FreeTDS Development Group

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

On Sat, 19 Jul 2003 19:06:44 -0500, GUTIERREZ Pablo TESIS-TAMSA
<Tesgup AT tamsa.com.mx> wrote:
> I think that "defined (__WATCOMC__)" is necessary because this patches
> are targeted only to QNX 4 (WATCOM COMPILER is the only compiler
> available for QNX4 - QNX4 and Watcom compiler are very linked). 
> 
> QNX 6 (the newest QNX) is completely different. Headers are different
> and it comes with a non-watcom compiler (and it also supports gnu
> compilers).(WATCOM stopped producing QNX compilers). If a QNX6 user
> tries to compile freetds it will match __QNX__ and the compilation will
> fail when trying to use a WATCOM function (such as _bprintf)

Understood.  That's why my tests are formed the way they are:

> #if defined (__QNX__) 
> # if !HAVE_VSNPRINTF && HAVE__VBPRINTF
> #  define vsnprintf _vbprintf

Disregarding the compiler, we substitute _vbprintf for vsnprintf only if
we have one and not the other (and only for QNX).  I'm testing directly
for the existence of the functions; you're using the compiler as a proxy
for the version of QNX.  It's not impossible that Watcom will reënter the
market, or that someone using Watcom would have added vsnprintf(3).  

> What's more for making the patch that I posted minutes ago, I had to
> port gnu diff first, since QNX diff is very limited (no context
> information, and no patch application)

Thank you for going to the trouble.  Your patch applied perfectly.  

> I hope these long paragraphs didn't bore you. ;-)

Not at all.  I learn portability through vicarious experience.  (!)

> For those reasons I wont be able to generate a configure script.

http://www.schemamania.org/projects/freetds/configure

Use it in good health.  

--jkl