FreeTDS Development Group

[Frediano Ziglio <freddyz77 AT tin.it>]

> > - test variable-byte (utf8/big5)
> 
> Yes.  Except I don't know if we have to handle stateful encodings.  If your
> algorithm is ready for prime time, OK.  
> 

Perhaps my algorithm is perfect. However it's only in pseudo-language...

> > - test all string length computation and rewrite some code (many code
> > assume fixed character conversions)
> 
> Much also assumes the first zero byte terminates a string...  And many OS's
> don't provide a wide-character equivalent of strlen() et. al.  Makes me wish
> for C++ and STL and basic_string<>, but I think we're going to get away with
> pointer-to-buffer + count-of-bytes.  
> 

We can use this method
- hold length in bytes
- is fixed size (min == max) ? len = bytes / min
- is utf8 ? use formula
- convert string to ucs4 (using a temporary buffer) and count converted
characters...


> > - SQLGetInfo
> 
> All of SQLGetInfo?  Or just just what's in TODO?
> 

All in TODO. Is not difficult to implement SQLGetInfo (just return what
odbc driver support).

freddy77