andreas-gruenbacher / richacl Goto Github PK

Hello, regarding #9
In case work is resumed on this project please be kind and read this blog post somebody wrote before going further
http://michael.orlitzky.com/articles/there_was_an_attempt_to_save_linux_filesystem_acls.xhtml

I think he wrote everything that had to be said about the linux ACL implementation pitfalls and why nobody wants to use these, probably fixing those issues would get more traction in an ACL system, this would be really helpful in container based workloads.

Current posix ACL's are unusable and RichACL can still be fixed.

Thank you for the work you've done anyway!

Why the owning group has only read access ? It should have write access ?
The following example is from "man richaclex"

Hi,

[ systemd-tmpfiles has support for acl setting (using traditional acls) [http://www.freedesktop.org/software/systemd/man/tmpfiles.d.html#a]. It seems that we should be able to add richacl support without much trouble, once richacl land in the kernel. I think we can allow entries for both normal acls and rich acls to be present, and pick one based on what the filesystem supports. ]

I was hoping that new library would solve a problem with libacl. When parsing acls, it would often help to know why parsing failed. For example, if the user or group cannot be resolved, this should be treated in a special way. This would allow smarter error handling in the initramfs where some users are not present and acls should be silently skipped. libacl doesn't really support this, because it just returns a single error code.

richacl_from_text and similar functions should set a meaningful error codes (ESRCH, ERANGE, etc, whatever is appropriate) to allow the caller to programatically distinguish reasons for failure.

The install documentation refers to two scripts (Makepkgs and install-sh) in the top level directory. Both those scripts don't exist anymore. I'm still looking into the installation process (for packaging) so not sure what else might be wrong.

Hi, since your domain bestbits.at is not available anymore and the RichACL Patches aren't reviewed yet by anyone who feels responsible for the VFS, how is the actual status of this Project? Are you still working on it?