Forum Discussion

Contributor

11 years ago

testcomplete 10.10 crash with ADO connectiont to oracle

Hi

I am in the progress of investigating upgrade efforts from TC9, to TC10..

Initialy I tested with one of the very first TC10 builds (10.0.xxx something). Which seemed to work ok.. There where a couple of issues that it couldn't find some buttons in our VB .net app that I am running regression tests on. Everything else seemed to work ok.

I then upgraded to TC 10.10.752.7, and it crashed totally on me. It's giving me "Catastrophic failure" now, everytime I have database query against oracle

code is as follows:

var query = "SELECT * FROM bc.order_line o WHERE o.order_line_status != 'PROCESSED' AND o.order_line_status != 'TERMINATED' AND o.create_date < SYSDATE - 10";

var qrySelect = ADO.CreateADOQuery();

qrySelect.ConnectionString = ProjectSuite.Variables.OpusDatabase;

Log.Message("ConnectionString: " + qrySelect.ConnectionString);

Log.Message("Executing query: " + query);

qrySelect.SQL = query;

qrySelect.Open();

the line spewing errors is the last one "qrySelect.Open()". This is working fine in TC9, and in the early TC10 that I tried. But in 10.10 it doesn't work..

Is this a known bug in testcomplete?

/ Thomas

Ryan_Moran
11 years ago
Usually you'll want to fill the flux capacitor with vitamin rich garbage before attempting time travel :). In regards to the Catastrophic failure however I have seen this in low memory situations. Make sure you are properly closing the ADO connection after you're done with it. If you hit the max (I believe 64) connections you'll cause rifts in the space-time continuum.

Ryan_Moran
Valued Contributor
11 years ago
Usually you'll want to fill the flux capacitor with vitamin rich garbage before attempting time travel :). In regards to the Catastrophic failure however I have seen this in low memory situations. Make sure you are properly closing the ADO connection after you're done with it. If you hit the max (I believe 64) connections you'll cause rifts in the space-time continuum.
tbom
Contributor
11 years ago
Just tried to reproduce the problem today, with report generator enabled. And now it works ok.

So it seems that it was a temporary time reversal of the flux capacitors energy flow, in a distant galaxy, that caused the problem.

Ie. it works as expected again.

/ Thomas
TanyaYatskovska
SmartBear Alumni (Retired)
11 years ago

Hi Thomas,

I haven't heard about this issue in TestComplete 10.10. Could you please follow some additional steps to help our team resolve the issue?

1. Run the Report Generator tool located in <TestComplete>\Bin\ReportGenerator.exe.

2. Select the "Launch/attach to TestComplete" option.

3. Reproduce the TestComplete crash.

4. Click the "Generate..." button in the Report Generator tool and save the report to a local folder.

5. Send the report to our TestComplete Support team.

Thanks in advance.
tristaanogre
Esteemed Contributor
11 years ago
EPIC response... :-)
tbom
Contributor
11 years ago
It seems that this is comming back to me like a boomerang.. The very first query I throw at the oracle server, fails with catastrophic failure.

Even after a complete reboot of the machine, where testcomplete resides, results in catastrophic failure when it tries to query the db.

I have checked the oracle server for persistent connections, and there is none. I have tried to use the same credentials in my standard query tool (toad), from another machine, this also works fine. So the user is not locked out of the server.

I even tried to make a small test script, that just connects to the oracle db, queries a table and then closes the connection. on my own machine it works fine. From the testpc it fails.

Seems to me, like there is some configuration, or persistent storrage of something that is kept during a reboot.. But where the *BIIP* is it located?

If anyone has a hint for this, he, or she, will become a part of my evening prayer :)

/ Thomas