Theodore Ts'o <tytso@xxxxxxx> writes: > On Sat, Nov 11, 2017 at 08:55:40AM +0100, Marius Bakke wrote: >> Commits 257f009d1034737096f197dc875931207bcdfc75 and >> 2b50c18a8a12781c26b71873f4368fe61711d9f9 missed one pass in expected >> output. >> >> Signed-off-by: Marius Bakke <mbakke@xxxxxxxxxxxx> > > Hi, what architecture did you discover this on? I don't need this > change on x86, but I was seeing some mysterious regression test > failures on a big-endian system which seemed to look like this. I > didn't have time to follow up on it then, but I suspect that this may > be what you are running into. Thanks for the reply. I get this on x86, but upon further investigation, it only happens in the Guix build container. Running "make check" interactively in a terminal works. Do you have any idea what may cause these two tests to extend the inode table when run in the isolated build chroot?
Attachment:
signature.asc
Description: PGP signature
