On Wed, Jan 19, 2022 at 06:50:41AM -0800, Andrew G. Morgan wrote: > Thanks for finding this. > > Fixed in: > > https://git.kernel.org/pub/scm/libs/libcap/libcap.git/commit/?id=f25a1b7e69f7b33e6afb58b3e38f3450b7d2d9a0 Thanks! FWIW, I agree that we should not have depended on it being initialized to zero at the beginning program. It was just wierd since it wasn't, and I was wondering if it was some exotic (musl, dietlibc) C library that was triggering it --- and then was surprised when it turned out to be glibc. The Jon Postel principle of "Be conservaive in what you send, and liberal in what you accept" is good one to follow, so while it's nice to some userspace programs that might not be as well-behaved as to not depend on the status of errno when main() is called, we should fix the calling program in xfstests as well. Cheers, - Ted
