home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.unix.sysv386:16608 comp.unix.sys5.r4:549
- Newsgroups: comp.unix.sysv386,comp.unix.sys5.r4
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!uwm.edu!linac!uchinews!machine!chinet!les
- From: les@chinet.chi.il.us (Leslie Mikesell)
- Subject: Re: tar problems (and cpio)
- Message-ID: <BxxJGn.ry@chinet.chi.il.us>
- Organization: Chinet - Public Access UNIX
- References: <1992Nov12.164004.9850@netnews.whoi.edu> <1992Nov16.232139.27765@dg-rtp.dg.com> <1992Nov17.185512.516@fwi.uva.nl>
- Date: Wed, 18 Nov 1992 20:55:34 GMT
- Lines: 18
-
- In article <1992Nov17.185512.516@fwi.uva.nl> janw@fwi.uva.nl (Jan Wortelboer) writes:
- >GNU cpio support's also SVR4 cpio (New MagicNumber) and SVR4 tar.
- > ^^^^^^^^^ ^^^^^^^^
- Great! R4's cpio is a real botch. If you have a mix of r3 and r4 machines
- connected with NFS or RFS there is no way to use a single command (say
- a shell script executed over the network mount point) to make an archive
- of files that will be readable on any of the machines. At least not
- without some seriously ugly contortions.
-
- Also, I use a script to copy directory trees around that ends up using
- cpio -p to do the work and I've noticed that copying from an r4 UFS
- filesystem to an r3 sysV filesystem over RFS leaves the date wrong on
- the destination. For some reason make doesn't work well when all of
- the targets exist with dates 20 years in the future and you just touched
- all the sources. Will GNU cpio fix this as well?
-
- Les Mikesell
- les@chinet.chi.il.us
-