From: David Mazieres (dm@reeducation-labor.lcs.mit.edu)
Date: 03/07/00-09:16:30 AM Z
Date: Tue, 7 Mar 2000 10:16:30 -0500 (EST) Message-Id: <200003071516.KAA31785@reeducation-labor.lcs.mit.edu> From: David Mazieres <dm@reeducation-labor.lcs.mit.edu> Subject: Re: attributes of renamed file > Date: Tue, 07 Mar 2000 16:05:06 +0100 > From: Michael Salmon <Michael.Salmon@uab.ericsson.se> > > The ctime is only updated when the inode changes and the file's name is not > coupled to the inode. File names only exist in directories, a file doesn't > even need to have a name. Perhaps in part for historical reasons (since rename used to be a non-atomic link and unlink), rename bumps the ctime of a file. It would be very difficult to do things like incremental backups of a directory without this feature, however, so I think it's not a bad thing. More concretely, though, I just tried this out on a solaris box, and rename definitely bumps the ctime of a file, whether or not you agree with the behavior. It has been my experience that all Unixes do this. David
This archive was generated by hypermail 2.1.2 : 03/04/05-01:48:06 AM Z CST