Pcbnew->Edit->Global Deletions causes segmentation fault

[Maarten]

1. Start pcbnew (from console)
2. Open design
3. Edit -> Global Deletions -> Enable Tracks -> Ok -> Yes
4. Segmentation fault

Running KiCad on Fedora 24

Version:
kicad-4.0.2-2.fc24.x86_64

[Maarten (madebr)]

stacktrace created from core dump

[Nick Østergaard (nickoe)]

Please answer the following questions:
1. Does this also happen in the nightlies?
2. Does this happen for any board or just a specific board 100% of the time?
3. Do you have an example board that exhibits this segfault you can share?

[Maarten (madebr)]

1. The segfault also happens in the current nightly on Fedora 24 (using my pcb layout)
2. It happens all the time on my current lay-out
3. I attached my (partially routed) lay-out showcasing the segfault. The first version of it was created using the kicad version bundled with fedora 22 (2015-mar-04 BZR unknown)-product

[Nick Østergaard (nickoe)]

I can reproduce the issue on 4.0.2 on archlinux, but not on 6996.

[Nick Østergaard (nickoe)]

Hmm, I have release build of 6996 where it also crashes, but I shall note that I don't hvae the python scripting stuff enabled there either.

[Nick Østergaard (nickoe)]

This also appears on 6996 with scripting and release build type. I have been trying to set some breakpoints before the code section, but I don't know why it does not seem to work. It seems to not break even if I set it at the previous frame function and proceedes to the segfault.

[Michael Steinberg (decimad)]

The BOARD class would access the first Track of its tracklist no matter if it has any Tracks at all.

[Michael Steinberg (decimad)]

Fixing the patch after Nick's response.

[Nick Østergaard (nickoe)]

I can confirm that Michael's v2 patch works now. I do not get the segfualt on global track delete anymore ;), tested on 6996.

[Maciej Suminski (orsonmmz)]

Thank you Michael, your fix has been committed in revision 7003.