See the patches here: https://www.redhat.com/archives/libosinfo/2018-September/msg00151.html And Fabiano's response here: https://www.redhat.com/archives/libosinfo/2018-October/msg00000.html If say win8 XML lists ich6 device, win8.1 should not list it, otherwise we end up with duplicate and misleading data everywhere. The patch series fixed some issues found by semi-manual inspection. But ideally we would have a test case of some sort that would detect duplicate device definitions in an OS hierarchy
Patches have been posted to the libosinfo ML: https://www.redhat.com/archives/libosinfo/2018-November/msg00062.html
master: 4c9224889be89213f4a64dc5ffc170041d0eb598