2010/3/19 Choongyeol Yoo <spcvar@gmail.com>
Hi.
I made ECL 10.3.1 installed on windows xp and linux.
To test CJK characters with ECL, I tried simple test like  (princ "ひらかな").
On linux build, simply it succeeded.
But, on windows build, built with msvc 2008, unicode enabled,  it was failed as shown below.
Any solution or suggestion needed to the error.
Thanks for this great implementation.

I have identified a couple of problems associated to Unicode, but I can not debug completely the one you mention.

The terminal is opened using a simple encoding :LATIN-1. It has to be changed to a diferent encoding in order to allow I/O of asiatic characters

(EXT:STREAM-EXTERNAL-FORMAT-SET *standard-input* ....)
(EXT:STREAM-EXTERNAL-FORMAT-SET *standard-output* ....)

where the "..." is a keyword naming the actual locale that you suspect is the correct one.

I am unsure about the appropriate choice of encoding for a Windows terminal. In older experiments it seemed to me that Windows does not use UTF-8. It would be nice if you could help us identify whether this is the case using some simple program, such as

(loop for i = (read-byte *standard-input*) do (print i))

and seeing whether your CJK characters are actually UTF-8 encoded in the terminal or not. Then you may experiment with STREAM-EXTERNAL-FORMAT-SET to find out the right encoding among the many that ECL supports.

--
Instituto de Física Fundamental, CSIC
c/ Serrano, 113b, Madrid 28006 (Spain)
http://tream.dreamhosters.com