- Boost-users - lists.boost.org

[boost] boost::function regression in 1.34.0 wrt -fno-exceptions
by Peter Soetens 13 Nov '07

13 Nov '07

Hi, Upto boost 1.33.x, I could compile my program with the gcc option -fno-exceptions. Since boost 1.34.0, this is no longer possible due to code changes in operator= in boost/function/function_template.hpp The following snippets fail to compile: Line 561-571: operator=(Functor BOOST_FUNCTION_TARGET_FIX(const &) f) { this->clear(); try { this->assign_to(f); } catch (...) { vtable = 0; throw; } return *this; } and lines 589-602: // Assignment from another BOOST_FUNCTION_FUNCTION BOOST_FUNCTION_FUNCTION& operator=(const BOOST_FUNCTION_FUNCTION& f) { if (&f == this) return *this; this->clear(); try { this->assign_to_own(f); } catch (...) { vtable = 0; throw; } return *this; } The 1.33.0 and older versions used swap() instead of this try / catch / throw construct. These versions also used the boost::throw_exception function instead of a plain throw. Any rationale for this decision ? Should some/all boost functions compile exception free as well ? I'm using -fno-exceptions on embedded systems. Peter -- Peter Soetens -- FMTC -- <http://www.fmtc.be>

1 0

building boost.python
by Maik Beckmann 13 Nov '07

13 Nov '07

Hello I want to help those who have problems with using bjam to build boost.python (or other libs). I assume you're - running linux with a proper installed gnu toolchain or - running Windowshave a proper installed mingw incl. g++ and mingw-make32 Copy this Makefile to your boost folder: <Makefile> WITH_LIBS= #WITH_LIBS+= --with-regex #WITH_LIBS+= --with-filesystem #WITH_LIBS+= --with-program_options #WITH_LIBS+= --with-thread #WITH_LIBS+= --with-signals #WITH_LIBS+= --with-test WITH_LIBS+= --with-python ########## BJAM executable # set BJAM_EXE to /path/to/prebuild/bjam(.exe) # #### Windows: got it from http://sourceforge.net/projects/boost/ BJAM_EXE=boost-jam-3.1.15-1-ntx86/bjam.exe # ##### Linux: build myself by $ cd tools/jam/src && sh build.sh # x86 (32-Bit): #BJAM_EXE=tools/jam/src/bin.linuxx86/bjam # x86_86 (64-Bit): #BJAM_EXE=tools/jam/src/bin.linuxx86_64/bjam BJAM:=$(BJAM_EXE) -j2 -d+2 BOOST_VARIANT=release # no need to set PYTHONPATH on Linux, but # if I'm on Windows export PYTHONPATH=C:\Python25 all: $(BJAM) variant=$(BOOST_VARIANT) $(WITH_LIBS) --toolset=gcc stage </Makefile> Change the paths (i.e. PYTHONPATH) to make them point to the right location. Now start a command terminal, cd to the boost folder an run - make or respectively - mingw32-make It works without an user-config.jam file at your $HOME folder. Porting this Makefile to nmake shouldn't be hard, but I don't have access to a recent msvc. Note that this Makefile is simplified from the one I actually use to make it more newbie friendly. If some of you are using Linux and cygwin/MSYS/cli on Windows I can post a more intelligent one. Best Regards, - Maik Beckmann

1 0

Re: [Boost-users] Boost-users Digest, Vol 1449, Issue 1
by Vial, Florent 13 Nov '07

13 Nov '07

>Be very clear Well you don't seem to be clear yourself... Arnold & Richter Cine Technik GmbH & Co. Betriebs KG Sitz: Munchen - Registergericht: Amtsgericht Munchen - Handelsregisternummer: HRA 57918 Personlich haftender Gesellschafter: Arnold & Richter Cine Technik GmbH Sitz: Munchen - Registergericht: Amtsgericht Munchen - Handelsregisternummer: HRB 54477 Geschaftsfuhrer: Prof. Dr. Dejan Ilic; Franz Kraus; Jens Schafer Beiratsvorsitzender: Prof. Dr. Hans-Jorg Bullinger

1 0

[ptr_container] Using bind with ptr_map
by Bill Buklis 13 Nov '07

13 Nov '07

Is there a way to use bind with ptr_map? The following fails to compile (but does work with std::map): typedef boost::ptr_map<int,int> TESTMAP; boost::bind( &TESTMAP::value_type::first, _1 ); This errors with pointer to reference member is illegal (C2634 in VC++ 8). Strangely enough this also generates an internal compiler error in VC++ 8 (along with the proper error messages). But that's another story. -- Bill --

2 1

Boost.Date_Time and quarters, incomplete (or unknown) dates
by Christoph 13 Nov '07

13 Nov '07

Is there special support in Boost.Date_Time to support quarters, or (just) years? Consider the following scenario: I have a customer (or patient) whose (exact) date of birth is not known, only that 'he was born in 1960' or in 'May 1960' (w/o attached information on month or days). Is there another way to handle this (and preserve knowledge about the incomplete information) besides pretending that month (and perhaps even day) is 1 and storing additional flags about the information being not actually present? Best regards Christoph

3 2

[rfc] getting the dataflow library ready for submission
by Stjepan Rajko 13 Nov '07

13 Nov '07

Hello all, I am getting close to having the Dataflow library ready to be submitted for review, and would like to respark some interest. The current version of the library offers a generic layer for dataflow programming that can be applied to various data transport mechanisms, and a Dataflow.Signals support layer for Boost.Signals that provides a number of general-purpose components and easy connectability. In the near future I will try to put together more real-world examples of how this library can be useful, but for now the docs have two possibly interesting examples: * an example of how to provide a support layer for another mechanism, with VTK as the victim. * an example of how to use the Dataflow.Signals layer with Boost.Asio and Boost.Serialization to make a distributed dataflow network. The docs are uploaded to: http://dancinghacker.com/code/dataflow/ Code is available from the boost sandbox (SOC/2007/signals), as well as the boost vault (Dataflow directory). All tests (with 1 exception) pass on GCC 4.0.1/darwin and MSVC8. The docs (and the library to some extent) still need work, but I hope what's out there is good enough to paint a picture of what this library is about and how it might be useful. All feedback is welcome! I will next work on an example that uses Boost.GIL to construct an image processing dataflow network. Best regards, Stjepan

2 2

[Newbie] Help on thread
by Olivier Tournaire 12 Nov '07

12 Nov '07

Hi all, I would like to know if there is a way to write the same piece of code using threads (one for each shared_ptr creation ) : shared_ptr< Profile > p2 ( new Profile(origine+(double)i*width_sur_nb_fois_wdir,rect->GetHeightDirection(),MPPParameters_.ProfilesNbPoints_,MPPParameters_.ProfilesDistancePoints_ ) ); shared_ptr< Profile > p3 (new Profile(origine+(double)(-i)*width_sur_nb_fois_wdir,rect->GetHeightDirection(),MPPParameters_.ProfilesNbPoints_,MPPParameters_.ProfilesDistancePoints_ ) ); rect_profiles.push_back( p2 ); rect_profiles.push_back( p3 ); Regards -- Le temps des cerises reviendra. Dans l'immédiat, c'est le temps des noyaux. Courage.

1 0

undefined reference to `boost::system::get_system_category()'
by Vivek Nallur 12 Nov '07

12 Nov '07

I just moved to using boost_asio_0_3_8 because of issues with asio_0_3_7's macro redefinition of FD_SETSIZE. Now, there's no issue with compiling the source, however while linking I get an error saying: undefined reference to `boost::system::get_system_category()' I see from the source that classes "posix_error_category" and "system_error_category" are defined in error_code.cpp (boost_asio_0_3_8/libs/system/src), but there's no actual library (*.so or *.a) to link to. I'm including /path/to/boost_asio_0_3_8 while compiling, but am not doing anything for linking, since all of the code (well almost, since error_code is a cpp file) is .hpp What am I doing wrong? Should I be including anything while linking? regs Vivek

1 0

boost program options positional arguments in config file?
by Zachary 12 Nov '07

12 Nov '07

I've been using program_options quite happily and successfully for a short while now and would like to take this opportunity to express my gratitude. Is there a way to extract postional args from config file, though? The headers seem to indicate that it's possible include_positional flag, but I haven't been able to find something workable in the documentation or mailing lists. Thanks, Zak

2 2

Problem with shared version of Boost.Regex library
by ancode tran 12 Nov '07

12 Nov '07

Hi all, I compiled a shared version (DLL) of Boost.Regex on Windows2000 Pro with gcc 3.4.5, mingw 3.13, STLport 5.1.4 and IBM's ICU 3.8. Details are as below: 1. Open file "boost/config/platform/win32.hpp", change the line: #if !defined(__GNUC__) && !defined(BOOST_HAS_DECLSPEC) to: #if (!defined(__GNUC__) || defined(__MINGW32__)) && !defined(BOOST_HAS_DECLSPEC) I did this because mingw (at least with my version) supports both __declspec(dllexport) and __declspec(dllimport), and __GNUC__ is also defined. 2. Open file "boost/regex/user.hpp" and uncomment 2 lines: "#define BOOST_HAS_DLL_RUNTIME 1" and "#define BOOST_REGEX_DYN_LINK 1". 3. Edit file "libs/regex/build/gcc-shared.mak" and then use it to build Boost.Regex from MSYS shell with GNU Make. After the building process completed, I wrote a small program to test the output. It ran nicely but I got the following warning when linking against Boost.Regex: ../../include/boost/regex/pending/static_mutex.hpp:116: warning: 'bool boost::scoped_static_mutex_lock::locked() const' defined locally after being referenced with dllimport linkage It seems my compiler doesn't like to see "inline" and "__declspec(dllimport)" in the same place. So I have some questions: 1. Should I remove "inline" and rebuild Boost.Regex to skip this warning ? 2. Should I continue to use my own shared version of Boost.Regex ? May it result in such thing like undefined behavior, especially in multithread world ? Can anyone give me a hint, please ? Thanks in advance. P.S: STLport and ICU were built as shared libraries, I therefore dont like a static build of Boost.Regex at all. --------------------------------- Thức ăn hữu cơ - thức ăn lành mạnh có lợi cho sức khỏe như thế nào? Khám phá tại Yahoo! Hỏi & Đáp

2 1