Rats nest for unconnected pads in not displayed properly in OpenGL view

[Art]

Two separate issues that I noticed which are not present in the default view.

1. I have an rats net line present that supposedly shows that that pad is unconnected. The pad however is connected and it passes DRC check. When I press Hide Ratsnest button, it goes away, so it is a ratsnest line. The line is not present in the default view.

2. I had another pad that was not connected. When I ran DRC check in OpenGL and then clicked on the error in the report the cursor jumped to some other pad next the the unconnected one and highlighted that other net. When repeated, the cursor would always jump to the other pad but sometimes only portion of the track would highlight, sometimes it wouldn't highlight at all. When DRC check is run under the default view, the cursor would jump to the correct pad. When the problem was fixed and pad was connected, the ratsnest line was still present and wouldn't go away in the OpenGL view, although it would pass DRC check just fine.

See pictures attached

Running KiCad 6097 on Windows 7 64 bit

[Maciej Suminski (orsonmmz)]

Hi Art,

Is it possible to get a board that exposes the problem?

Regards,
Orson

[Art (diametrix)]

After opening this file again I noticed that the artifact was gone. Took me a while to figure out what was going on. If you have a board is fully routed and you change the footprint that requires you to flip a footprint (i.e. swap the pin numbers on a resistor or a diode) when the new netlist is read by PcbNew and you flip that footprint, the rat's nest lines don't go way. Even when you do rebuild board connectivity the lines are still there. You have to restart PcbNew for them to disappear.

[Maciej Suminski (orsonmmz)]

Hi Art,

I have just tried to follow the steps you described, but I could not reproduce the problem (revision 6142). Could you try a newer revision or provide a simple board that exposes the issue? Thank you in advance.

Regards,
Orson

[Art (diametrix)]

Seems to be working properly now. Just one more thing that I noticed. Not really a show stopper, but previously a net was considered to be making a connection when any part of it was touching any part of the pad. Now, it seems that the track must be snapped precisely to the center of the pad. I always do that in the default mode. But I changed the footprint for the SMD resistors, which moved the center of the pads slightly. Now most of the resistors have tiny ratsnest lines in the center of the pads. (See picture)

[Maciej Suminski (orsonmmz)]

Thank you for the feedback. I have just fixed the issue you described in the post #5 in revision 6143.