Re: libpq build problem with <io.h> on MS VC++

Bruce Momjian wrote:
> Bruce Momjian wrote:
> > Tom Lane wrote:
> > > Bruce Momjian <[email protected]> writes:
> > > > The only other option I can think of is to #undef those to defines,
> > > > include io.h, then re-include port.h?  Is that better?
> > > 
> > > How about not #define'ing rename() etc in port.h in the first place?
> > > 
> > > We could put
> > > 	#ifdef WIN32
> > > 	#define rename(x) pgrename(x)
> > > 	#endif
> > > into those very few .c files that need it.  (I'm assuming that
> > > forgetting to include this in a file that calls rename() will yield an
> > > obvious build failure on Windows --- if not then the idea is no good.)
> > 
> > I think they have rename --- it just isn't reliable like ours is.
> 
> One other thing we could do is to do the out-of-order includes only for
> MS VC. That is the only platform that needs it, and let the other
> Win32's do the regular include.

Ah, one thing we have done is to reference everything as pg* and define
it to be the libc function on unix and give a compatibility function on
Win32.  We do that with pgpipe.  That might be our best solution.

Personally I never liked defining _away_ from libc function names
(rename -> pgrename), but I think defining _to_ libc names is fine.

Let me fix this tomorrow.  I always felt we had too many tricks in
adding these functions and was never sure we had it consistent.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  [email protected]               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
      joining column's datatypes do not match