On Sunday, January 29, 2023 11:04:22 PM CET Paul Eggert wrote: > On 2023-01-29 03:06, Kamil Dudka wrote: > > On Wednesday, January 25, 2023 11:01:45 PM CET Paul Eggert wrote: > >> On 2023-01-25 13:56, Ondrej Valousek wrote: > >>> But it's not the same meaning. What I am trying to explain here is that > >>> > >>> Cp -p (or cp --preserve=mode) also retains ACLs. This fact is not > >>> obvious, but yet it's happening > >> Then I'm afraid I don't understand. In what sense do ACLs differ from > >> xattr here? > > As I understand it, `cp -p` now preserves ACLs but not xattr (unlike `cp > > -a`). > > OK, the light is slowly dawning on me. Though I'm still confused. > > Why are ACLs treated differently from extended attributes? Shouldn't the > two be treated the same (assuming they're both supported)? > > In other words, what's the underlying model and motivation here? It's > more important to document that, than to document little pieces of it.
I cannot speak for Ondrej. My understanding is that ACLs can be seen as extension over permission bits whereas the extended attributes can store pretty much anything. I am not saying which approach is (more) correct though. Kamil