Conserve memory option causes extra lines in report

[david_jenkinson]

Hello,
We have a stresstest sequence that runs 72 hours.  We need to check the "conserve memory" box while reporting on-the-fly.  The problem is that when we do, we get a lot of extra lines that are not normally in the html report when running with on-the-fly on or off (without conserve memory checked).

I have attached 3 reports, one with on the fly disabled, one with on the fly enabled, one with on the fly enabled and conserve memory enabled.  In the last one, you will see that there are extra lines that reflect the sequences/subsequences that are being launched, but we do not want that information there.  We are setting the report filter expression programatically in the setup of our sequence, such that anything other than "passed", "done", or "skipped" is output to the report.  Can anyone tell me how or why this is happening?  We are running a modified version of the Single Pass entry point sequence, in which we have added infomration for the header of the report, but haven't really touched it otherwise.  We are running Teststand 4.0.

Thanks for any help
David Jenkinson

[Anand_Jain]

Hi

Can you please post a sample sequence file which reproduces this issue. Also post your report options file.

Thank you
Anand Jain
National Instruments

Message Edited by Anand Jain on 07-16-2007 09:31 PM

[david_jenkinson]

OK,
Here is the sequence we are running.
Of course, all the modules that are called by the sequence are not included.  Attached also is the TeststandModelReportOptions file. 

Thanks
Dave J

[Mikef]

I've been watching this thread for the last few days, and hoping to see some sort of a solution.  We have a similar problem that we had to solve by modifying the TSOTFRG.dll code.

Is there another solution?

Mike

[James M]

Hi David,

This was reported to R&D (# CAR 3MFCAB3S) for further investigation and is slated to be fixed in a future release of TestStand.

A possible workaround would be to disable reporting on the sequence via Properties>Run Options>Record Results. As another option (although it is not recommended), you could also attempt to modify the source code for the on-the-fly server located in:
<TestStand>\Components\NI\Models\TestStandModels\TSOTFRG

Thanks for the feedback!

Regards,

[david_jenkinson]

So this has really begun to cause us problems.  We are attempting to run for 72 hours.  If we do not check "conserve memory", we run out of memory because the resultlist maintained by teststand does not get purged.  If we check "conserve memory", the above mentioned bug causes there to be HUGE html results, which are sometimes very diffucult to impossible to navigate, because they are so large. 

As I mentioned before, NOT checking conserve memory, the html results look fine.  We have it filtered to only report failures.  So we are in a catch 22 situation.

You mentioned "you could also attempt to modify the source code for the on-the-fly server".  Well, we have spent many thousands of dollars on this software and licenses, we feel NI should be fixing this issue, not us.  Where is this bug in it's status, and is there a forthcoming patch? 

[hansf]

Hi James, et.al.,

We need a fix for this before the next version of TestStand, especially since we don't know when the next version will be released.  Disabling reporting on the sequence files is not an acceptable work around because we need the sequence steps in the report when they fail.  That leaves modifying the source code for the on-the-fly server as the only viable option left on the table. 

Since user modification of the source code is not recommended per NI, and since many users need the fix, it makes much more sense for NI to modify the source code and make the dll available to all users as a patch than for each user to individually make their own fix.

We are paying for support for multiple licenses, and in return we expect correctable problems such as this to be resolved in a reasonable manner.  Please let us know if NI can provide a patch.

Thanks,

Hans

[AllenP]

Currently we are investigating the issue to try to develop a patch for this issue.  I would like to clarify that James incorrectly stated that we do not recommend you modify the source code provided.  The source code is provided specifically so that customers can modify the code to change the behavior.  We are currently investigating a patch for TestStand 4.0.  However, if anyone else is expiriencing this problem in other versions of TestStand, please post in this thread with the version of TestStand you are using.  I will let you know as soon as there are any updates.

Allen P.
TestStand R&D

[Josh_W]

The behavior discussed above where extra lines appear in the report when the conserve memory option is checked was reported to R&D (# CAR 3MFCAB3S) for further investigation and has been fixed.  For users who are using TestStand 4.0.1 who require this fix, there is a patch available. 

Please contact NI directly (ni.com/contact then choose Request Support) to request the patch.