Net labels seem to be broken in version 0.9.3b

#1

Steps I took that resulted in the problem:

Attempting to create an Arduino Nano projects - I created net labels for most of the pins, then created matching net labels for several screw terminals and male connection points. Everything seemed ok at first, then I began to notice that several nets with different names were being treated as the same net on the pcb layout. I deleted the faulty net labels and created new ones. That seemed to fix the issues, till I found another one that was cross-connected. So I deleted those and created new ones and that fixed that issue, but fixing that one caused several others to become cross connected.

What I expected should have happened instead:

Each uniquely named net label should be a unique net, with no cross-connections between labels of different names.

My version of Fritzing and my operating system:

0.9.3b (I understand that this is a beta version, but this is a pretty bad problem, so I thought I should report it.)

Please also attach any files that help explaining this problem

#2

The current production version is 0.9.4b

That description is not enough to try to identify the cause.

What operating system environment are you using?

A sample sketch that has the problem might help, combined with specifics about which nets/labels have the problem.

Better would be the steps needed for someone else to be able to reproduce the problem.

This is very likely going to be completely ignored, unless it can be reproduced in version 0.9.4 and 0.9.5.

This might not have anything at all to do with net labels. There are other reports about incorrect connections that have been blamed on creating connections (wires) in different views (not creating in a single view, then ONLY creating wires in other views that match ratsnest lines), then deleting some of those wires in different views. Certain sequences of add/delete seem to leave behind invisible wires.

#3

Thanks for the reply @microMerlin.

I was running on Windows 10 - I’ve since moved my project over to my Linux box.

0.9.3b was the latest version I could find on the internet that I could download without making a contribution. I love to contribute to good projects, but this was my first introduction to Fritzing and so I wanted to “kick the tires” so to speak before committing cash.

As it happens, I love Fritzing - in spite of the bugs. I’m 35 year veteran software engineer with open source experience, so I can always pull the source and fix issues if I want to put the time into it. So I went to the official site and paid to download the latest.

After upgrading, I didn’t see any more net label issues, but that is perhaps because I manually fixed them all before loading my project into the latest version. Fixing the problem involved deleting all the nets with issues and recreating them manually - there were a half dozen - it was not hard.

Creating a small sample with the problem is difficult because I don’t know what I did to cause it. I didn’t do anything particularly strange, but one thing may have contributed to the cause - I discovered ctrl-d (duplicate), which is an incredible time saver. Of course, this creates ports with the same net label as the source object, so there’s a good chance of creating an object that has some connection with the source object.

The problem doesn’t show up in the schematic view. I discovered it when I moved over to the PCB view to create my board and found several nets were interconnected (It was strange to see a microcontroller with all the digital pins tied together!)

I started writing an EDA program very similar to Fritzing (before I discovered it) last year after playing with the free stuff that was out there - i.e., KiCAD, etc. Most software was clearly written by people the really understood EDA, but didn’t have much experience designing software. My program was even using QT. I got as far as the canvas and a few parts, but gave up when I realized how much time it was going to take. Needless to say, i was thrilled to discover Fritzing - and that it was open source!

Regards,
John

#4

I started to enter a bug in the issue board and found this is a long-standing issue (3 years):