03-22-2017 09:20 AM - edited 03-22-2017 09:31 AM
I was working with an array of double going into a cluster with a few different data types. I inserted "to single" on the top wire. All of my coersion dots disappeared and it caught me by surprise.
Note that this is strictly a visual bug... and saving the VI brings the coersion dots back - which is why I posted a screenshot instead of a snip. Still thought it worth mentioning.
Edit: Removed useless information...
03-22-2017 10:00 AM
Interestingly enough, if I do a BD cleanup, LabVIEW is coerced into showing the dots again. (Pun definitely coerced and intended.)
03-22-2017 10:23 AM
Which version of LabVIEW is this?
Can you share the VI that does this?
What if you don't do a BD cleanup, but just scroll the window so that that part of the code goes off screen and comes back again.
(Wondering if it is some sort of redraw bug and forcing a redraw fixes it.)
03-22-2017 11:03 AM - edited 03-22-2017 11:16 AM
Along the same lines an RavensFan:
I'm wondering if they are redrawn poorly. I can't test it myself right now (I did something bad to LabVIEW playing around with a VIA test vi so, I may need a reboot to blow out a vi that blows up the IDE when loaded) ARGH
So, try to change the bundle by name to:
EDIT: LabVIEW 2016 up again. Al you need is one wire bend to work around the issue
And it does not really matter which wire is bent going into the bundle by name
03-22-2017 11:09 AM - edited 03-22-2017 11:11 AM
@RavensFan wrote:
Which version of LabVIEW is this?
Can you share the VI that does this?
What if you don't do a BD cleanup, but just scroll the window so that that part of the code goes off screen and comes back again.
(Wondering if it is some sort of redraw bug and forcing a redraw fixes it.)
I actually just made a simple cluster of a few different types of numerics for the experiment above. I am using LV 2015 SP1. I tried the ol' scroll trick (that was my first guess) and I couldn't coerce LabVIEW into showing the dots again. Interestingly enough, it also happens if the result of the coercion also needs to be coerced. In this case, you get a coercion dot on that cluster input, but all the others go away.
03-22-2017 11:24 AM
@JÞB wrote:
Along the same lines an RavensFan:
I'm wondering if they are redrawn poorly. I can't test it myself right now (I did something bad to LabVIEW playing around with a VIA test vi so, I may need a reboot to blow out a vi that blows up the IDE when loaded) ARGH
So, try to change the bundle by name to:
- Short/ No names
- Bundle (not by name)
EDIT: LabVIEW 2016 up again. Al you need is one wire bend to work around the issue
And it does not really matter which wire is bent going into the bundle by name
Interestingly enough, this DOESN'T work in LV 2015! But doing just about anything to the bundle by name fixes it, inclding deleting the output cluster or adding one.
03-22-2017 01:27 PM - edited 03-22-2017 01:27 PM
LabVIEW 16f2, moving the function to force wire bends did not recharge the coercion dots.
Hitting Ctrl and the run arrow forced a recompile and that brought the dots back. Also interesting, I have constant folding turned on. The constant folding hashmarks disappeared with the coercion dots. Forcing a recompile or adding subtracting a wire caused the dots to show correctly and the constant folding hashmarks to return.
This feels to me a somewhat of a bug where certain changes aren't being completely detected in the compile process, and only doing something that forces a recompile updates the display correctly. So more of a "compile" visual bug rather than a simple graphical visual bug.
03-22-2017 03:06 PM - edited 03-22-2017 03:13 PM
@RavensFan wrote:
LabVIEW 16f2, moving the function to force wire bends did not recharge the coercion dots.
I must have been unclear. (I can see that) start with a coersion dot that is on the BBN top or bottom.
There is some other additional factor I can't duplicate. Interesting find on constant folding indication. Toss an AC in there and
Amazingly copying a constant does not produce a constant