- Boost-users - lists.boost.org

regex traits for multibyte strings
by Ben Hutchings 09 Sep '03

09 Sep '03

Is it possible to define a traits class for regex that supports multibyte strings? I am intending to do so, but I'm unsure that it is possible because there are several required functions that take a parameter of type char_type, which wouldn't really hold a complete character.

2 1

aCC: HP ANSI C++ B3910B A.03.37 and boost
by Markus Werle 09 Sep '03

09 Sep '03

Hi! Before I start the tedious task of a reconfigure: Anyone with experience on how to handle that compiler? I need the -AA option, so the compiler gets really picky about _everything_: Warning (anachronism) 823: "/opt/FREE-SOFTWARE/boost-1.30.0/boost/mpl/if.hpp", line 172 # Redundant preprocessing concatenation operation results in two valid preprocessing tokens. Depending on this implementation defined behavior will result in non-portable code. BOOST_MPL_AUX_VOID_SPEC(3, if_) Any help appreciated. aCC has some problems with e.g. this code snippet, which passed intel's 7.1: What's wrong here? typedef typename boost::mpl::apply_if_c< SAME_TYPE(T, Simpler1), boost::mpl::identity<pair<T, Overloader<0> > >, boost::mpl::apply_if_c< (SAME_TYPE(Simpler1, Simpler2) || SAME_TYPE(Simpler1, SimplerN)), boost::mpl::identity<pair<Simpler1, Overloader<1> > >, boost::mpl::apply_if_c< SAME_TYPE(Simpler2, SimplerN), boost::mpl::identity<pair<Simpler2, Overloader<2> > >, boost::mpl::identity<pair<SimplerN, Overloader<3> > > > > > ::type Decision; Error 24: "../daixtrose/src/daixtrose/Simplify.h", line 164 # ':' expected instead of '::'. > ::type Decision; ^^ Error 585: "../daixtrose/src/daixtrose/Simplify.h", line 164 # Expected an operator before 'Decision'. > ::type Decision; ^^^^^^^^ Error 172: "../daixtrose/src/daixtrose/Simplify.h", line 164 # Undeclared variable 'type'. > ::type Decision; ^^^^ Error 19: "../daixtrose/src/daixtrose/Simplify.h", line 167 # Unexpected 'Decision'. COMPILE_TIME_ASSERT ^^^^^^^^^^^^^^^^^^^ -- Build your own Expression Template Library with Daixtrose! Visit http://daixtrose.sourceforge.net/

2 1

BOOST_MPL_AUX_APPLY2 and aCC
by Markus Werle 09 Sep '03

09 Sep '03

Hi! aCC A.03.37 complains about the macro BOOST_MPL_AUX_APPLY2, why? Any fix at hand (whether I defien BOOST_NO_CONFIG or not has no influence)? aCC -DHAVE_CONFIG_H -I. -I../daixtrose -I. -I../daixtrose/src -I/opt/FREE-SOFTWARE/boost-1.30.0 -I/opt/FREE-SOFTWARE/loki-2003May28 -DBOOST_NO_CONFIG -DBOOST_MPL_NO_PREPROCESSED_HEADERS -AA -DAportable +W823 -D__HPACC_THREAD_SAFE_RB_TREE -DRWSTD_MULTI_THREAD -DRW_MULTI_THREAD -D_REENTRANT -D_THREAD_SAFE -c -o test_simple_get_value_1-main.o `test -f '../daixtrose/src/demos/SimpleGetValue.1/main.C' || echo '../daixtrose/'`../daixtrose/src/demos/SimpleGetValue.1/main.C Error 422: "/opt/FREE-SOFTWARE/boost-1.30.0/boost/mpl/bind.hpp", line 377 # 'apply' is used as a type, but has not been defined as a type. Perhaps you meant 'apply' as in "template <typename U,typename U2,typename U3,typename U4,typename U5> struct apply" ["/opt/FREE-SOFTWARE/boost-1.30.0/boost/mpl/bind.hpp", line 372]. : BOOST_MPL_AUX_APPLY2(F,T,U) ^^^^^^^^^^^^^^^^^^^^ Error 318: "/opt/FREE-SOFTWARE/boost-1.30.0/boost/mpl/bind.hpp", line 377 # A template name was expected instead of '<type inserted by compiler>'. Did you forget to define the template? : BOOST_MPL_AUX_APPLY2(F,T,U) ^^^^^^^^^^^^^^^^^^^^ Error 445: "/opt/FREE-SOFTWARE/boost-1.30.0/boost/mpl/bind.hpp", line 377 # Cannot recover from earlier errors. : BOOST_MPL_AUX_APPLY2(F,T,U) ^^^^^^^^^^^^^^^^^^^^ make: *** [test_simple_get_value_1-main.o] Error 2 Markus -- Build your own Expression Template Library with Daixtrose! Visit http://daixtrose.sourceforge.net/

2 1

win32.hpp has a using std::max, min
by Charles Schwartz 09 Sep '03

09 Sep '03

win32.hpp has using std::min and using std::max directives. (win32.hpp in fact overrides these functions). These directives hinder use of the max function elsewhere in a client program, for example, by #included c libraries, which are expecting to use the global (non overridden) max function. Indeed, Sutter's More Exceptional C++ #40 (p. 236) asserts that using directives/declerations should never appear in any headers (certainly not at file scope!). Will this be removed in future releases?

2 1

Building Boost with Intel compiler 7.0 and STLPort on windows
by Andre du Toit 09 Sep '03

09 Sep '03

Hello. Does anyone know whether it is currently possible to use Boost with the Intel Compiler v7.0 and the latest STLport, in windows? It seems I still have to hack files in date_time/posix_time (specifically time_formatters.hpp), to reduce my compile errors. Like defining BOOST_MSVC and then undeffing it, before line 43 of this file is processed. Otherwise I get ambigious operator definitions. Also, is it possible to actually build the date_time library with the Intel compiler and STLport at the same time? Why is there a msvc-stlport toolset but no intel-stlport toolset? O; another thing. Has anyone perhaps built boost in MSYS before? If MSVC + Intel setup is giving me too much hassle, I am going to try and build boost in MSYS. Greetings, Andre

3 2

lambda and smart pointers
by Rodolfo Lima 08 Sep '03

08 Sep '03

What's the proper way to use lambda with smart pointers? I'm trying to compile the following: class Dados { public: Dados(int _a) : a(_a) {}; int Pega() const { return a; } private: int a; }; int main() { shared_ptr<Dados> pDados(new Dados(5)); cout << (bind(&Dados::Pega, _1)(pDados)==5) << endl; return 0; } I've already tried with *_1 with no success. In this case, the compiler (VC7.1) returns "error C2039: 'iterator_category' : is not a member of 'boost::shared_ptr<T>'". Any hope? Thanks, rod

1 0

Compile errors boost 1.30.2
by Andre du Toit 08 Sep '03

08 Sep '03

Hi. I get the following error when compiling the latest version of boost, using Intel compiler 7.0 in visual c++: ..\..\..\boost\boost/type_traits/is_integral.hpp(39): error: class "boost::is_integral<unsigned short>" has already been defined BOOST_TT_AUX_BOOL_TRAIT_CV_SPEC1(is_integral,wchar_t,true) Now the offending line sits between a "#ifndef BOOST_NO_INTRINSIC_WCHAR_T". The simple fix of defining BOOST_NO_INTRINSIC_WCHAR_T solves this problem, but causes other problems in other locations. All I want is to compile and link against boost::date_time library, using my compiler setup. Thats all. Any suggestions? Greetings, Andre

1 0

RE: [Boost-users] pyste and incomplete types
by Jason.Sibthorpe＠aculab.com 05 Sep '03

05 Sep '03

I have just checked out the 'latest' version from cvs. I notice that the 'problem' has been fixed. Thanks -J -----Original Message----- From: Jason.Sibthorpe(a)aculab.com [mailto:Jason.Sibthorpe@aculab.com] Sent: 05 September 2003 17:21 To: boost-users(a)lists.boost.org Subject: [Boost-users] pyste and incomplete types Hi, I have encountered a problem with pyste and incomplete types. From earlier posts and the doco I think that pyste should generate BOOST_PYTHON_OPAQUE_SPECIALIZED_TYPE_ID(it) But is does not currently (boost boost-1.30.2). invoking bjam without the BOOST_PYTHON_OPAQUE_SPECIALIZED_TYPE_ID(it) results in the output attached (output1) with BOOST_PYTHON_OPAQUE_SPECIALIZED_TYPE_ID(it) results in the output attached (output2) In either case It does not build. Am I misunderstanding something? Thanks Jason //////////////////////////////////////////////////////////////////////////// /// //// incomplete_type.h //////////////////////////////////////////////////////////////////////////// /// typedef struct incomplete_type it; it *func1(void); //////////////////////////////////////////////////////////////////////////// /// //// incomplete_type.pyste //////////////////////////////////////////////////////////////////////////// /// itype = AllFromHeader("incomplete_type.h") set_policy(itype.func1, return_value_policy('return_opaque_pointer')) // Note the stringified 'return_opaque_pointer' //////////////////////////////////////////////////////////////////////////// /// //// itype.cpp - pyste output: //////////////////////////////////////////////////////////////////////////// /// // Includes =================================================================== #include <boost/python.hpp> #include <incomplete_type.h> // Using ====================================================================== using namespace boost::python; // Module ===================================================================== BOOST_PYTHON_MODULE(itype) { def("func1", &func1, return_value_policy< return_opaque_pointer >()); class_< incomplete_type, boost::noncopyable >("incomplete_type", no_init) ; } Jason Sibthorpe Software Engineer Aculab Tel: +44 (0) 1908 273 869 Fax: +44 (0) 1908 273 801 Email: jason.sibthorpe(a)aculab.com Website: <http://www.aculab.com>

1 0

Re: [Boost-Users] ostream_iterator and std::copy - better solutions?
by Victor A. Wagner, Jr. 05 Sep '03

05 Sep '03

is this what you had in mind?? I've considered this from time to time (first time I've formalized it tho) especially thinking of alternatives to the 3 delimiter arguments (substitute one? "{,}" ??) or what-have-you maybe we should boostify it? template <typename InputItr> ostream& range_streamer(ostream& to, InputItr start, InputItr stop, const char* sdelim, const char* mdelim, const char* edelim) { // output the beginning delimiter to << sdelim; // output the 1st element if it exists if (start != stop) to << *start++; // output the rest of the list (if any) separated by mdelim while (start != stop) { to << mdelim << *start; ++start; } //output the ending delimiter and return return to << edelim; } and of course, if you're going to do it, do it all the way: template <typename Container> inline ostream& container_streamer(ostream& to, Container const& cont, const char* sdelim, const char* mdelim, const char* edelim) {return range_streamer(to, cont.begin(), cont.end(), sdelim, mdelim, edelim);} At Wednesday 2003-09-03 09:07, you wrote: >Before rolling my own, does boost have a facility to stream out an iterator >range without the dangling delimiter at the end? For example > > >std::ostream& operator<<( std::ostream& s, const MyClass& aC ) >{ > const std::string lDelim( "," ); > > std::vector<int> lInts; > > lInts.push_back(1); > lInts.push_back(2); > lInts.push_back(3); > > std::ostream_iterator<int> lItr( s, lDelim.c_str() ); > > std::copy( lInts.begin(), lInts.end(), lItr ); >} > >yields: > >1,2,3, > >I'm thinking something like the following usage would be nice: > > s << range_streamer( lInts.begin(), lInts.end(), "{", "," "}" ); > >yielding: > >{1,2,3} > >Any thoughts? > > >-- > >Jeff Flinn >Applied Dynamics, International Victor A. Wagner Jr. http://rudbek.com The five most dangerous words in the English language: "There oughta be a law"

2 2

pyste and incomplete types
by Jason.Sibthorpe＠aculab.com 05 Sep '03

05 Sep '03

Hi, I have encountered a problem with pyste and incomplete types. From earlier posts and the doco I think that pyste should generate BOOST_PYTHON_OPAQUE_SPECIALIZED_TYPE_ID(it) But is does not currently (boost boost-1.30.2). invoking bjam without the BOOST_PYTHON_OPAQUE_SPECIALIZED_TYPE_ID(it) results in the output attached (output1) with BOOST_PYTHON_OPAQUE_SPECIALIZED_TYPE_ID(it) results in the output attached (output2) In either case It does not build. Am I misunderstanding something? Thanks Jason //////////////////////////////////////////////////////////////////////////// /// //// incomplete_type.h //////////////////////////////////////////////////////////////////////////// /// typedef struct incomplete_type it; it *func1(void); //////////////////////////////////////////////////////////////////////////// /// //// incomplete_type.pyste //////////////////////////////////////////////////////////////////////////// /// itype = AllFromHeader("incomplete_type.h") set_policy(itype.func1, return_value_policy('return_opaque_pointer')) // Note the stringified 'return_opaque_pointer' //////////////////////////////////////////////////////////////////////////// /// //// itype.cpp - pyste output: //////////////////////////////////////////////////////////////////////////// /// // Includes =================================================================== #include <boost/python.hpp> #include <incomplete_type.h> // Using ====================================================================== using namespace boost::python; // Module ===================================================================== BOOST_PYTHON_MODULE(itype) { def("func1", &func1, return_value_policy< return_opaque_pointer >()); class_< incomplete_type, boost::noncopyable >("incomplete_type", no_init) ; } Jason Sibthorpe Software Engineer Aculab Tel: +44 (0) 1908 273 869 Fax: +44 (0) 1908 273 801 Email: jason.sibthorpe(a)aculab.com Website: <http://www.aculab.com>

1 0