R devel - Sep 2006 - (PR#9201) Unable to save a plot containing Chinese (two-byte)

This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--27464147-689190176-1157401193=:15799
Content-Type: TEXT/PLAIN; CHARSET=ISO-8859-1
Content-Transfer-Encoding: 8BIT
Content-ID: <Pine.LNX.4.64.0609042121261.15799 at gannet.stats.ox.ac.uk>

On Mon, 4 Sep 2006, Pavel Stranak wrote:
> On 4.9.2006, at 21:16, Prof Brian Ripley wrote:
> 
> >Please can we have a full reproducible example, including the locales
> >used, and exactly how you 'saved' the plot?  (As we do ask.)
You have _still_ not told us what we asked, including the locales!
> I have used GUI command "File->Save As .." Im Mac GUI.
That may well not work, but reports on the Mac GUI are inappropriate for 
R-bugs.
> It works perfectly with the same plots, when I don't use Chinese.
> It also works fine, when I use Czech accented characters. At least some of
> them are also two-byte in UTF-8.
Yes, but Chinese chars are not.
> But the problem also occures when I direct pdf device right into the file:
> > pdf(file="Rplot.pdf")
Incorrect usage.
> > plot(1:20, 1:20, main="P??li? ?lu?ou?k? k?? ?p?l ??belsk?
?dy.")
> # Czech accents work fine
> > plot(1:20, 1:20, main="?P??li? ?lu?ou?k? k?? ?p?l ??belsk?
?dy.")
> Error in title(...) : conversion failure in 'mbcsToSbcs'
> # One Chinese character => Error
> 
> >
> >If Chinese characters are 'two-byte', then this likely will not
work (but
> >at least European locales on MacOS X are UTF-8, so I expected Chinese
ones
> >to be also).
> As far as I know, UTF-8 contains at least 1byte and 2byte characters.
and 3 byte and 4 byte chars, and potentially up to 6.
> All the chinese characters are represented by two bytes in UTF-8.
Please give your reference!  They are in the Unicode ranges from 2F00, and 
characters above 07Ff need 3 or more bytes in UTF-8.

http://www.cl.cam.ac.uk/~mgk25/unicode.html

> The characters I have used (but that fail to display in your bug-tracking
> system)
> are two characters "Chinese language".
> 
> My locales are set to UTF-8.
That is *not* a locale but a character encoding.
> I have not set any fonts in R because all the Chinese characters display
fine
> both in the console and the plot.
But not in pdf, as you said.  You really have failed to read the 
references, including the help page and tutorial I very kindly pointed you 
to.

In case it has still not got through to you: PDF is not written in UTF-8 
and most fonts do not include Chinese ideographs, including Helvetica, the 
default font for PDF.

-- 
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
--27464147-689190176-1157401193=:15799--