April 2008 - Boost-users - lists.boost.org

Unordered not 1.35.0?
by Andreas Klöckner 10 Apr '08

10 Apr '08

Hi all, I just updated to 1.35.0 from an old-ish svn snapshot, only to learn that boost/unordered_{map,set}.hpp are not in the release. I thought unordered had gotten accepted *way* before 1.35? Thanks, Andreas

3 2

any advantage of using property_map over std::map
by Abir Basak 10 Apr '08

10 Apr '08

Though it is a stupid question, is there any advantage i will get if i use boost property_map instead of std::map for my name value property pairs ? I am working for a statistical system, where i have certain parameter names (20-50 names at max) mostly fixed defined in terms of compile time constants (enums) but sometimes are from a variable set, and values which changes during adaptation (a small statistical distribution struct contains mean & sigma usually). so far i have used things like, enum height{ class1, class2,class3}; struct dist{ float mean; float sigma; }; typedef std::map<height,dist> height_prop; ( i also thought of having a compile time lookup based on a lookup table, but not sure if that is worth esp when search is among small no of elements) and similarly other properties. so will i get any benefit using boost::property_map instead of std::map (or unordered_map ) . in future i may want these properties to be associated with a Boost graph node & edge. thanks abir

1 0

[1.35] Weird behavior while building bjam under linuxx86_64 platform.
by Lian Cheng 10 Apr '08

10 Apr '08

Hi, all I'm trying to build boost 1.35 within a RedHat x86-84 box. The gcc version information is here: $ gcc -v Reading specs from /usr/lib/gcc/x86_64-redhat-linux/3.4.5/specs Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-java-awt=gtk --host=x86_64-redhat-linux Thread model: posix gcc version 3.4.5 20051201 (Red Hat 3.4.5-2) $gcc --version gcc (GCC) 3.4.5 20051201 (Red Hat 3.4.5-2) Copyright (C) 2004 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. While building bjam by invoking $BOOST_ROOT/tools/jam/build_dist.sh, I got error messages as following: ./bootstrap/jam0 -f build.jam --toolset=gcc --toolset-root= clean ...found 1 target... ...updating 1 target... ...updated 1 target... ./bootstrap/jam0 -f build.jam --toolset=gcc --toolset-root= ...found 47 targets... ...updating 2 targets... [MKDIR] :bin.linuxx86_64 <--- *NOTICE THIS LINE* [COMPILE] :bin.linuxx86_64/bjam filemac.c:115:5: warning: C++ style comments are not allowed in ISO C90 filemac.c:115:5: warning: (this will be reported only once per input file) /tmp/ccjqZ5Qv.o(.text+0x903): In function `var_expand': : undefined reference to `path_parse' /tmp/ccjqZ5Qv.o(.text+0x9ee): In function `var_expand': : undefined reference to `path_build' /tmp/ccjqZ5Qv.o(.text+0xa1d): In function `var_expand': : undefined reference to `path_parent' [...] You see, the bin directory name is ":bin.linuxx86_64" rather than "bin.linuxx86_64". Weird? I've build boost 1.35 successfully under cygwin, and the bin directory name is exactly "bin.cygwinx86". I think it is this wrong directory name that causes following linkage errors. Is it a bug or I missed something obvious in the documentation? Any suggestion is appreciated! -- Cheers Cheng

1 0

Boost.Exception
by Olaf Peter 10 Apr '08

10 Apr '08

with interest I did follow the discussion about boost.exeception some times ago. At http://www.revergestudios.com/boost-exception/boost-exception.html I read: Note: Boost Exception was accepted as a Boost library on November 7, 2007; however it has not yet been part of an official Boost release. Why is it not in release 1.35?? Is it safe to use it out of the boost distribution (concerning API changes). Thanks, Olaf

1 0

boost.variant as argument
by Olaf Peter 10 Apr '08

10 Apr '08

Hi, I've wrapped a variant<double,bool> into a tuple. Unpacking the tuple works as far I can see. I want to print the tuple for doubles with one pretty print functor and for booleans with another functor (not shown here). Anyway, variant_print takes a long argument list as the error message shows. How can I pass it convinient? I guess, inside the variant_print I have to use the vistor to dispatch the print functors, isn't? BTW, fusion could be used? On debian/linux 1.35 is not released yet, so I use 1.34.1. Thanks, Olaf ---8<--- #include <boost/variant.hpp> #include <boost/tuple/tuple.hpp> #include <boost/tuple/tuple_io.hpp> #include <iostream> struct pretty_print { pretty_print( double d ) : m_d( d ) { } std::ostream& print_on( std::ostream& os ) const { os << m_d; return os; } double m_d; }; std::ostream& operator<<( std::ostream& os, const pretty_print& p ) { return p.print_on( os ); } struct variant_print { // ????????? variant_print(::boost::variant<double, bool>& v) { } }; using namespace std; using namespace boost::tuples; using namespace boost; int main() { typedef variant<double, bool> double_variant; typedef tuple<double_variant> double_tuple; double_tuple v1; double_tuple v2; get<0>( v1 ) = 3.14; get<0>( v2 ) = true; // This works cout << pretty_print( get<double>( get<0>( v1 ) ) ) << endl; // This not cout << variant_print( get<0>( v1 ) ) << endl; } --->8--- $ LANG=en g++ -Wall tuple.cpp -o tuple && ./tuple tuple.cpp: In function 'int main()': tuple.cpp:55: error: no match for 'operator<<' in 'std::cout << variant_print(((boost::variant<double, bool, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_>&)((boost::variant<double, bool, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_>*)boost::tuples::get [with int N = 0, HT = boost::variant<double, bool, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_>, TT = boost::tuples::null_type](((boost::tuples::cons<boost::variant<double, bool, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_>, boost::tuples::null_type>&)(& v1.boost::tuples::tuple<boost::variant<double, bool, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_>, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type, boost::tuples::null_type>::<anonymous>))))))' ...

2 4

i18n
by Olaf Peter 09 Apr '08

09 Apr '08

Hi, what would be the best approach for a new software project running on unix/win/osx regards to i18n? As may point of view there are gettext using with boost.format or gnu::autosprintf. What would be the adavantages/disadvantages of these? Any experience, other possibilities? Thanks, Olaf

4 3

[asio] Implementation of multi-threaded support
by Yang Zhang 09 Apr '08

09 Apr '08

Hi - at least on Linux, how does boost asio implement multithreaded support for io::run? Is there a single thread blocking on select() (which I believe isn't thread-safe given the same sets), which others interrupt (EINTR) via pipes? Thanks for any explanation!

1 0

[boost-users] Windows Installer for Boost 1.35.0
by Klaim 09 Apr '08

09 Apr '08

Hello, I'd like to know if an installer for boost 1.35.0 is planned? Currently only 1.34.1 and previous versions of the installer are provided on the boost-consulting website.

1 0

[python] howto expose iterator / range with parameter?
by Jörg Becker 09 Apr '08

09 Apr '08

Hello, I want to expose a begin and end method which takes one argument and returns an iterator as a python iterator. Here is an example: struct range_test { typedef std::list< int > my_type; typedef my_type::iterator iterator; my_type m_vals; my_type::iterator begin ( int i ) { return m_vals.begin(); } my_type::iterator end ( int i ) { return m_vals.end();} }; class_< range_test >( "range_test" ) .def( "get_iter", range( &range_test::begin, &range_test::end ) ) ; If I import this, python exists immediately without any error message. Trying it with wrappers, I get a (gcc) compile error saying my wrapper function is not a class, struct, or union type: range_test::iterator begin( range_test & t, int i ) { return t.begin( i ); } range_test::iterator end( range_test & t, int i ) { return t.end( i ); } class_< range_test >( "range_test" ) .def( "get_iter", range( &range_test::begin, &range_test::end ) ) .def( "iter_wrapper", range( begin, end ) ) ; So I'm wondering how to expose such iterators, e.g. a method taking a predicate and returning a filter_iterator. Thanks for any help Jörg

1 0

[serialzation] Portable binary archives
by Daniel Oberhoff 09 Apr '08

09 Apr '08

Hi, I was just wondering if there is any chance of having portable binary archives soon again, or to which version of boost I have to revert to use those in the vault. I am just asking as I want to javatize some of my work for another project, and I would like to use boost binary archives for portable serialization (portability of the data is really vital for this project). Best Daniel

4 12