On Fri, 2008-03-28 at 16:45 +0100, Denys Vlasenko wrote: > On Friday 28 March 2008 16:22, Joakim Tjernlund wrote: > > > > On Fri, 2008-03-28 at 09:44 -0500, Kevin Holland wrote: > > > You have to use something like > > > cat myfile > /dev/mtdblock/1 > > > because you don't have a filesystem mounted to that mtdblock I'm > > > assuming. If there is a filesystem you would mount it then copy the > > > file to the mountpoint. > > > Kevin > > > > Yeah, I know the cat trick works but so should cp too, I think. Earlier > > I used GNU cp and that worked like that. Compare with symlinks, cp > > copies the contents, not the symlink itself(unless -d or -P is given) > > Well, GNU cp also copies TO dest symlink's target too, > which is incredibly careless. Hell knows where that symlink points - > /etc/passwd? /dev/sda? Cool, eh?
Very :) But then again perhaps this is the right thing to do. Why follow the symlink in one direction but not the other? > > Instead of wanting cp to be a mix of copy and cat, why don't you use > cat when you want to say "please open and read from this file/device/pipe"? > That would be unambiguous. (Same holds for writing TO things - use >file). maybe, but then cp should be fixed to not copy the contents of a symlink? What about cp myfile /dev/mtdblock1? Currently cp copies the contents of myfile to the flash. > > As it stands now, cp's code is already a maze of heuristics > "what user actually wants, dammit?" Maybe cp should just open and copy? No special handling by default. Jocke > -- > vda > _______________________________________________ busybox mailing list [email protected] http://busybox.net/cgi-bin/mailman/listinfo/busybox
