cp: odd behaviour; does not preserve symlink

[Cristian Ionescu-Idbohrn]

On Tue, 9 Jun 2009, Denys Vlasenko wrote:

> Yes. It's logical. cp *copies files*. IOW: it *creates a copy
> of an existing file*. Copy of a file should be a file.
>
> In this light,
>
> cp file symlink - should not write into linked file
> cp file device  - should not send file's bytes into the device

So, what you're saying is that you're in disagreement with gnu & co. Is
that correct?

> but both should either refuse to copy or delete 2nd param,
> and create *an ordinary file*.

Right.  But, that's not real life.
Don't get me wrong.  I don't disagree with you.

> Apart from that, "cp file symlink" is a security risk.
> Think about this:
>
> cp /backup/home/joe/dissertation.htm /home/joe
>
> What if malicious Joe created /home/joe/dissertation.htm symlink
> pointing to /etc/shadow? Or to /dev/sda1?

But of course.  Symlink attacks were not discovered yesterday.  We keep
getting daily (security) tips on how to keep away from them.  Still...

> I know that POSIX and friends do not do that. I do not know
> why they chose to do stupid things and have security risks
> instead of prescribing that cp is a copy operation.

Ok.  That's fine with me.  What you're saying is the busybox behaviour is
intentional.  True?

> If you want to dump bytes into an arbitrary entry in a directory,
> the natural way is "cat >dest".

Yes.
Now how do we preach the gospel to the non-believers?  POSIX and gnu
folks?  My collegues.  Are Denys and Cristian the true prophets? ;-)


Cheers,

-- 
Cristian