On 02/11/2013 08:47 PM, Assaf Gordon wrote: - if (errno != 0) + /* EINVAL can happen if 'base' is invalid (hardcoded as 10, so can't happen), + or if no conversion was performed (on some platforms).
We plan to release coreutils-8.20 early next week, so any testing you can do between now and Monday would be most welcome. Here's the NEWS: http://git.sv.gnu.org/cgit/coreutils.git/tree/NEWS coreutil
Fedora 15 - x86_64 coreutils gnulib RHEL 6.1 gnulib solaris - sparc coreutils All the above due to: cp -a x y from within /tmp (swap) which warned about acl failures. truss shows: acl("y/x", SETACL,
... ... ... ... ... Thanks for all the tests. FYI, I've confirmed all tests pass on these (all x86_64): Fedora 17 Fedora 18 Fedora rawhide Debian unstable Also just tested on an i386 with minimal Fed
Here's a preview of what coreutils-8.18 should look like. I noticed one spurious test failure reported by the Nixos Hydra autobuilder: http://hydra.nixos.org/build/2924158 It's in the most recently a
Hmm, I think this is not enough: df can also fail due to other reasons, e.g. when a mount point is not reachable by the user: $ src/df > /dev/null ; echo $? src/df: ‘/root/backup’: Permission den
Good idea, but doesn't work either: $ src/df . ; echo $? src/df: Warning: cannot read table of mounted file systems: No such file or directory Filesystem 1K-blocks Used Available Use% Mounted on - 35
But then it won't fail. It's tricky :) Updated version that explicitly matches the error is attached. cheers, Pádraig. Attachment: df-tests.diff Description: Text document
I thought about that case again, and to me something's still wrong when mtab is missing. When mtab cannot be read, then df cannot determine the file system type. Example (/tmp/mnt is an nfs mount): $
Oops, sorry, one of the cases in the new test failed. I fixed it: -LD_PRELOAD=./k.so df --total || fail=1 +LD_PRELOAD=./k.so df --total && fail=1 Additionally, as this patch didn't make it into 8.18,
Thanks. It came a little too late for 8.18. Here is the patch that I expect to squash in. Most are nits, but there was an exploitable bit of (debugging?) code in the new test. diff --git a/NEWS b/NEW
Thanks for the rewording both in NEWS and cu.texi. Often it's still hard for me to find the right wording in English ... I wonder if there's something like "make diff-syntax-check" which enforces CU'