Bug 5618 - Uncaught CompletionNodeFound exception doing code assist
Summary: Uncaught CompletionNodeFound exception doing code assist
Status: RESOLVED FIXED
Alias: None
Product: JDT
Classification: Eclipse Project
Component: Core (show other bugs)
Version: 2.0   Edit
Hardware: PC Windows 2000
: P2 normal (vote)
Target Milestone: 2.0 M1   Edit
Assignee: David Audel CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2001-11-07 11:02 EST by Philipe Mulet CLA
Modified: 2002-01-11 09:22 EST (History)
0 users

See Also:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Philipe Mulet CLA 2001-11-07 11:02:11 EST 
Build 207 + JDTCORE from HEAD content (as of 20011107).

1 org.eclipse.core.resources 4 Unhandled exception caught in event loop.
Log: Wed Nov 07 17:00:46 CET 2001
4 org.eclipse.ui 0 Failed to execute runnable 
(org.eclipse.jdt.internal.codeassist.complete.CompletionNodeFound)

Was trying to perform a method declaration completion inside an anonymous
class:

Inside org.eclipse.jdt.internal.core.CompilationUnit
...
protected void 
codeComplete(org.eclipse.jdt.internal.compiler.env.ICompilationUnit cu, 
org.eclipse.jdt.internal.compiler.env.ICompilationUnit unitToSkip, int position, 
final ICodeCompletionRequestor requestor) throws JavaModelException {
	codeComplete(
		cu,
		unitToSkip,
		position,
		new ICompletionRequestor(){
			<CODEASSIST HERE>
		}
}
Comment 1 Philipe Mulet CLA 2001-11-07 11:05:30 EST 
Actually the declaring class name was 'Openable'
Comment 2 Philipe Mulet CLA 2001-11-07 11:30:49 EST 
Was able to reproduce on a simpler example:

public class X {

        void foo(){
                
                System.out.println(
                        new Runnable(){
                                <CODE ASSIST HERE>
                        }
                );
        }

}

Then select method 'run()' and insert it, the walkback occurs right after it in 
the UI (suspecting it triggers automatically another code assist at this point, 
which is the one crashing us).
Comment 3 Philipe Mulet CLA 2001-11-14 10:24:02 EST 
Protection was added, but real defect should still be fixed
Comment 4 David Audel CLA 2001-11-26 11:16:08 EST 
Fixed
Comment 5 Philipe Mulet CLA 2001-11-26 12:47:04 EST 
I do not buy the fix, and re-released the original (still wrong) behavior.

The introduced flag isn't reinitialized when consumed and when doing recovery. 
The nestedMethod() action should be investigated more if offending.
Comment 6 David Audel CLA 2001-11-27 06:53:29 EST 
push BlockInvocationPtr twice for method and constructor (and pop twice)