On Thu, Sep 20, 2007 at 11:19:35AM -0400, Wayne Topa wrote: > Andrew Sackville-West([EMAIL PROTECTED]) is reported to have said: > > well its interesting that it cares what the parent directory perms are > > (from your other mail in this thread). I would think the .ssh > > directory would be sufficient. I'm not sure where to report the > > problem as well. Things that occur to me: does ssh-copy-id assume that > > /home/$USER has the correct perms without checking? > > It must, as I used ssh-copy-id to send the public key from 5 boxes to > the to 3 accounts on the server, including the one with the bad perms. > It did not throw any errors. > > > > > if so, on what is > > that assumption based and who sets that perm (adduser perhaps?)? > > That's where I see the breakdown. Either ssh-copy-id isn't doing a > > sufficient job of checking or its assumptions are faulty.
Is the bug that ssh-copy-id doesn't check the perms of /home/$USER, or is it that ssh cares what the permissions of /home/$USER are? There may be valid reasons for individuals wanting to have more than one user able to access one user's directory but to also be able to ssh as that user. Sure, you can get around it by creating a separate shared directory, but should ssh enforce this? Doug. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
