PR# 11027 Selecting call stack element when stopped in routine called from invariant causes crash

Problem Report Summary
Submitter: prestoat2000
Category: Debugger
Priority: Medium
Date: 2006/08/16
Class: Bug
Severity: Serious
Number: 11027
Release: 5.7.62488
Confidential: No
Status: Analyzed
Responsible: jfiat_es
Environment: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.7.10) Gecko/20050726 Firefox/1.0.6 Solaris 9 on SPARC
Synopsis: Selecting call stack element when stopped in routine called from invariant causes crash

Description
Compile with attached classes and config file.
Put break on {TEST4}.show.  Execute system.
When execution stops at breakpoint, double-click on `make'
at bottom of call stack.  Estudio crashes with attached trace.

To Reproduce

										
Problem Report Interactions
From:jfiat_es    Date:2006/09/18    Status: Analyzed    Download   
Dear David,

This issue is partly fixed to be honest, we modified the debugger to prevent EiffelStudio to crash, however the real issue still exists and we are working on it.

For now we keep this issue as "Analyzed" for our need.

From:prestoat2000    Date:2006/09/17    Status: Closed    Download   
Although I didn't hear anything further, this seems to be fixed in 5.7.63625.
Perhaps some other change fixed it.  Closing this report.

From:jfiat_es    Date:2006/08/28    Download   
Indeed, I've been able to reproduce it, 
this seems to be due to the use of precompilation and the invariant..
I will tell you when it is fixed.

From:prestoat2000    Date:2006/08/27    Download   
The problem is still reproducible (every time) with 5.7.63002 on Solaris 10 on x86.

From:prestoat2000    Date:2006/08/17    Download   
Yes, the crash happens every time (at least on Solaris 10 on x86),  Can't
remember if it happens on Solaris 9 on SPARC, but I suspect it does.  If the
memory corruption fix doesn't fix it, then we can pursue it further.

By the way, I won't be back until Saturday August 26 so these are my last
few responses before my vacation starts.

From:jfiat_es    Date:2006/08/17    Status: Analyzed    Download   
We were unable to reproduce this issue, and according to the stack trace, it seems the cause could be a memory corruption. Next release contains fixes for memory corruption issues. Then it may be fixed by next release.

However I am wondering if you are able to reproduce this crash easily ?

From:prestoat2000    Date:2006/08/16    Download   
Attachments for problem report #11027

Attachment: test.ecf     Size:1292
Attachment: test.e     Size:95
Attachment: test4.e     Size:189
Attachment: stack17.txt     Size:2681