On 1/29/07, Roland Mainz <roland.mainz at nrubsig.org> wrote:
> William James wrote:
> > On 1/26/07, David Korn <dgk at research.att.com> wrote:
> > > > How can I convert a character like ? (small u, grave accent) to the
> > > > integer character value of the current locale, i.e. per
> > > > http://saftsack.fs.uni-bayreuth.de/html/HTML_2.0/iso8859-1.html ? is
> > > > 249
> > >
> > > ksh93 supports character constants in integer expressions, thus
> > > $(('`'))
> > > will expand to the character value for grave and $(('u')) will
> > > give the character value for small u. However, character values
> > > can be signed so that
> > > $(('?'))
> > > expands to a signed number. However, you can do
> > > printf "%u\n" $(('?' &0xff ))
> > > to get the unsigned value.
> >
> > That works in the en_US.iso885915 locale but in en_US.utf8 it fails:
> > $ printf "%u\n" $(('1'))
> > 49
> > $ printf "%u\n" $(('?'))
> > ksh: '?': arithmetic syntax error
>
> Umpf... it seems the code uses |char| where it should use |wchar_t| to
> handle locale-specific character values > 127 (AFAIK using |wchar_t| is
> the best choice in this case... David ?).
>
David said he's going to fix that for the next release.
Cheers,
William
--
@,,@ William James
(\--/) williamjamesgnusolaris at gmail.com
(.>__<.) GNU/Solaris hacker
