified for 3.7.1 RC2 with build M20110824-0800. Verified for 3.8M2 using build I20110911-2000. |
BETA_JAVA7 import java.util.*; interface A { int get(List<String> l); } interface B { int get(List<Integer> l); } interface C extends A, B { int get(List l); // name clash error here } This code compiles fine with javac 7b147, not with ecj.