Re: Adding -l option to cp

[christos%astron.com@localhost (Christos Zoulas)]

In article <20110123225319.GA4566%britannica.bec.de@localhost>,
Joerg Sonnenberger  <joerg%britannica.bec.de@localhost> wrote:
>On Sun, Jan 23, 2011 at 10:35:24PM +0000, Christos Zoulas wrote:
>> In article <00BED867-3DBD-42EE-8773-EB490EB369BB%gmail.com@localhost>,
>> Adam Hamsik  <haaaad%gmail.com@localhost> wrote:
>> >
>> >Just for the record it can be done with pax -r -w -l.
>> 
>> Does it really matter? Don't we have tools that have overlapping 
>> functionality
>> already? It is simple enough to add, and it does not add a lot of maintenance
>> overhead. In terms of userland utility innovation there is solaris at one
>> end of the spectrum and linux at the other. I think it is better to be
>> somewhere in the middle.
>
>The most important point is that the patch doesn't seem to make cp
>behave as advertised. It makes cp not *copy*. So if the goal is to have
>a tree-level link create program, that sounds like a good reason to add
>-R to ln, but not a good reason to add it to cp.

I agree with that, it sounds saner. And like mv uses cp to do the dirty work
cross device, ln can re-use code from cp.

christos