Bug: Diagram keeps massively reorganising itself

[Norbert_B]

---

@big5824 wrote:

I'm afraid I'm having trouble getting the Labview scripting package to work, I'll keep trying.

---

In current versions of LV, activation of VI Scripting is simply activating one checkbox in the LV Options (VI Server). If you have an older version of LV, you might require some "outside activation". So again the question: What version of LV are you using?

Norbert

[big5824]

Ah my apologies, I looked everywhere for that setting except VI Server. I've enabled that setting, now how do I access the VI scripting palette? It hasn't appeared on my right click menu.

I'm using LV 2015.

[jcarmody]

There's a thread that discusses a BD issue in LV2015.  I posted a way to reproduce a bug related to auto-growing structures here - http://forums.ni.com/t5/LabVIEW/block-diagram-growing-in-LV-2015/m-p/3259693/highlight/true#M950738.

[big5824]

This looks similar, however the solution posted there doesn't work for me.

[big5824]

I've attached a zip of my project, the file in question here is RT Main, it should be obvious where the problem is occuring.

[RavensFan]

Please attach an actual zip file so that we don't need to download a 3rd party app to open a .7z file.

[altenbach]

---

@big5824 wrote:

I've attached a zip of my project, the file in question here is RT Main, it should be obvious where the problem is occuring.

---

One thing I can tell right away is the presence of a huge number of race conditions due to blatant overuse of local variables. Does the code actually work reliably?

For example in the second frame of the timed sequence you are reading from and writing to the same local variables in random order (e.g. angular rate, acceleration, etc) but chances are that the reading occurs before the writing, getting you stale data. LabVIEW does not operate left to right if there is no data dependency. Eliminate the locals by just wireing between the two locations. It is also much more efficient.

Now I'll look into the cleanup problem....

[Norbert_B]

Hm, i checked your RT Main.vi and it seems that, besides its huge size, that it is not necessarily out of bounds....

However, running the BD cleanup on a selection does increase the BD size, which indicates that the guess about "insert space" by BD cleanup is correct.

What i find surprising is that this happens also if removing the sequence structure and using the BD clean up on the whole BD once. After that, the behavior is still that BD clean up adds room for the updated code which moves the other code attached to terminals.

Doing some additional testing shows that this seems a misbehavior of the BD clean up with timed structures having a sequence in it.

If you do not need the sequence frames in the timed loop, the solution for now is to remove them. NOT using variables would be an easy way for you to be able to skip that sequence frames....

Norbert

[big5824]

Don't worry about the race conditions, I was in the process of completely reorganising my code layout when I decided to make this post, I'm well aware of this.

[altenbach]

---

@Norbert_B wrote:

Select your affected VI and execute this on it. I would expect the BD to be totally out of bounds....

---

Left=24, Top=-140, Right=7340, Bottom=2051