On Jan 15 15:03, Corinna Vinschen wrote: > On Jan 15 13:21, Roland Schwingel wrote: > > [...] > > 11. cp -rp //server/share/path/to/your/home/tmp/test /tmp > > 12. Error message appears: > > > > cp: will not create hard link: '/tmp/test/symlink' to directory > > '/tmp/test/subdir' > > > > Obviously cygwin 1.5 is now detecting the symlink correctly while 1.3 was > > not because it copied > > the symlinked folder as folder not as link (even when I did cp -a). > > Cygwin 1.5.25 uses the i-node numbers returned by > Samba > 3.x as ... i-node numbers. Cygwin 1.3.x generated always pseudo > i-node numbers on remote drives. > [...] > cygwin-1.5.25$ ls -li > total 0 > 8808978157900 drwxr-xr-x 1 corinna root 0 Jan 15 14:46 subdir > 8808978157900 drwxr-xr-x 1 corinna root 0 Jan 15 14:46 symlink > [...] > I have no idea why Samba doesn't return the correct i-node number for > symlinks. This is most unfortunate.
A look into the Samba source code shows why. Samba always calls stat(2) on files if the client is a Windows client. It uses lstat(2) only for UNIX CIFS clients. That explains the identical i-node numbers in Cygwin. Sigh, a bit more intelligence would have been nice :( Eric, maybe cp(1) can be pushed up a bit by checking the number of hardlinks on the file? Two files with identical i-nodes are only actually hardlinks to the same file if the number of links is equal and > 1? Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Project Co-Leader cygwin AT cygwin DOT com Red Hat -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
