Re: isprint(3)

[Joerg Sonnenberger <joerg@xxxxxxxxxxxxxxxxx>]

On Mon, May 02, 2005 at 07:24:50PM +0200, Joerg Sonnenberger wrote:
> On Mon, May 02, 2005 at 09:50:57AM -0700, Matthew Dillon wrote:
> > 
> > :I see the problem. isprint is currently true for:
> > :isalnum() || ispunc() || isblank()
> > :
> > :\t is in isblank(), but it is also a control char. I have to investigate
> > :whether it is better to not include blank (only space), split the define
> > :into two parts are explicitly mask off cntrl. The first should be allowed
> > :by SUS for the POSIX locale, but I have to check it carefully.
> > :
> > :Joerg
> > 
> >     It seems to me that the most generic solution is to mask off cntrl.
> 
> That means an and more and possible double evalution of the
> argument. Actually, it is allowed by SUS to have a control character
> as printable. Question is, are there non-tab non-space white spaces, which are
> control chars and not graphs?

After reading SUS once again, I think we should go with including only
space in print. All control-chars should be part of blank instead.

Joerg