Nicolas Williams wrote: > On Wed, Dec 17, 2008 at 05:59:00PM +0100, Pavel Filipensky wrote: > >> Just a comment - I think that the path can contain apart from space also >> newline + tab. >> > > bquote() handles tabs! > > Newlines, not so much :( > > But we can add support for newlines too. It's a bit more complicated, > but not impossible. > > >> Maybe changes similar to bquote () would handle that, but: >> - it needs a really mastership in shell programming >> > > That's OK, that's what we're here for :) > > >> - the umounatll(1M) would make a BIG effort to get correctly the data >> out from the kernel, just to pass it later back to the kernel >> >> I second Frank's idea to file an RFE for redesign of umountall(1M) and >> if this is implemented, then 6786256 might be simpler to fix. >> > > I agree that a syscall would be better, more reliable. > > But if you need a quick fix, then stay on your current path. > > So let's at least get 6778894 fixed, and you could fix 6786256 for > everything but newlines using the code I sent you. That would add value > with little effort. > > Nico > Yes - the plan is to fix 6778894 now, it is independent of fix for 6786256. I propose to keep 6786256 open till one of these happens: a) there is a real need for the fix (it gets higher priority or get escalated) b) done as part of redesign of umountall(1M)
I have recorded your code to Public Comments section of 6786256 in case a) happens. Thanks, Pavel
