FreeTDS Development Group

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

On 30 Mar 2003 21:43:22 +0200, Frediano Ziglio <freddyz77 AT tin.it> wrote:
> Il dom, 2003-03-30 alle 21:32, James K. Lowden ha scritto:
> > read.c::tds_get_string() contains what looks like a silly loop:
> > 
> >     char temp[256];
> >     char *p, *pend;
> >     /* ... */
> >     while (string_len > 0 && p < pend) {
> >             in_left = string_len > (sizeof(temp) / 2) ? 
> >                     (sizeof(temp) / 2) : string_len;
> >             tds_get_n(tds, temp, in_left * 2);
> >             p += tds7_unicode2ascii(tds, temp, in_left, p, pend - p);
> >             string_len -= in_left;
> >     }
> >     return p - dest;
> > 
> > Basically, if the input string is more than 128 characters, we call
> > tds_get_n() and tds7_unicode2ascii() in 128-character chunks.  
> > 
> > Why do we do this?  I can't see any reason.  
> 
> Cause you don't use malloc?
> tds_get_string can't return a "I'm sorry but your computer is out of
> memory" results, so avoiding malloc fix the problem.

Thanks, I see.  I think I misread the logic.  temp is the
"preconversion" buffer, the place where the UCS-2 data are parked
before converting them to ASCII.  It has to have a size, and there's
no advantage to allocating dynamically.  

--jkl