Bug 26912

Summary:	'null == null' fooling blank final analysis
Product:	[Eclipse Project] JDT	Reporter:	Philipe Mulet <philippe_mulet>
Component:	Core	Assignee:	Philipe Mulet <philippe_mulet>
Status:	VERIFIED FIXED	QA Contact:
Severity:	normal
Priority:	P3
Version:	2.1
Target Milestone:	2.1 M5
Hardware:	PC
OS:	Windows 98
Whiteboard:

Description Philipe Mulet

2002-11-21 17:48:41 EST

Build 2.1-M3

The following code should be rejected:

public class X {
	final int blank;
	{ 	if (null == null) { 	
			blank = 1;	
		}	
	}	
	X(){	// blank may not be initialized (since null==null cannot be 
inlined)
	}	
}

Comment 1 Philipe Mulet

2002-11-22 17:04:40 EST

Fixed along with changes to flow analysis.

Comment 2 David Audel

2002-12-17 11:35:35 EST

Bug still occur in 2.1 M4 candidate

reopen and change target milestone to M5

Comment 3 Philipe Mulet

2002-12-17 14:45:25 EST

Seems like we still inline null==null where we shouldn't.
Need to investigate. Not critical

Comment 4 Philipe Mulet

2003-01-29 05:28:33 EST

Fixed. Was positioning an optimizedConstant were it shouldn't.

Comment 5 David Audel

2003-02-07 09:06:49 EST

Verified.