Pádraig Brady <[email protected]> writes: >> Hi Collin, >> Thanks for your hint. >> Yes you are right, our glibc was built against 7.1. > So this would have failed if the previous coreutils 9.11 was built on kernel > 7.2 also. > It sounds like gl_CHECK_STRERRORNAME_NP could be augmented to checking EFTYPE > like it does for other problematic errnos.
Carlos O'Donell doesn't consider it a bug [1]. We would need to check every possible errno to make sure strerrorname_np doesn't return NULL. And then a kernel upgrade to could invalidate that anyways. I don't think it is worth the effort. The test failure is annoying, though. New errno values aren't added to often, I'd assume. Collin [1] https://bugzilla.redhat.com/2533579
