Re: [Boost-users] Re: Q: W4003 with vc7 boost=1.32 in date_time lib

From: Jeff Garland (jeff_at_[hidden])
Date: 2004-12-01 07:57:25

On Tue, 30 Nov 2004 17:49:27 -0800, Eric Niebler wrote
> Dan Dimerman wrote:
> > 
> > The warning reads:
> > C:\Boost\include\boost-1_32\boost\date_time\gregorian\greg_facet.hpp(293) : 
> > warning C4003: not enough actual parameters for macro 'max'
> > 
> 
> This is a date_time bug. Boost has guidelines for avoiding clashes 
> with the min/max macros, which you can read at 
> http://boost.org/more/lib_guide.htm.

Yes, I'm afraid you are right -- we apparently missed this after you made the
other boost min/max changes :-(  Sadly non of our regression procedures caught
this and if your primary test platform isn't Windows it's an easy mistake to
make...
 
> Requiring users to #define NOMINMAX is not satisfatory. Sadly, there 
> are 3rd party libraries, and even platform headers, that depend on 
> those macros. To play well in the real world, Boost headers should 
> compile cleanly even in the presence of the min/max macros.

.rant on
While I understand your point, I think we should push for reform in Redmond. 
99.9% of the folks at MS know better than this, but we (the users) always have
to deal with the 0.1 percent that do stupid stuff like this.  Same thing goes
with the DLL definitions.  If MS wanted to, they could easily eliminate the
requirement that classes to be exported be written with macros (eg: class
SOME_MS_MACRO_MAGIC foo) and just fix the tools to allow external definitions
of dll linking.  I'm sick of wasting my time dealing with non-standard MS
requirements.  So I say, submit your bug reports to Wa.
.rant off

Jeff
ps: I'm not an MS hater, except when they waste my time...