On Fri, Dec 14, 2012 at 09:00:56PM +0000, Kevin Chadwick wrote: > > > Greg, can you write back to this message with specific examples of what > > > would need to be customized so that separate /usr would work right > > > without an initramfs? I have tried to explain multiple times that this > > > is a mis-conception that udev caused it, but I am getting nowhere. > > > > It's not my job to do this, nor yours, or fix any of these issues. It's > > up to the people who wish to keep a separate /usr partition without an > > initramfs to do this work. > > So even though you keep stating things without being specific like > udev is not a blocker, you have just admitted that the udev package > does violate the Filesystem Hiearchical Standard as well as the latest > draft when installing.
Specifically how does udev violate it? And also note, FHS is a trailing standard, documenting how things are done, not how they should be done. It can be changed. And since when did Gentoo start worrying about FHS and LSB? > I can understand following the current trend (some of which I agreed > with) but what is the justification for that which didn't already have > an optional solution? I don't understand, what in the udev package, or source, goes against FHS? > p.s. embedded does not equal mobile and android uses a leaner init > than /sbin/init and experiments posted to the buildroot list found > systemd to be slower, guessed to be due to increased cycles but perhaps > memory usage on even some mobile level processors which accounts for a > fraction of linux potential in embedded applications. POSIX compliance > is also a requirement by some major industries. What does POSIX have to do with anything here? And when did I bring up systemd and boot times? That's not what this thread is about, sorry. greg k-h
