On Jan 7 15:56, cyg...@kosowsky.org wrote: > > Corinna Vinschen wrote: > > On Jan 6 16:11, cyg...@kosowsky.org wrote: > > It is. I realized belatedly, that 3da9e136.acl is apparently a > > directory, not a file. > > It's actually a file...
This is weird. The meaning of the OI and CI markers are "Object inheritance" and "Container inheritance". These bits only make sense for directories and they control how ACEs are inherited by child objects (files) and child containers (subdirs). Consequentially, if I use `icacls /restore' on a file with the DACL saved by you, the OI and CI bits are simply ignored. After /restore, if I call /save again the resulting file looks like this: $ cat aclfile-after-restore.sav acltest D:PAI(A;;FA;;;SY)(A;;0x1200a9;;;WD)(A;;FA;;;BA) However, this gave me a clue. If this is really a file, it's a good chance that the inheritance flags are restricted to directories at one point in either the Cygwin DLL itself, or the getfacl tool. I'll have a look into the sources later, but I sure would prefer if I could create such a file locally. Thanks, Corinna -- Problem reports: https://cygwin.com/problems.html FAQ: https://cygwin.com/faq/ Documentation: https://cygwin.com/docs.html Unsubscribe info: https://cygwin.com/ml/#unsubscribe-simple
