On 12/06/18 09:44, Joerg Schilling wrote: > Wols Lists <antli...@youngman.org.uk> wrote: > >> On 11/06/18 09:54, Joerg Schilling wrote: >>> Well, "Windows ACLs" is the only ACL system that is standardized (as part >>> of >>> the NFSv4 standard). The old proposal in POSIX.1e from 1993 from Sun has >>> been >>> withdrawn in 1997 since the customers did not like it. >>> >> Ummm - just because it's standard doesn't mean it's any good :-) > > Is is a result of a common discussion. At the same time, when Sun introduced > NFSv4 ACLs, IBM and Apple did the same for their local filesystems. > >> This version I'm talking about dates from about 1983. The company making >> it went bust in 1991. > > What are you talking about?
Pr1me. Okay, I don't remember most of the dates accurately, but Pr1mos 19.4 had a working Access Control List setup. I was using that on their Pr1me-2250 machines, at a company I left in 1984. (Wikipedia says the 2250 was released in 1982. I can't find a date for 19.4.) > > IIRC, the first ACLs have been on VMS in the late 1980s. > >> I've just had a quick look at the NFS v4 RFC, and almost the first thing >> I see is DENY entries. These ACLs don't have deny, because it's >> pointless. And DENY is exactly why I think Posix/Windows ACLs are >> confusing and hard to use. > > Your text looks confusing. You claim DENY entries and no DENY entries in the > same paragraph without explaining what you are talking about. The RFC talks about deny entries. Pr1me ACLs didn't have deny, because it doesn't make sense in that context. > > Jörg >
