Re: acl: request for testing

[Bruno Haible]

Eric Blake wrote:
> On FAT32 (which has no ACL support), one failure and one spurious output:
> 
> file_has_acl("tmpfile0") returned no, expected yes
> FAIL: test-file-has-acl.sh
> 
> setfacl: illegal acl entries
> PASS: test-copy-acl.sh

Thanks for testing. Can you show the complete output of

   TMPDIR=/fat32dir VERBOSE=yes make check

At first sight, I would guess that the setfacl program should signal an
error, and that in this case the test-file-has-acl.sh test should skip
the following statements. But one needs to look at the details.

> on NTFS, one failure:
> 
> files tmpfile0 and tmpfile4: different ACL entry #1: different permissions
> 000 and 004
> FAIL: test-copy-acl.sh

Here, can you show
  getfacl $TMPDIR/gl..../tmpfile0
and
  getfacl $TMPDIR/gl..../tmpfile4

Then
  TMPDIR=/fat32dir VERBOSE=yes make check
to see which program is executed (probably
  gltests/test-copy-acl $TMPDIR/gl..../tmpfile0 $TMPDIR/gl..../tmpfile4
), and then you can start debugging it.

Bruno