Hello, Joan Lledó via Bug reports for the GNU Hurd, le dim. 16 févr. 2020 13:23:35 +0100, a ecrit: > El 15/2/20 a les 15:02, Samuel Thibault ha escrit: > > Why a capital b? It's be more coherent to have -d -b -s -f all > > small-caps. > > > > Because of this notice in --help: > > "More than one permission scope may be specified. Uppercase options > create a new permission scope if the current one already has a value for > that option. If one node is covered by more than one permission scope, > only the first permission is applied to that node."
Ok. The text is really not easy to understand, but with your examples I understand what you meant. This seems to me quite difficult for users to understand, and to me it unnecessarily conflates the two unrelated notions of PCI specification and scope. I would say it would be much simpler to make the -U and -G options create new scopes, and let -d/-b/-s/-f be lowercase. I.e. we'd have sets of -d/-b/-s/-f options separated by sets of -U/-G options, so instead of : > -B 0 -s 5 -f 0 -B 0 -U 1000 we'd have: -b 0 -s 5 -f 0 -U 1000 -b 0 > This will create two permission scopes: > 1.- 00:05.0 belongs to nobody (Error) > 2.- 00:*.* belongs to 1000 (OK) And this confusing example: > -B 0 -s 5 -f 0 -s 4 -U 1000 would be -U 1000 -B 0 -s 5 -f 0 -s 4 which is more obviously bogus: -s is set twice in the same uid=1000 scope. Samuel
