From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Sat, 6 Jun 2015 02:46:43 +0200 Subject: [Buildroot] [PATCH 04/12] fs/iso9660: enable Joliet extension In-Reply-To: <20150605211840.GD3641@free.fr> References: <1433430330-2166-1-git-send-email-thomas.petazzoni@free-electrons.com> <1433430330-2166-5-git-send-email-thomas.petazzoni@free-electrons.com> <20150605211840.GD3641@free.fr> Message-ID: <20150606024643.15fb2bc7@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Dear Yann E. MORIN, On Fri, 5 Jun 2015 23:18:40 +0200, Yann E. MORIN wrote: > Only tangentially related to your change: > > -R is for RockRidge, but it seems using -r would be better: > > -r This is like the -R option, but file ownership and modes are set > to more useful values. [--snip--] > > Reading the full entry for -r is instructive, and makes it sound like > that's what we would want to use instead of -R. I'm actually not sure. Let's list what -r is doing compared to -R: * Setting uid:gid to 0:0 to all files. We already do that using fakeroot, and we also set some other ownership for specific files using _PERMISSIONS and the device table. So getting all of that reset to 0:0 is wrong. * Changing permissions all over. Why do we care? We already take care of dependencies, and -r makes everything world-readable, which we clearly don't want. * It apparently clears the set-id bit, which would make all setuid application to break. So, in fact I'm rather convinced that we should keep -R and not use -r. Best regards, Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com