LabVIEW
BUG Block diagram cleanup fails completely? Generates very bad results
@billko wrote:
I can only guess that he doesn't mean that they literally break, but that they make some funny turns even when they apparently don't need to.
No, they break. I have had cases where BDC will leave dangling bits of wires. It was a really ugly diagram (not mine). I have been getting used to BDC as a quick way to make reasonably readable code. It has its oddities but it does save a lot of moving things 1 pixel at a time.
The instance I had, I think it was dangling bits. But there may have also been completely isolated bad wires with no connections at all.
I don't think it broke any connections BUT the diagram I used as a massive test case was some of the worst programming I have ever seen. It was about 5 X 6 large screens but wires and icons were on top of each other and completley unreadable. The system kept setting and reading values through properties and must have had horrible race conditions. When CBD got through with it, 5 minutes later, it was about 20 X 30 screens in size but at least the wires were visible.
It was a stress test of BDC and yes it has been passed to NI as an example of a BDC that fails. These should be given to NI so that they can find these weird edge cases and fix them.
@sth wrote:
The instance I had, I think it was dangling bits. But there may have also been completely isolated bad wires with no connections at all.
I don't think it broke any connections BUT the diagram I used as a massive test case was some of the worst programming I have ever seen. It was about 5 X 6 large screens but wires and icons were on top of each other and completley unreadable. The system kept setting and reading values through properties and must have had horrible race conditions. When CBD got through with it, 5 minutes later, it was about 20 X 30 screens in size but at least the wires were visible.
It was a stress test of BDC and yes it has been passed to NI as an example of a BDC that fails. These should be given to NI so that they can find these weird edge cases and fix them.
I think there is actually a limit to how big the BD can be:
Maximum Size for Front Panel and Block Diagram
You are probably very near it and weird things happen when you get close to it. I'm guessing the BD coughs up a hairball in that VI for the same reason.
@billko wrote:
You are probably very near it and weird things happen when you get close to it. I'm guessing the BD coughs up a hairball in that VI for the same reason.
Might be. I doubt that was the problem but that some where there are wire "loops" that got left dangling. However that article is LV 7.1 and we might have moved to 32 bit canvases since then. That seemed to be a signed 32 bit number for the window size.
I will check it out.
