Boost logo

Boost :

From: Eric Niebler (eric_at_[hidden])
Date: 2004-06-30 19:22:09


Martin Wille wrote:
> Doug Gregor wrote:
>
>> The new handling of min/max seems to break under GCC 2.95.3 with
>> STLport. I'm not quite sure _why_ it doesn't work (I don't have the
>> compiler around), but you can see the failure here:
>>
>> http://tinyurl.com/ytdlf
>>
>> Perhaps someone with that compiler/lib combination could get
>> preprocessed output for the problem area so we could see what's
>> happening?
>
>
> Maybe this experiment sheds some light on the problem:
> I tried this code with gcc-2.95.3 and STLport:
>
> #include <boost/minmax.hpp>
>
> int foo()
> {
> BOOST_USING_STD_MIN();
> return min BOOST_PREVENT_MACRO_SUBSTITUTION(3, 4);
> }
>
> namespace notboost
> {
> template <typename T>
> T const &
> proxy_min(T const &p, T const &q)
> {
> BOOST_USING_STD_MIN();
> return min BOOST_PREVENT_MACRO_SUBSTITUTION(p, q); // line 16
> }
> }
>
> int bar()
> {
> return notboost::proxy_min(3, 4); // line 22
> }
>
>
> namespace test
> {
> int baz()
> {
> BOOST_USING_STD_MIN();
> return min BOOST_PREVENT_MACRO_SUBSTITUTION(3, 4);
> }
> }
>
> int main() { foo(); bar(); test::baz(); }
>
>
> The compiler emits these error messages:
> min.cpp: In function `const int & notboost::proxy_min<int>(const int &,
> const int &)':
> min.cpp:22: instantiated from here
> min.cpp:16: `min' undeclared (first use this function)
> min.cpp:16: (Each undeclared identifier is reported only once
> min.cpp:16: for each function it appears in.)
>
>
> So it looks like gcc 2.95.3 doesn't handle the using declaration
> inside template functions properly.
>
>
> HTH,
> m

Ugh. Thanks Martyn. Can you tell me if there is something funny about
the way STLPort defines std::min/max? Does it use macros? Or are they
defined globally and imported into the std:: namespace with using
declarations?

In the worst case, I suppose for this compiler/library combination we
can import std::min/max into the boost namespace with a using
declaration and make BOOST_USING_STD_MIN() expand to nothing. Martyn,
can you try that and see if it works?

-- 
Eric Niebler
Boost Consulting
www.boost-consulting.com

Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk