On 10 May 2009 at 19:14, Theo de Raadt wrote:

> > Hi Damien,
> > 
> > > I think we should fix dd(1).
> > 
> > I like that.
> >
> > Not allowing uppercase violates the principle of least surprise for "M"
> > in particular, but accepting uppercase for all four multipliers is probably
> > most convenient indeed (without documenting it, of course).
> 
> And I disagree *VERY STRONGLY* since it leads people to write scripts
> which are not portable.  In fact, people who used this tomorrow would
> be writing scripts which will not work on OpenBSD 4.5, so there is the
> compatibility answer right there.
> 
> Extensions should not be added except when they are very pervasive or
> required.  The upper case versions are not pervasive, and they are not
> required.
> 
> The example where the M was used was obviously written by someone who
> did not realize that documentation should be accurate.  Why punish the
> rest of us who value portabilitty for their error?
> 
> 

More likely the M example was written by someone that is familiar with 
one of the other variants of dd (such as possibly GNU). A quick check 
of some other systems, including Linux flavors, shows me that while 
they also require the b, c, and w suffixes in lower case, the M suffix 
is expected to be upper case. So perhaps that one (and only that one) 
should be allowed (or required?) in upper case.

Reply via email to