[R] New problem printing °C in plots

[Prof Brian Ripley]

On Wed, 2 Feb 2005, Martin Maechler wrote:

>>>>>> "BDR" == Prof Brian Ripley <ripley at stats.ox.ac.uk>
>>>>>>     on Tue, 1 Feb 2005 23:33:37 +0000 (GMT) writes:
>
>    BDR> That this prints as an octal escape was always the
>    BDR> intention:
>
> excuse me Brian, but "always" is not entirely correct:
> Originally (say 6-8 years ago), the intention was really something like
>
>  `` iso-latin-1 (= iso-8859-1) should work everywhere;
>     we don't care yet for anything else ''

We do say (and have for at least 7 years) that which are valid characters 
was locale-dependent.  So we did not implement your policy, instead using 
isxxxx functions to do the testing, not entirely consistently.

It seems that older systems allowed non-ASCII Latin-1 characters in the
C locale, which is not correct according to the POSIX definition of that 
locale (but I think allowed by ISO C90).

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595