On Dec 21 14:13, Thomas Wolff wrote: > On 18.12.2015 20:38, EXT Corinna Vinschen wrote: > >On Dec 18 18:11, Corinna Vinschen wrote: > >>On Dec 18 17:14, Thomas Wolff wrote: > >>>I wrote: > >>>>... > >>>>After removing SYSTEM write permission with setfacl, > >>>>it was effectively removed for SYSTEM but the other groups got > >>>>write permission ADDED instead (as also properly indicated by ls) − > >>>>which is kind of the opposite of the intended operation. > >>>cygwin-2.4.0-0.11, sorry > >>In that case the behaviour is by design. Try the same on Linux and the > >>result will be the same. Every time you change group perms, the mask > >>will be changed to reflect the maximum permissions given to any group or > >>seccondary user. You always have to check the mask or set it explicitely > >>to the desired value. > >I'm sorry, but I forgot to mention an important part: Recomputing the > >mask is *not* done in the kernel or, in our case, Cygwin. Rather this > >functionality is part of the setfacl tool. Setfacl recomputes the mask > >by default. There's a new option -n/--no-mask as on Linux to retain the > >current mask setting, e.g. > > > > $ setfacl -n -m g:wheel:r-x file > > > >Try setfacl --help for a comprehensive description of all options. > > > > > >HTH, > Yes, thank you. > Just pondering: > "...the maximum/union of all permissions..." could well be interpreted as > "... all *effective* permissions"
Uh, no. The effective permissions are a *result* of applying the mask, so they can't constitute the mask. Stimulus/response are unambiguously defined here. > which would make a difference in the presented case. > Anyway, you are right, this is an upstream design issue. And upstream in > this case seems to mean referring to a standard that isn't even officially > available anymore... Heh, yes. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Maintainer cygwin AT cygwin DOT com Red Hat
signature.asc
Description: PGP signature
