[Oorexx-bugs] [ oorexx-Bugs-2953752 ] REXX.EXE crashes at REXX script end

[SourceForge.net <no...@so...>]

Bugs item #2953752, was opened at 2010-02-17 15:49
Message generated for change (Comment added) made by ltdan66
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=684730&aid=2953752&group_id=119701

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Interpreter
Group: v4.0
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: oorexxpert (oorexxpert)
Assigned to: Nobody/Anonymous (nobody)
Summary: REXX.EXE crashes at REXX script end

Initial Comment:
After my REXX program exits, I get a pop-up that says  that REXX has encountered a problem and needs to close....  The error is in REXX.EXE, ModVer: 4.0.0.5088 Offset 00005bfe.  It does not happen to all scripts, just one of my scripts.  It's probably my error but I have no idea as to how to isolate the cause.  The script does invoke our own C code API to our database.  However, that code has been around for years and hasn't changed in quite a while.  It may be related or not, I can't tell.

----------------------------------------------------------------------

Comment By: LT Dan (ltdan66)
Date: 2010-05-13 11:07

Message:
I have taken over for oorexxpert and I am the one that writes the C code
that gets called in our REXX.  I can confirm that this is failing in 4.0.0
and 4.0.1.  The problem was this line:  
     call value "MONITOR",monitor,"";

where variable monitor was some string.  I know that monitor is reserved
but it should not fail on termination - here is the stack - our issue is
fixed though -- thanks


KERNEL32! 7c812afb()
REXX! 00338a4c()
REXX! 002f01de()






----------------------------------------------------------------------

Comment By: oorexxpert (oorexxpert)
Date: 2010-02-18 18:15

Message:
OK, I will try to re-create the bug without any of our APIs and common
function.  However, I believe that at some point, the error will disappear.
 That knowledge may be helpful or not.

The API interface to REXX is the same as was used on 3.2.  Has there been
any changes to the API that we need to consider.  I am not a C programmer
and don't want to be so I can't be of much help.  If there is anything you
would like me to ask our C programmer, let me know.  I don't think he is
even using C++, just vanila C.

----------------------------------------------------------------------

Comment By: Rick McGuire (bigrixx)
Date: 2010-02-18 18:07

Message:
The location in REXX.EXE is a worthless piece of information.  That
executable is just a stub that calls the RexxStart() API.  An exception
there would likely indicate some sort of memory overlay occurred, but not
give you any information on what might have caused the overlay.  This
should be easy to debug once a sample program that can reproduce the
problem is available. 

----------------------------------------------------------------------

Comment By: oorexxpert (oorexxpert)
Date: 2010-02-18 17:46

Message:
The problem is happening to more and more of my scripts.  Since the
location of the error was identified at offset 5BFE into REXX.EXE, I would
think that someone could tell me what REXX was trying to do at that point
so that I would have a clue as to isolate a test case for the failure.

----------------------------------------------------------------------

Comment By: oorexxpert (oorexxpert)
Date: 2010-02-17 16:43

Message:
The program is not using regular expressions.

I will try to create a testcase that fails.  It will probably take quite a
while.  Is there a later version of REXX I should be using?

----------------------------------------------------------------------

Comment By: Mark Miesfeld (miesfeld)
Date: 2010-02-17 16:32

Message:
This could be:

2857264  ooRexx may crash on exit when RegularExpression class is use  

which is fixed in trunk.

Does the one script use the regular expression class?

----------------------------------------------------------------------

Comment By: Rick McGuire (bigrixx)
Date: 2010-02-17 16:04

Message:
You're going to have to supply us with some way to recreate this if you
want to have a fix.  Just telling us that something crashed does not
constitute sufficient means for debugging this. 

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=684730&aid=2953752&group_id=119701