[lug] cp and rm

John Hernandez John.Hernandez at noaa.gov
Wed Aug 1 17:40:45 MDT 2001

Maybe you've discovered a bug in bash.  Or maybe in cp.  If you trace down the bug to a particular program, we'll know who to get upset with.  Did calling /bin/cp directly fix the problem?  If not, perhaps it's not unalias that's failing you, maybe something else is resposible for this.  noclobber bash option?  Just guessing now...

David wrote:
>    Going back to where all this started, I really think RH's move, namely aliasing cp and rm to safer variants, was a fair and good decision. ...
> As already stated, personally I like the -i option, and essentially
> for the reason you give.  However, that is not the starting point of
> this thread; the starting point was the lack of reliability of code
> from one release to another.  Also I accept the "good practice"
> doctrine that you give.  But that, too, avoids facing the issue of
> unreliable code.  In this case I think that I have provided code that
> demonstrates that cp does not perform according to its documentation.
> That is bad, even if there are a hundred ways around the problem.
> To consider the problem, reflect on this code snippet and the result.
> type cp
> unalias cp
> type cp
> cp is aliased to `cp -i'
> cp is /bin/cp
> copying to Archive2
> cp: overwrite `/Archive2/home/bozo/.Xdefaults'?
> Are you now so sure about this?  How about POSSIBLY?
>    One thing we can PROBABLY count on is that the basic set of /bin programs will always be there, and they always do what we intend.
> dajo
> _______________________________________________
> Web Page:  http://lug.boulder.co.us
> Mailing List: http://lists.lug.boulder.co.us/mailman/listinfo/lug


  - John Hernandez - Network Engineer - 303-497-6392 -
 |  National Oceanic and Atmospheric Administration   |
 |  Mailstop R/OM12. 325 Broadway, Boulder, CO 80305  |

More information about the LUG mailing list