I started looking into it, but quickly found that all or almost all the failures are all due to the rpm upgrade, and the failures all happen on three old distros: centos 7, debian 8, ubuntu 16.04.

What is particularly odd is that the centos/debian failures are non-deterministic: in some cases rpm-native fails to build due to missing omp.h, in other cases it builds, but the resulting binaries segfault. This non-determinism seems to happen even on the same worker, e.g.:
https://autobuilder.yoctoproject.org/typhoon/#/builders/75/builds/1812
https://autobuilder.yoctoproject.org/typhoon/#/builders/15/builds/2073
(one is failing in do_compile, another compiles fine, but then produces segfaults)

Ubuntu always segfaults, but maybe it's just coincidental.

I am not sure how to take this further, but maybe the best course of action is to defer the rpm/dnf update once more, until those old distros are finally retired.

Alex

On Sun, 26 Apr 2020 at 20:29, Alexander Kanavin via lists.openembedded.org <alex.kanavin=gmail.com@lists.openembedded.org> wrote:
On Sun, 26 Apr 2020 at 19:35, Richard Purdie <richard.purdie@linuxfoundation.org> wrote:

There were some failures from this series:

https://autobuilder.yoctoproject.org/typhoon/#/builders/83/builds/890

It may be easiest for you to have a look through that build result as I
think all the failures are from your patches as that is the main thing
in -next! :)

Thanks Richard, I'll work through these one by one. Hopefully the amount of actual issues is a lot less than the amount of red in that build.

Alex