Re: mkfs and fsck in /sbin

2002-11-09 Thread Thomas Bushnell, BSG
Robert Millan <[EMAIL PROTECTED]> writes: > On Sat, Nov 09, 2002 at 01:57:31PM -0500, Richard Kreuter wrote: > > > > Also, unfortunately, at the moment, the body of FHS requires mkfs, > > fsck, et al. in /sbin, if these files exist on a system (section > > 3.14.2), so Debian GNU/Hurd may need t

Re: mkfs and fsck in /sbin

2002-11-09 Thread Thomas Bushnell, BSG
Jeff Bailey <[EMAIL PROTECTED]> writes: > My point was that if we can reasonably show that on a GNU/Hurd system > that most (say 90-95%?) of the commands in sbin would be reasonable for > a power user to use there is probably value in just having that in the > general users path. You seem to be a

Re: mkfs and fsck in /sbin

2002-11-09 Thread Richard Kreuter
On Sat, Nov 09, 2002 at 08:12:01PM +0100, Robert Millan wrote: > On Sat, Nov 09, 2002 at 01:57:31PM -0500, Richard Kreuter wrote: > > > > Also, unfortunately, at the moment, the body of FHS requires mkfs, > > fsck, et al. in /sbin, if these files exist on a system (section > > 3.14.2), so Debian

Re: mkfs and fsck in /sbin

2002-11-09 Thread Neal H. Walfield
> 1) /sbin should be added to every users path on i386-gnu systems. The > concept of a binary that is completely unusable for regular users is > almost unheard of for us. (The only few that come to mind is init, > fdisk, and grub). Why would a normal user not want to run fdisk or grub? Think bo

Re: mkfs and fsck in /sbin

2002-11-09 Thread Marcus Brinkmann
On Sat, Nov 09, 2002 at 02:22:43PM -0500, Jeff Bailey wrote: > On Sat, 2002-11-09 at 14:09, Marcus Brinkmann wrote: > > > Certainly not. There is no point in sbin if it is i every users path. > > It is our believe that the distinction between sbin and bin is useful. > > Who is 'our' in this case

Re: mkfs and fsck in /sbin

2002-11-09 Thread Jeff Bailey
On Sat, 2002-11-09 at 14:09, Marcus Brinkmann wrote: > Certainly not. There is no point in sbin if it is i every users path. > It is our believe that the distinction between sbin and bin is useful. Who is 'our' in this case. The FHS committee? My point was that if we can reasonably show that o

Re: mkfs and fsck in /sbin

2002-11-09 Thread Marcus Brinkmann
On Sat, Nov 09, 2002 at 04:26:52PM +0100, Robert Millan wrote: > If a program hard-codes /sbin it means that it's being run as normal > user (without /sbin in PATH) but still needs that utility. so it's > just a workaround for the problem which is the lack of the utility > in users' PATH. No, the

Re: mkfs and fsck in /sbin

2002-11-09 Thread Robert Millan
On Sat, Nov 09, 2002 at 01:57:31PM -0500, Richard Kreuter wrote: > > Also, unfortunately, at the moment, the body of FHS requires mkfs, > fsck, et al. in /sbin, if these files exist on a system (section > 3.14.2), so Debian GNU/Hurd may need to provide these files here. And the second paragraph

Re: mkfs and fsck in /sbin

2002-11-09 Thread Marcus Brinkmann
On Sat, Nov 09, 2002 at 10:04:41AM -0500, Jeff Bailey wrote: > I think I see two potential solutions to this: > > 1) /sbin should be added to every users path on i386-gnu systems. The > concept of a binary that is completely unusable for regular users is > almost unheard of for us. (The only few

Re: mkfs and fsck in /sbin

2002-11-09 Thread Richard Kreuter
On Sat, Nov 09, 2002 at 04:26:52PM +0100, Robert Millan wrote: > On Sat, Nov 09, 2002 at 09:26:04AM -0500, Richard Kreuter wrote: > > On Sat, Nov 09, 2002 at 12:24:59PM +0100, Robert Millan wrote: > > > > > > On the GNU system it will be usual for normal users to create > > > and maintain their own

Re: mkfs and fsck in /sbin

2002-11-09 Thread Thomas Bushnell, BSG
Jeff Bailey <[EMAIL PROTECTED]> writes: > It's questionable that they should be artificially hidden in the first > place, but hey. =) The purpose of /sbin is not to "hide" anything, but to avoid cluttering users' command namespace with commands they can't usefully ever use.

Re: mkfs and fsck in /sbin

2002-11-09 Thread Jeff Bailey
On Sat, 2002-11-09 at 12:35, Robert Millan wrote: > do you mean that someday we eventualy won't need /sbin at all? if that's > the case i don't mind working the problem around by adding /sbin to PATH I think that's a nice long term ideal. As soon as you take the idea that users are gods of their

Re: mkfs and fsck in /sbin

2002-11-09 Thread Robert Millan
On Sat, Nov 09, 2002 at 10:04:41AM -0500, Jeff Bailey wrote: > > I think I see two potential solutions to this: > > 1) /sbin should be added to every users path on i386-gnu systems. The > concept of a binary that is completely unusable for regular users is > almost unheard of for us. (The only

Re: mkfs and fsck in /sbin

2002-11-09 Thread Jeff Bailey
On Sat, 2002-11-09 at 11:05, Robert Millan wrote: > > > Let's file on general, wait a reasonable period of time, then start > > > filing individual bugs with the certainty that noone will complain > > > on us filing mass-bugs. > > A bug on 'general' will be fowarded to debian-devel anyway and I s

Re: mkfs and fsck in /sbin

2002-11-09 Thread Robert Millan
On Sat, Nov 09, 2002 at 04:35:52PM +0100, Michael Banck wrote: > > Let's file on general, wait a reasonable period of time, then start > > filing individual bugs with the certainty that noone will complain > > on us filing mass-bugs. > > A bug on 'general' will be fowarded to debian-devel anyway a

Re: mkfs and fsck in /sbin

2002-11-09 Thread Michael Banck
On Sat, Nov 09, 2002 at 04:33:00PM +0100, Robert Millan wrote: > On Sat, Nov 09, 2002 at 03:34:00PM +0100, Michael Banck wrote: > > On Sat, Nov 09, 2002 at 12:24:59PM +0100, Robert Millan wrote: > > > Do you agree with that? In that case we should start fixing it > > > by filing a bug against packa

Re: mkfs and fsck in /sbin

2002-11-09 Thread Sean Neakums
commence Robert Millan quotation: > On Sat, Nov 09, 2002 at 03:34:00PM +0100, Michael Banck wrote: >> On Sat, Nov 09, 2002 at 12:24:59PM +0100, Robert Millan wrote: >> > Do you agree with that? In that case we should start fixing it >> > by filing a bug against package "general" and waiting for a

Re: mkfs and fsck in /sbin

2002-11-09 Thread Robert Millan
On Sat, Nov 09, 2002 at 03:34:00PM +0100, Michael Banck wrote: > On Sat, Nov 09, 2002 at 12:24:59PM +0100, Robert Millan wrote: > > Do you agree with that? In that case we should start fixing it > > by filing a bug against package "general" and waiting for active > > maintainers to fix their packag

Re: mkfs and fsck in /sbin

2002-11-09 Thread Robert Millan
On Sat, Nov 09, 2002 at 09:26:04AM -0500, Richard Kreuter wrote: > On Sat, Nov 09, 2002 at 12:24:59PM +0100, Robert Millan wrote: > > > > On the GNU system it will be usual for normal users to create > > and maintain their own filesystems, typicaly on removable media. > > > > Hence they need mkfs,

Re: mkfs and fsck in /sbin

2002-11-09 Thread Jeff Bailey
On Sat, 2002-11-09 at 06:24, Robert Millan wrote: > Hence they need mkfs, fsck and other filesystem-related utilities > available in the /bin directory instead of /sbin. > Do you agree with that? In that case we should start fixing it > by filing a bug against package "general" and waiting for ac

Re: mkfs and fsck in /sbin

2002-11-09 Thread Michael Banck
On Sat, Nov 09, 2002 at 12:24:59PM +0100, Robert Millan wrote: > Do you agree with that? In that case we should start fixing it > by filing a bug against package "general" and waiting for active > maintainers to fix their packages. You don't fix things by filing bugs on 'general', trust me. Micha

Re: mkfs and fsck in /sbin

2002-11-09 Thread Richard Kreuter
On Sat, Nov 09, 2002 at 12:24:59PM +0100, Robert Millan wrote: > > On the GNU system it will be usual for normal users to create > and maintain their own filesystems, typicaly on removable media. > > Hence they need mkfs, fsck and other filesystem-related utilities > available in the /bin director

mkfs and fsck in /sbin

2002-11-09 Thread Robert Millan
Hello, On the GNU system it will be usual for normal users to create and maintain their own filesystems, typicaly on removable media. Hence they need mkfs, fsck and other filesystem-related utilities available in the /bin directory instead of /sbin. Do you agree with that? In that case we shoul