Guenther Posted September 3, 2008 Report Share Posted September 3, 2008 Hi guys, I just discovered a serious bug in the bundle function in LV 8.5.1. Looks like the output cluster doesn't match the input values. The VI should generate dummy data for testing the queued state machine in the main program. When the QSM started to act funny, I found that the generated data are wrong There are 3 nested loops, the output should contain a state for the QSM and the indices of the loops in a variant. The first 6 iterations of the inner loop are ok, but after that the value of the middle index is wrong in almost every other iteration. The boolean is just there to prevent constant folding, so I could actually watch what is going on in the loops. See the screenshot, where the bundle inputs are "1 1 0" while the cluster output reads "1 0 0". This is the first step that is wrong. After that, plenty more. Funny thing, if I wire the output cluster to the case structure or the loop frame, everthing works fine. Could someone try to reproduce this behavior to make sure it is not my copy of LV? In LV 8.6. I could not reproduce this bug, everything works as it should. Good reason to switch to LV 8.6. Cheers, Guenther Download File:post-3733-1220344764.vi[ Quote Link to comment
LAVA 1.0 Content Posted September 3, 2008 Report Share Posted September 3, 2008 I assume that the fact you use unnamed cluster elements is a fact LabVIEW dislikes. It probably is some variant optimization bug. Ton (I get similar results) Quote Link to comment
Guenther Posted September 3, 2008 Author Report Share Posted September 3, 2008 QUOTE (Ton @ Sep 2 2008, 11:17 AM) I assume that the fact you use unnamed cluster elements is a fact LabVIEW dislikes. It probably is some variant optimization bug.Ton (I get similar results) Hi Ton thanks for reproducing this bug. I was starting to worry about my copy of LV. The names don't make the difference. If I feed an unnamed cluster constant into the cluster input, the VI works. But that is just the reason for using this simple bundle function, that you don't have to use a cluster template. Looks more like a cluster initialization issue. Cheers, Guenther Quote Link to comment
vugie Posted September 3, 2008 Report Share Posted September 3, 2008 I didn't get such result in 8.2.1 (although I only reproduced simplified diagram) Are you sure that it is not a probe related issue? Try to index cluster through loops to generate 3d array and check the array. Quote Link to comment
Guenther Posted September 3, 2008 Author Report Share Posted September 3, 2008 QUOTE (vugie @ Sep 2 2008, 12:42 PM) I didn't get such result in 8.2.1 (although I only reproduced simplified diagram)Are you sure that it is not a probe related issue? Try to index cluster through loops to generate 3d array and check the array. It is not probe related, the output array shows the same errors, even if I output a 3D array. That was the original design, when I first noticed this issue. Interesting that it works fine with LV8.2. Cheers, Guenther Quote Link to comment
LAVA 1.0 Content Posted October 3, 2008 Report Share Posted October 3, 2008 Roy (Head of LV R&D) posted in the NI Bug Thread found here the following; " Regarding Ben's entry: Issue with "Bundle" node in LV 8.5This code can be found on LAVA in this thread. I've confirmed the issue discussed in the LAVA thread was previously reported to R&D (CAR # 113018) and subsequently fixed in 8.6. Roy " Ben Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.