Re: 16bit ctype table

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Sat, Oct 06, 2012 at 03:56:32PM +0000, YAMAMOTO Takashi wrote:
> hi,
> 
> > On Sat, Oct 06, 2012 at 03:56:25AM +0900, Takehiko NOZAKI wrote:
> >> > I still don't see the point in redefining the ctype stuff. Does it
> >> > *really* break on any platform we care about?
> >> 
> >> if you think so, it is the matter. maybe only in your mind.
> >> please clarify what are you warried about.
> >> 
> >> it is far away from ``technical'' discussion.
> >> i'm getting sick of such uncertainty.
> > 
> > It is quite a technical issue. I don't see how any platform definition
> > of _RUNETYPE_A and friends can invalidate the definitions from
> > runetype_file.h, since it is pretty much the last thing included in
> > mklocale's sources. Duplicating the definitions in the source is
> > certainly bad.
> 
> are you talking about _CTYPE_FOO and _NB_CTYPE_FOO?
> given the invariant _RUNETYPE_FOO == (_CTYPE_FOO << 8), your version of
> patch has same amount of redundant definitions there, hasn't it?

I'm talking about having two definitions for it in the same file for
non-NetBSD systems. Relationship between rune masks and ctype masks is a
separate question.

Joerg