Bug in qt5-3d

[eringerli]

According to this bug report (https://bugs.archlinux.org/task/65449), a bug in GCC is responsible for a SEGFAULT on the start of even a simple qt3d program. I hit the exact same spot with my software, but I can't fix it myself. The bug report of it is under https://forum.manjaro.org/t/compile-qt5 ... -3d/133958

The fix is (relatively) simple: update the GCC in the toolchain from 9.2.0 to 9.3.0 or higher. This has to be fixed upstream so Manjaro can pull the changes.

[eringerli]

Is an update of the toolchain planned? Because I can't really go further with my project until this is fixed.

Also every piece of software which uses qt3d is affected, this includes QGIS. I develop a software which uses precision RTK-GPS (like this: https://www.u-blox.com/en/product/zed-f9p-module) to provide automatic steering for farming equipment like a tractor. Another usecase of this setup is using the precise GPS (which is exact to worst case 1-2cm, see the data sheet of the F9P) to find boundary stones or drainage systems in the fields. This is normally done with an rod, which has the antenna on top. At least in Germany and Switzerland, you can get the officially measured coordinates for this from the goverment. Without QGIS, is it difficul to transcode from the various map standarts to WGS84, which is used in GPS measurement. Also with QGIS you can do a lot of other stuff important to farming like measure the exact area of a crop, as Google Maps or other webservices with satelite images have some differences to the real word depending on the location and/or the fotos are sometimes quite old and don't reflect the land usage anymore. If you miss the boundary by a couple of meters by drawing it by hand/mouse, it makes a lot of area if the field is long enough, which costs you money in fertilizer/seeds or work done by other people like harvesting (which is paid by area).

So it is quite important for my project and the farming community for a working qt3d on the raspberry pi 3/4, which in turn depends on a fixed bug in gcc >=9.3.

[eringerli]

The log here https://archlinuxarm.org/packages/armv7h/gcc/log shows an update to 9.3.0-1 six weeks ago (also here: https://github.com/archlinuxarm/PKGBUIL ... c/PKGBUILD), but the version here https://archlinuxarm.org/packages/arm/gcc is still 9.2.0-4. Can someone actualize the packages? Or is there a reason not to?