Broken VI with no block diagram (case #1052503), Need to contact R&D (as per Applications Engineer)

[emerino]

Hello,

  I was in contact with NI (case #1052503) with a problem with a VI that I created.

 Initially it was not clear if the problem was with my installation of Labview or with the VI itself. After testing the VI in question in several computers, I determined that the problem is with my VI that somehow got corrupted.

In Labview 8.0 it would open the VI and run the program, but it would crash (it would close Labview abruptly) if I tried to show the block diagram (Ctrl-E).

When I tried to open it on Labview v8.0.1, it gave a message saying that the VI was broken and it did not have a block diagram.

The last message that the applications engineer gave me (over the email) was that I needed to report that error to R&D.

How do I report it?

In case someone from R&D check this post, I am including the corrupted VI as an attachment. (Or if anyone else knows a way of repairing it, you are welcome to try!)

Thanks!!

[tst]

Are you sure they said you needed to report it?

The AE should be the one to escalate it, since they have the channels for it. I suggest you talk to them again.

[emerino]

Yup....see the Applications Engineer message below:
" ...I apologize for this error occuring.  Again you are going to need to report
that error to R&D.  From my end, there isn't anything I can see to fix your
VI."

The AE did say that was gonna be in training next week, so maybe I'll have to call and have someone else "escalate it" like you say...

[rolfk]

---

@emerino wrote:
Yup....see the Applications Engineer message below:
" ...I apologize for this error occuring.  Again you are going to need to report
that error to R&D.  From my end, there isn't anything I can see to fix your
VI."

The AE did say that was gonna be in training next week, so maybe I'll have to call and have someone else "escalate it" like you say...

---

He probably meant the R&D of the company or departement that created that VI. You as a user have no way to directly contact LabVIEW R&D and he as NI employee can only hope that you or whoever wrote that VI has a backup that is not yet corrupted. The chance that the AE or LabVIEW R&D can do much in this case are very dim. They may want to see the VI to try to understand how it got corrupted but it might be quite impossible to fix it after the fact.

Rolf Kalbermatter

Message Edited by rolfk on 10-07-2007 12:24 PM

[emerino]

Hmmm....I see....well, I was the one that wrote the VI......and I think I do have a backup from 2 days before it got corrupted.....I'll see what I can do and repost later with my conclusions.....

[Yi Y]

Hi Emerino,

I opened up your VI and tested it out on my computer, and it would crash LabVIEW 8.0 and won't open in 8.2.1 with the same error message you mentioned above. Unfortunately, it does appear that this particular VI is corrupted. Do you remember how the VI got corrupted? If the procedure can be reproduced to corrupt a VI, I will report the issue to National Instruments R&D.  It will probably help R&D eliminating the issue if we can find the source of the issue.
TST and Rolf are correct that customers cannot file a bug report directly to R&D.  Applications Engineers handles the communication with R&D.  There might be some miscommunication there, and I believe the engineer you talked to meant that you might have to talk to your R&D department (in this case, as it turned out, that would be you) and start from the last back up you have that's intact.
It's very important for us to address any issue with National Instruments product appropriately. We apologize for any misunderstanding, and we appreciate your patience with this issue.
Please feel free to post any follow up questions in this thread.

Yi Y.
Applications Engineer
National Instruments
http://www.ni.com/support

[emerino]

Hi,

 Yi, you are correct, the AE made a typo and she will contact R&D about it...


To expand on the events leading to the breaking of the VI....and on my findings....

I was using LV 8.0, installed fresh just under a month ago.

I had been working on this VI for several days, but from Monday thru Wednesday I worked some more on it doing cosmetic improvements and documenting (my bad luck being that it got corrupted just when I had declared that I was finished with it).
I always had the block diagram open and it was not until thursday that at one point labview crashed. Several things happened before crashing, so I don't know if one in particular was the culprit, a combination of factors or none of them.... one of the things that happened being that I "stupidly" put a message box inside a while loop, with no way to exit the loop (doh!)... and since there was no way to stop the program I opened up Task Manager and ended labview. After that I'm pretty sure I restarted Windows (XP Pro), just in case the system became unstable from the abrupt ending of LV.
In any case, the VI opened ok after that event and seemed to run ok. I may have run it several times and stopped normally by clicking on an ON/OFF switch on the front panel...at one point, though, LV crashed on its own when I clicked on the ON/OFF switch. That's when the trouble started....or so I thought!
Since then, I tried opening my VI and it would open the front panel only. If I tried to run the VI it would run perfectly (as always) but would crash LV when I clicked on the ON/OFF button again.
If I tried to show the block diagram (Ctrl-E) it would crash LV.

Everytime I started LV again (after a crash) a system message would say that LV ended abruptly last time on a .cpp process and if I wanted to report the error, but information about that error didn't give any answers.

After contacting support for the first time, I decided to determine if the problem was the VI or my installation of LV, so I tried to open it on another computer that had LV 8.0 and the same thing happened. I also tried opening on a LV 8.0.1 version and in this case it wouldn't wanna open it, saying that the VI was broken and it did not have a block diagram.

At the lab where I work, they make backups everyday, so I though I had lost at least one day of work...it turns out it was more than that! I asked them to restore from the day before the crash (when it had never crashed) and it crashed as well...and LV 8.0.1 said the same thing....I had to go back to a backup done on Monday to get a version that didn't have any errors. I'm still puzzled why the VI didn't crash labview before thrusday, if the VI had been broken since tuesday....

I have since updated my LV to 8.0.1, but on my other remaining computer with 8.0 I did several things today:
I opened up the VI and run it....it crashed LV as soon as I hit the ON/OFF switch
I opened the VI again (after starting LV again of course) and pressed Ctrl-E, it crashed LV
I opened and tried to print the block diagram....same result LV crashed
I tried to "save as"....    LV crashed
I tried to save as a previous version......LV crashed
3 other things that I noticed:
* The grid on the front panel was gone
* I could not edit the front panel (right click didn't work at all), I couldn't add anything to it or move or delete (or even select anything for that matter) It would only show the little hand to change the state of buttons but not the arrow to select...and yes the arrow was selected on the tools.
* On one of the menus there would be an option that said "switch to edit mode" if I hit it it would crash LV

I am now working to (remember and) restore the functions that I had added since the last 'good' backup and will certainly have it all done in another day or so. So it's not like I lost something irreplaceable, mostly annoyed though and intrigued to why it happened, if it will happen again (which is the reason why I just updated to 8.0.1) and most of all if in the end it was just some misplaced character in the header of the VI that could've easily been fixed and had the whole thing repaired....specially since the fact remains that you could open and run the VI, so it definetly is not completly broken.
I have seen at least 2 or 3 other posts in the forum about people with similar problems, so it'd be nice if R&D came up with a VI repair utility, I mean the same way there are compressed .ZIP file repair utilities, or .JPEG or windows registry repair, or .PPT repair tools, etc, etc, etc. Just do a google search for "repair utility" and you'll see that all major software vendors have repair utilities for their file formats.

 I will certainly make several backups during the day now, and will try to pay more attention to when LV behaves strangely to try and give a better reproduceable error next time.

[Yi Y]

Hi emerino,
Thank you very much for the very detailed response.  I will pass the message along and make sure the additional information is added to the bug report.  Meanwhile, if you have not sent the error log file to the engineer that you talked to on the phone earlier, could you please post the error log here? If you are using Windows XP, the error log should be stored at C:\Documents and Settings\<your user name>\My Documents\LabVIEW Data\lvfailurelog. The information presented in the log files are for the most part information to help the developers trace back as to what went wrong when a bug is reported.  That's why they might be a little less informative for users.
VIs should only be corrupted while it's being written to (i.e. saving).  I suspect that the numerous crashes you were experiencing might be linked to the file being corrupted at one point during save, copying or transmitting over the network or sometimes even harddrive failure.
What does the "ON/OFF" switch control? Also, it's possible to make a source destribution VI with the block diagram removed.  I don't think that's your issue, but the "switch to edit mode" option makes me curious.
It's a great idea to have a VI repair utility. I would highly recommend making a product suggestion at www.ni.com/contact, because customer suggestions are reviewed directly by R&D when they consider features to release for future versions.
I hope this information helps, and please feel free to post any follow up questions/information on this thread.
Thank you!

Yi Y.
Applications Engineer
National Instruments
http://www.ni.com/support

[emerino]

I don't have the log files anymore (I deleted them, out of frustration). However, it appears that the VI became corrupt 2 days before LV finally crashed....eventough I restarted LV several times in between, so the log files wouldn't tell you why the VI got corrupted in the first place, just when did LV noticed it.

The ON/OFF switch was just a way of getting out of the main while loop (and therefore finish execution). That was its only function.

If I tried to stop execution by pushing the "abort execution" button on LV it would also crash.

[NoahR]

Hi emerino,

Thanks for all of the information.  We will continue to look into it and reproduce the issue to make sure that R&D knows about this.  We will post back on here to let you know what we have concluded about your question.  Thanks!