Hi, commit e28d75ed7204d7fac2c0fac13978098530b0574e dropped some logic during refactoring. PCI-based programmers now fail to initialize IFF the desired PCI device not the last one enumerated by libpci.
Given that this refactoring was done to make unit tests easier, I'd expect the unit tests to have caught that problem. This raises a few questions: - Why did unit tests not find that problem? - If there are no unit tests testing for such a problem, why was the code refactored in the first place? - Do we need CI tests with real hardware or at least with qemu to avoid such problems in the future? Regards, Carl-Daniel _______________________________________________ flashrom mailing list -- flashrom@flashrom.org To unsubscribe send an email to flashrom-le...@flashrom.org
