From: Tim <obe...@we...> - 2004-05-29 01:01:26
|
Hi! On Fri, 2004-05-28 at 21:05, Kern Sibbald wrote: > I think it is best that the user specifically request to backup the > ACLs. This can be handled at a higher level. Yes, I agree. > A few questions: > > - Are these ACLs on Linux systems? I'm testing on Linux at the moment. But ACL support is available since Kernel 2.6.x or (2.4.x with a patch), but it is optional. So I suppose in most Distributions ACL support must explicitely activated. This was the case in mine (Gentoo Linux) and a few others like Red Hat 9.0, SuSE Linux 9.1 Pro / Personal, SuSE Linux Enterprise Server 8. > - Is it easy to detect that they exist or not? I don't know, but I think it will be easy. I'll check that. > - Are there any other kinds of ACL systems implemented? The ACL library implements the abandoned POSIX standard: "IEEE Std 1003.1e draft 17 ("POSIX.1e", abandoned)" And because, the library itself is ported from FreeBSD I think my code will work there too. > - If libacl does not exist will Bacula link correctly? If not, we will > need some way to automatically detect libacl in configure and set a > variable so that they can be #ifdefed off. No, the code will not link correctly without libacl, so an automatic detection is the best way I think. Bye, Tim |