FreeTDS Development Group

["James K. Lowden" <jklowden AT schemamania.org>]

I *think* I incorporated our discussion in code.  Freddy, I used ideas
from your patch, but of course I couldn't apply it.  

I committed the ct-lib univarchar patch separately.  On faith. No way for
me to test.  

The tds UTF-8 unit tests now pass.  

There's just one problem: I've been confusing EILSEQ and EINVAL:

  1.  An invalid multibyte sequence is encountered in the
  input. In this case it sets errno to EILSEQ....
        ...
  3.  An incomplete multibyte sequence is encountered in the
  input, and the input byte sequence terminates after it. In
  this case it sets errno to EINVAL....

I had them reversed in my head.  I've swapped them now, I think.  But I
have to go over the code again to undo my confusion.  

Our "chunking" discussion regards EINVAL: an incomplete character at the
end of the buffer due to giving tds_iconv() an arbitrary chunk to work on
(i.e., one not necessarily respecting character boundaries).  

A bit of s/EILSEQ/EINVAL/g should help.  

--jkl