On 02/20/2016 05:37 PM, Andreas Gruenbacher wrote:

Hi Michael,

thanks again for all the feedback. I've followed all your suggestions;
again, please see the github repo for the latest version:

  https://github.com/andreas-gruenbacher/richacl

On Sun, Feb 14, 2016 at 10:31 PM, Michael Kerrisk (man-pages)
[off-list ref] wrote:

quoted

Hi Andreas,

Here's a few more comments on the current richacl(7) page
that I fetched from the git repo.

[...]

quoted

quoted

.HP
.BR dir_inherit "\ (" d ):
The entry is inheritable for directories.

"When this flag appears in the ACL entry of a directory, then that entry is
inherited by new subdirectories created in the directory."

Is that text that I propose correct?

It's not entirely wrong, but only a small part of the truth. Entries
are inherited such that the permissions in file_inherit entries become
effective for access checking for new files, and the permissions in
dir_inherit entries become effective for access checking for new
directories, recursively.

Your feedback prompted me to look into the inheritance flag
computation more closely once again, and I found bugs. The steps of
the algorithm are explained in the section "Permissions at
file-creation time", by the way.

Hmmm... writing and refining documentation helps find code bugs. 
Who knew? :-)

Cheers,

Michael

-- 
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/