Bug 329589

Summary:	[1.4/1.5][compiler] Class comparison using '==' has issue with java.lang.Class and the 1.4/1.5 mixed mode
Product:	[Eclipse Project] JDT	Reporter:	Olivier Thomann <Olivier_Thomann>
Component:	Core	Assignee:	Srikanth Sankaran <srikanth_sankaran>
Status:	VERIFIED DUPLICATE	QA Contact:
Severity:	normal
Priority:	P3	CC:	tjwatson
Version:	3.7
Target Milestone:	3.7 M4
Hardware:	PC
OS:	Windows 7
Whiteboard:

Description Olivier Thomann

2010-11-05 16:25:22 EDT

Using HEAD code, some code that is comparing Class objects using '==' has some issues when 1.4 and 1.5 types are mixed up.
Regression test to follow.

Comment 1 Olivier Thomann

2010-11-05 16:26:44 EDT

Regression test in:
org.eclipse.jdt.core.tests.compiler.regression.MethodVerifyTest._test329589()

Comment 2 Olivier Thomann

2010-11-06 12:10:41 EDT

The regression test still fails for this one with the fixes for the other issues, but there is no longer any error in the OSGi workspace.

Tom, could you please try JDT/Core from HEAD with your own workspace and confirm that everything compiles fine ?

Thanks.

Comment 3 Srikanth Sankaran

2010-11-06 13:56:04 EDT

(In reply to comment #2)
> The regression test still fails for this one with the fixes for the other
> issues, but there is no longer any error in the OSGi workspace.

This regression test is "unreal" and cannot be a correct model
of the user's setup, since in this test case, the class 'Class'
will appear both as BinaryTypeBinding (1.5 type) and SourceTypeBinding
(1.4 type) to the compiler. There can be no equivalence between
two such types.

I think the appropriate test case should have Class as a generic binary
type from 1.5 jar and again as a non generic 1.4 binary type.
I suspect the underlying problem is the same as the one in bug 329593
to do with type equivalence in the presence of plain, raw and parameterized
versions of a type.

> Tom, could you please try JDT/Core from HEAD with your own workspace and
> confirm that everything compiles fine ?

Olivier, Tom, Appreciate all the help with triaging and testing.
Thanks also for your patience, this area has been the veritable hornet's
nest that has been stirred up.

Comment 4 Srikanth Sankaran

2010-11-06 22:32:26 EDT

I deleted the junit as it is not the right test case.

Closing as duplicate of bug 329593.

*** This bug has been marked as a duplicate of bug 329593 ***

Comment 5 Thomas Watson

2010-11-08 08:57:35 EST

(In reply to comment #2)
> Tom, could you please try JDT/Core from HEAD with your own workspace and
> confirm that everything compiles fine ?

My workspace is compiling fine using JDT/Core from HEAD.  Thanks for the quick turnaround!

Comment 6 Olivier Thomann

2010-12-07 12:14:24 EST

Verified using I20101207-0250 (4.1 I-build)