bb.donnay-software.com

Roger,

We have our first client on 2.0. They tested with 10-15 users for 30 days and now all users are using 2.0 (about 100 users).

Contant errors and crashing, which was not occurring with just a few users, or at least not noticed if it was.

This error is in the error log about once an hour, all on different programs. Do not understand what it is.

ERROR LOG of "T:\oiswin.20\OPMENU.EXE" Date: 03/01/2016 09:49:33
Xbase++ version : Xbase++ (R) Version 2.00.554
eXPress++ version : eXPress++ (c) Version 2.0.262
Operating system : Windows Server 2008 R2 06.01 Build 07601 Service Pack 1
Operating system : Windows Server 2008 R2 06.01 Build 07601 Service Pack 1
Database path : Z:\oiswin
Default Directory : Z:\oiswin
Current Directory : Z:\oiswin
User : SHAUNTELCONDLEY
ERROR OBJECT:
oError:args :
-> VALTYPE: O CLASS: xbpTestIcon
oError:description : Internal data structures corrupted
oError:filename :
oError:genCode : 41
oError:operation : :destroy
CALLSTACK:
Called from XBPTESTICON:DESTROY(160)
Called from _ISICON(12211)
Called from DC_GETBITMAP(11947)
Called from DC_XBPGET:CREATE(614)
Called from DC_GETLIST:CREATEOBJECT(11754)
Called from DC_GETLIST:READGUI(2021)
Called from DC_READGUI(113)


This top portion, prior to our program names is always the same. Do not know what xpbtesticon is? The only Icon we use is the desktop icon, if that is relative.
Have you seen this, or do you know what it means?

(I have not looked at the xppfatal log get for the constant crashing, some while not even doing anything)

Fred
Omni

Look at the below code in _DCGETBX.PRG.
It appears that J. Duijf's contribution is causing the problem.

Remove everything except: oIcon := XbpIcon():new():create()

Then run BUILD20.BAT to rebuild DCLIPX.DLL

Ok, did that and will replace it on their server later.

Now the xppfatal. Checked and there is maybe one or two fatal errors a week prior to 2.0. On 2.0 there are 10-15 per day or more.

All on different menu options, random places. some doing nothing but clicking exit.

The only similarity on the error log is this one line, but the thread number and number inside paren is different
(Two different examples within 5 minutes.

Call Stack of Thread 8 (3788):

Call Stack of Thread 3 (1436):

As a note, last week when they were running 1.9 on most users and 2.0 on 10-15 users, all fatal errors were on 2.0. None on 1.9.
Same programs,etc, except for version used to compile and the necessary dll files.

Fred

Have you looked at your XPPFATAL.LOG ?

IDSC's can detabilize an application even if they don't cause an error. Not all IDSC's cause errors. What they do cause is memory corruption that can later lead to a fatal error.

My hope is that your fatals will disappear when the IDSC's go away by removing the icon test.

I got those errors from the falal error log. I was hoping the same thing, that the updated dclipx would resolve both, but did not tell the client to get his hopes up. Its just an inconvenience to his users, since on some occasions it will occur multiple times to the same user, just closes our menu completely.
I will know more by this time tomorrow.

Thanks

roger,

That new dclipx did resolve the normal errors, and about half or more of the fatal errors. They had 8 fatal errors in the last 24 hours.

6 of them had this error.

Call Stack of Thread 4 (1656):
@JD_MAGICHELPTHREAD@I@EXECUTE(95)
(B)@JD_MAGICHELPTHREAD@I@INIT(41)

The call stack was different, as were all the programs, but those two magic... lines were the same. Any thoughts?

Fred

I need to see your XPPFATAL.LOG.

It will have info about which thread caused the error.

Here is their most recent..not sure all of them are in this one, but most are.

Hope the attachment gets there. Last time it uploaded but I could not see it anywhere, so was not sure it got there.

Fred

Put it in a zip file.
Apparently PHPBB doesn't like to upload .LOG files.

Ok, trying now.