This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.

Bug 46682 - Tooltip evaluation of EL breaks debugging
Summary: Tooltip evaluation of EL breaks debugging
Status: CLOSED FIXED
Alias: None
Product: javaee
Classification: Unclassified
Component: Debugger (show other bugs)
Version: 4.x
Hardware: PC Windows XP
: P2 blocker (vote)
Assignee: Martin Grebac
URL:
Keywords: RELNOTE
Depends on:
Blocks:
 
Reported: 2004-07-28 13:31 UTC by L Martinek
Modified: 2006-03-24 10:07 UTC (History)
1 user (show)

See Also:
Issue Type: DEFECT
Exception Reporter:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description L Martinek 2004-07-28 13:31:10 UTC
[Build 200407271830, JDK 1.5.0]

Add following three lines to JSP:

aaa
1 + 2 = ${1 + 2} <br>
bbb

Put a breakpoint to first line and start debugger.
When the breakpoint is reached, move mouse over
${1 + 2} to show tooltip evaluation. Then press
F8. Next line of JSP ins't reached and debugger
session is now broken.
Comment 1 Martin Grebac 2004-07-28 17:06:41 UTC
I was not able to reproduce on 1.4.2_05. Sucessfuly reproduced on
1.5.0-beta2.

Reassigning to debuggerjpda for deeper look. Evaluation of EL
expression ${1+2} means evaluation of a statement (aka hidden watch)
below in the context of a jsp:

pageContext().getExpressionEvaluator().evaluate("${1+2}",
java.lang.String.class, (javax.servlet.jsp.PageContext)pageContext, null)

and from the behaviour it seems like the debugger gets into running
state while evaluating the expression.
Comment 2 L Martinek 2004-07-29 12:08:20 UTC
I can confirm that it's JDK1.5 specific.
Comment 3 Maros Sandor 2004-08-06 15:05:35 UTC
I can reproduce it on 1.5.0-b61. Debugee stops (for unknown reason) 
at org.apache.jasper.runtime.HttpJspBase.service:95 at which point 
debugger resumes it with smart stepping because no sources are 
available. More research is needed (may be JDK build-specific?). No 
problems with JDK 1.4.2.
Comment 4 Jan Jancura 2004-08-31 17:08:40 UTC
Looks like some bug in JDK.
Comment 5 Jan Jancura 2004-10-01 10:14:32 UTC
Reproducible on JDK 1.5.0 FCS.
Looks like to same root cause as in 49576.
Comment 6 Martin Grebac 2004-10-01 10:22:45 UTC
What's the bug number of JDK issue then? This is a serious issue (P2),
I believe you will not be able to waive it by saying 'this is jdk bug'. 
I just noticed that Jarda's issue is a P1!
Comment 7 Jan Jancura 2004-10-13 09:44:54 UTC

*** This issue has been marked as a duplicate of 50315 ***
Comment 8 L Martinek 2004-10-29 10:34:16 UTC
The bug which this bug is duplicate of is RESOLVED-WONTFIX, but I
can't agree with this solution. As user I don't care whether the bug
is in the IDE or JDK. And the JDK bug will not be likely fixed in
JDK1.5.0_01. Is there any possibility to implement some workaround?
Comment 9 Jan Jancura 2004-10-29 10:50:15 UTC
No known workarround on JDK1.5.0. You can use JDK1.4.x.

*** This issue has been marked as a duplicate of 50315 ***
Comment 10 L Martinek 2004-10-29 11:13:17 UTC
Sorry, I forgot to change component. I meant workaround in webapps code.
Comment 11 Martin Grebac 2004-10-29 11:36:54 UTC
 We may disable EL evaluation for 1.5 if that's what you would call a
workaround, but then this will never work in 4.0+jdk1.5 even when the
bug is fixed in later releases of jdk. Let me know what you think.
 No other workaround known. The only thing we can do is release note
this and push jdk team to fix the bug (I adjusted priority of the bug
already).
Comment 12 Jan Jancura 2004-10-29 17:29:03 UTC
I have probably found some workarround, which is not so painful.
The bug occurres during evaluation of expression loading a new
instances of classes.


Checking in src/org/netbeans/modules/debugger/jpda/expr/Evaluator.java;
/cvs/debuggerjpda/src/org/netbeans/modules/debugger/jpda/expr/Evaluator.java,v
 <--  Evaluator.java
new revision: 1.13; previous revision: 1.12
done
Comment 13 Martin Grebac 2004-10-29 18:25:52 UTC
Works fine for me.
Comment 14 L Martinek 2004-11-01 10:43:49 UTC
verified in 200410311900
Comment 15 L Martinek 2004-11-01 10:44:18 UTC
verified