- Boost-users - lists.boost.org

[Boost.Test] Compiling on Mac OS X
by Florian Winter 07 Jul '08

07 Jul '08

Hi, We have run into problems trying to compile Boost.Test on a Mac OS X 10.3 system using GCC version 3.3 (details below). We are using the "included" implementation of Boost.Test in a shared library that we build using our own build system. This works fine on various different platforms and compilers, however part of Boost.Test fails to compile on Mac OS X. We are including Boost.Test in our shared library in the following way: // Define this to declare all Boost.Test symbols as __declspec(dllexport) // rather than __declspec(dllimport). #define BOOST_TEST_SOURCE #define BOOST_TEST_DYN_LINK #include <boost/test/included/unit_test.hpp> The compiler errors (see below) indicate that some symbols related to signal handling are undefined. Either, some required system headers are not included, or these symbols are not portable and not available on Mac OS X (10.3). Has this problem been encountered before? Is there an easy workaround for it, such as defining certain symbols before including the Boost.Test headers? Could it be somehow related to the way we are building Boost.Test? Furthermore, it would be interesting for us to know what the status of Boost.Test on Mac OS X is in general. Unfortunately, the "Portability" section in the documentation http://www.boost.org/doc/libs/1_35_0/libs/test/doc/portability.html as well as the rest of the documentation do not provide a lot of information regarding support for different platforms. Below, you will find the error messages that we got, as well as some information about our system and the version of Boost.Test that we used. Thanks in advance for any help. Best regards, Florian *Boost Version* 1.35.0 *System* MacOSX 10.3 *Compiler* g++ (GCC) 3.3 20030304 (Apple Computer, Inc. build 1666) Copyright (C) 2002 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. *Errors* In file included from <...>/boost/test/included/unit_test.hpp:22, from <...>:18: <...>/boost/test/impl/execution_monitor.ipp: In member function `void boost::detail::system_signal_exception::report() const': <...>/boost/test/impl/execution_monitor.ipp:286: error: `ILL_ILLOPN' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:286: error: (Each undeclared identifier is reported only once for each function it appears in.) <...>/boost/test/impl/execution_monitor.ipp:291: error: `ILL_ILLADR' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:306: error: `ILL_PRVREG' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:311: error: `ILL_COPROC' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:316: error: `ILL_BADSTK' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:326: error: `FPE_INTDIV' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:331: error: `FPE_INTOVF' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:361: error: `FPE_FLTSUB' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:391: error: `BUS_ADRERR' undeclared (first use this function) <...>/boost/test/impl/execution_monitor.ipp:396: error: `BUS_OBJERR' undeclared (first use this function)

2 5

[BGL] why not breakup put_get_helper into three structs?
by Mostafa 06 Jul '08

06 Jul '08

Helllo, I was wondering why put_get_helper shouldn't be three structs (put_helper, get_helper, put_get_helper, with the latter deriving from from the formers) instead of one? My rationale is outlined below along with the proposed code. //Begin Rationale Problem: want a helper method for readable property maps. Solution1: use exist boost::put_get_helper and just never use the put function. Problem with solution 1: 1) the put function will still be "callable", and 2) calling the put function will result in an invalid assignment to a read-only variable compiler error, which is cryptic to decipher, ie, that even though the put function does exist it should not be called. The only way to clarify this would be via documentation, which then brings about the problem of keeping code in sync with comments. Solution2: make a helper method for readable property maps that won't make the put companion function a valid function call, ie, break up boost::put_get_helper to put_helper, get_helper, and put_get_helper which derives from the latter two. Problems with solution 2: None that I can see. Advantages of solution 2: 1) It's self-documenting whether the put or get functions are intended to be called, ie if I derive from get_helper then it's self-evident that I shouldn't be calling the put function (vice versa); and 2) given the above scenario, even if by chance the put function get's called the compiler error message will be easier to decipher, ie, something along the lines of: <no matching function for call to ‘put(...)'>. 3) Not really an advantage, but ... backward compatibility is maintained. //End Rationale. //Proposed code adapted from boost/property_map.hpp // A helper class for constructing a property map // from a class that implements operator[] template <class RetVal, class RvaluePropertyMap> struct get_helper { }; template <class LvaluePropertyMap> struct put_helper { }; template <class RetVal, class LvaluePropertyMap> struct put_get_helper : public put_helper<LvaluePropertyMap>, public get_helper<RetVal, LvaluePropertyMap> { }; template <class PropertyMap, class RetVal, class K> inline RetVal get( get_helper<RetVal, PropertyMap> const & pa, K const & k) { RetVal v = static_cast<const PropertyMap&>(pa)[k]; return v; } template <class PropertyMap, class K, class V> inline void put( put_helper<PropertyMap> const & pa, K k, V const & v) { static_cast<const PropertyMap&>(pa)[k] = v; } Thanks in advance, -Mostafa

1 0

newbie issues on Solaris 9
by PAASILA RISTO 06 Jul '08

06 Jul '08

Hello, I am trying to build the filesystem and regex libraries, but have a problem with the filesystem library. Refer to the log below. In short, I get errors like "libs/filesystem/src/operations.cpp", line 1034: Error: Cannot use std::pair<boost::system::error_code, int> to initialize std::pair<boost::system::error_code, unsigned long long>. and "libs/filesystem/src/path.cpp", line 46: Error: Could not find a match for std::use_facet<std::Facet>(std::locale) needed in::converter(). Is there any configure option, or setting in user-config.jam which will solve the compilation problems? My configuration is SunOS aals176 5.9 Generic_122300-13 sun4u sparc SUNW,Sun-Fire-V440 The compiler version is CC: Sun C++ 5.8 Patch 121017-07 2006/11/02 =========== cd boost_dir; ./configure --with-toolset=sun --with-libraries=filesystem,regex --prefix=$OLDPWD cd "/extdisk1/customer/sdeaptr2/lib/opensource/boost/" #HOST_P#cd boost_dir; ./configure --with-toolset=sun --with-libraries=filesystem,regex --prefix=$OLDPWD -n Building Boost.Jam with toolset sun... tools/jam/src/bin.solaris/bjam -n Unicode/ICU support for Boost.Regex?... not found. Backing up existing Boost.Build configuration in user-config.jam.1 Generating Boost.Build configuration in user-config.jam... Generating Makefile... =========== gmake -C boost_dir install ====== cd "/extdisk1/customer/sdeaptr2/lib/opensource/boost/" 2>/dev/null || mkdir -p "/extdisk1/customer/sdeaptr2/lib/opensource/boost/"; cd "/extdisk1/customer/sdeaptr2/lib/opensource/boost/" #HOST_P#gmake -C boost_dir install ./tools/jam/src/bin.solaris/bjam --user-config=user-config.jam --prefix=/extdisk1/customer/sdeaptr2/lib/opensource/boost --exec-prefix=/extdisk1/customer/sdeaptr2/lib/opensource/boost --libdir=/extdisk1/customer/sdeaptr2/lib/opensource/boost/lib --includedir=/extdisk1/customer/sdeaptr2/lib/opensource/boost/include --with-filesystem --with-regex install Building Boost.Regex with the optional Unicode/ICU support disabled. Please refer to the Boost.Regex documentation for more information (don't panic: this is a strictly optional feature). ...patience... ...found 11247 targets... ...updating 51 targets... sun.compile.c++ bin.v2/libs/system/build/sun/release/threading-multi/error_code.o CC: Warning: -xarch=native has been explicitly specified, or implicitly specified by a macro option, -xarch=native on this architecture implies -xarch=v8plusb which generates code that does not run on pre UltraSPARC III processors sun.link.dll bin.v2/libs/system/build/sun/release/threading-multi/libboost_system-sw- mt-1_35.so.1.35.0 common.copy /extdisk1/customer/sdeaptr2/lib/opensource/boost/lib/libboost_system-sw- mt-1_35.so.1.35.0 common.hard-link /extdisk1/customer/sdeaptr2/lib/opensource/boost/lib/libboost_system-sw- mt.so sun.compile.c++ bin.v2/libs/filesystem/build/sun/release/threading-multi/operations.o CC: Warning: -xarch=native has been explicitly specified, or implicitly specified by a macro option, -xarch=native on this architecture implies -xarch=v8plusb which generates code that does not run on pre UltraSPARC III processors "libs/filesystem/src/operations.cpp", line 1034: Error: Cannot use std::pair<boost::system::error_code, int> to initialize std::pair<boost::system::error_code, unsigned long long>. "libs/filesystem/src/operations.cpp", line 1036: Error: Cannot use std::pair<boost::system::error_code, int> to initialize std::pair<boost::system::error_code, unsigned long long>. "libs/filesystem/src/operations.cpp", line 1070: Error: Cannot use std::pair<boost::system::error_code, int> to initialize std::pair<boost::system::error_code, long>. 3 Error(s) detected. "CC" -fast -mt -KPIC -DBOOST_ALL_DYN_LINK=1 -DBOOST_ALL_NO_LIB=1 -DBOOST_SYSTEM_DYN_LINK=1 -DNDEBUG -I"." -c -o "bin.v2/libs/filesystem/build/sun/release/threading-multi/operations.o" "libs/filesystem/src/operations.cpp" ...failed sun.compile.c++ bin.v2/libs/filesystem/build/sun/release/threading-multi/operations.o... sun.compile.c++ bin.v2/libs/filesystem/build/sun/release/threading-multi/path.o CC: Warning: -xarch=native has been explicitly specified, or implicitly specified by a macro option, -xarch=native on this architecture implies -xarch=v8plusb which generates code that does not run on pre UltraSPARC III processors "libs/filesystem/src/path.cpp", line 46: Error: Could not find a match for std::use_facet<std::Facet>(std::locale) needed in::converter(). "libs/filesystem/src/path.cpp", line 63: Error: Could not find a match for std::use_facet<std::Facet>(std::locale) needed in static boost::filesystem::wpath_traits::imbue(const std::locale&, const std::nothrow_t&). 2 Error(s) detected. "CC" -fast -mt -KPIC -DBOOST_ALL_DYN_LINK=1 -DBOOST_ALL_NO_LIB=1 -DBOOST_SYSTEM_DYN_LINK=1 -DNDEBUG -I"." -c -o "bin.v2/libs/filesystem/build/sun/release/threading-multi/path.o" "libs/filesystem/src/path.cpp" ...failed sun.compile.c++ bin.v2/libs/filesystem/build/sun/release/threading-multi/path.o... sun.compile.c++ bin.v2/libs/filesystem/build/sun/release/threading-multi/portability.o CC: Warning: -xarch=native has been explicitly specified, or implicitly specified by a macro option, -xarch=native on this architecture implies -xarch=v8plusb which generates code that does not run on pre UltraSPARC III processors sun.compile.c++ bin.v2/libs/filesystem/build/sun/release/threading-multi/utf8_codecvt_fa cet.o CC: Warning: -xarch=native has been explicitly specified, or implicitly specified by a macro option, -xarch=native on this architecture implies -xarch=v8plusb which generates code that does not run on pre UltraSPARC III processors ...skipped <pbin.v2/libs/filesystem/build/sun/release/threading-multi>libboost_file system-sw-mt-1_35.so.1.35.0 for lack of <pbin.v2/libs/filesystem/build/sun/release/threading-multi>operations.o. .. ...skipped <p/extdisk1/customer/sdeaptr2/lib/opensource/boost/lib>libboost_filesyst em-sw-mt-1_35.so.1.35.0 for lack of <pbin.v2/libs/filesystem/build/sun/release/threading-multi>libboost_file system-sw-mt-1_35.so.1.35.0... ...skipped <p/extdisk1/customer/sdeaptr2/lib/opensource/boost/lib>libboost_filesyst em-sw-mt.so for lack of <p/extdisk1/customer/sdeaptr2/lib/opensource/boost/lib>libboost_filesyst em-sw-mt-1_35.so.1.35.0... common.copy /extdisk1/customer/sdeaptr2/lib/opensource/boost/lib/libboost_regex-sw-m t-1_35.so.1.35.0 common.hard-link /extdisk1/customer/sdeaptr2/lib/opensource/boost/lib/libboost_regex-sw-m t.so ln-UNIX /extdisk1/customer/sdeaptr2/lib/opensource/boost/lib/libboost_system-sw- mt-1_35.so ...skipped <p/extdisk1/customer/sdeaptr2/lib/opensource/boost/lib>libboost_filesyst em-sw-mt-1_35.so for lack of <p/extdisk1/customer/sdeaptr2/lib/opensource/boost/lib>libboost_filesyst em-sw-mt-1_35.so.1.35.0... ln-UNIX /extdisk1/customer/sdeaptr2/lib/opensource/boost/lib/libboost_regex-sw-m t-1_35.so ...failed updating 2 targets... ...skipped 4 targets... ...updated 45 targets... Not all Boost libraries built properly. ============================================== Thanks, Risto

1 0

[build] Library suffix on mingw/win32
by John Femiani 06 Jul '08

06 Jul '08

I have used boost for some time on windows with msvc, but I am trying to install & use it with mingw-g++. I'm no gcc expert, but it seems that ld requires specific extensions for libraries, and bjam is not generating the right ones. In particular, under mingw, the flag -lboost_unit_test_framework Will match boost_unit_test_framework.lib (the import library for the dll) Or libboost_unit_test_framework.a But NOT libboost_unit_test_framework.lib Which is what bjam produces. Here is an approach that WONT work: Using -llibboost_unit_test_framework Or -llibboost_unit_test_framework.lib Fails because the leading 'lib' is automatically stripped out by ld. As a solution I just renamed all lib*.lib files to lib*.a, but something seems basically wrong with the situation. I don't know WHY bjam used a '.lib' suffix, since I built the libraries using gcc. The static libraries I generate from gcc should be linkable from gcc... I use mingw + msys to build boost as follows: 1) I installed MinGW Automated Installer, MingW GCC 4.3.0, and then MSYS via its installer. - this creates C:\MingW and C:\msys\1.0. The gcc-4.3.0 files are extraced into C:\mingw 2) I open the MSYS shell by running msys.bat 3) I fetch boost/trunk into my home directory using svn. 4) I run tools/jam/build_dist.sh to create bjam 5) I copy bjam.exe to /mingw/bin 6) from within the msys shell, I run bjam --toolset=gcc --layout=system --build-type=complete --prefix=/mingw When I am done, under /mingw/lib I have files with names like: boost_*.dll boost_*.lib libboost_*.lib In particular, the static libraries have a '.lib' suffix. I have googled and searched the boost site, and I can find issues related to mingw/etc but nobody else seems to have _this_ problem. Is there something I am just doing wrong here? BTW: --toolset=mingw failed to find the mingw toolset, but when I build with gcc is seems to know what to do. Thanks, --John Femiani

3 4

gcc 4.3 support planned for Boost 1.36 ?
by Ruediger Berlich 06 Jul '08

06 Jul '08

Hi there, just a short question whether support for gcc 4.3 is planned for Boost 1.36 ? I'm seeing some strange compilation problems with my application (bost with boost 1.35 and the current trunk version) when I use gcc 4.3.1 (on an OpenSUSE 11 box). These are not there when using Boost 1.34.1 or Boost 1.35 with gcc 4.2.1 (on OpenSUSE 10.3). Best, Rüdiger

1 0

New to Boost C++ - Dynamic Arrays
by Jean-Sebastien Stoezel 06 Jul '08

06 Jul '08

Hello, I am interested in supporting dynamic arrays in my C++ application. I read about STL, which somehow brought me to Boost C++... I see Boost C++ provides the Array library, which only seems to support static arrays... Does Boost C++ support dynamic arrays? Any example I could use to understand how this work? By any means, would this dynamic array library be thread safe? Thanks! Jean

4 3

[Boost.MultiIndex] Parameterizing multi_index_container on a template parameter
by Julius Ziegler 06 Jul '08

06 Jul '08

Hello, I am having problems using a multi_index_container in the context of a template class, and I am hoping someone here can help me. This is a condensed example: template<class T> struct C { typedef multi_index_container<T> SomeSet; SomeSet container; // I learned I need typename, since this is an implicit type typedef typename multi_index_container<T>::nth_index<0>::type SomeSetIndex; void f() { SomeSetIndex& index = container.get<0>(); } }; The typedef for SomeSetIndex did not compile. g++-4.2 reported (I stripped something from the message that I think is not relevant): error: non-template 'nth_index' used as template note: use 'multi_index_container<T, [stripped] >::template nth_index' to indicate that it is a template Now I (blindly) tried this declaration: typedef typename multi_index_container<T>::template nth_index<0>::type g++ gets past it, but now reports a syntax error in function f, when index gets initialised. Can someone please explain that first error message to me? Any help is appreciated. -J

3 2

boost with CodeGear rad studio 2007
by Sean Farrow 06 Jul '08

06 Jul '08

Hi: I have the installer from boost-consulting, I would like to use boost with CodeGear rad studio 2007 as well as visual studio 2005, is this possible, how do I go about doing this, and what are the incompatibilies, if any? Cheers Sean. __________ Information from ESET NOD32 Antivirus, version of virus signature database 3239 (20080703) __________ The message was checked by ESET NOD32 Antivirus. http://www.eset.com

2 2

[build] Library suffix on mingw/win32
by John Femiani 05 Jul '08

05 Jul '08

I have used boost for some time on windows with msvc, but I am trying to install & use it with mingw-g++. I'm no gcc expert, but it seems that ld requires specific extensions for libraries, and bjam is not generating the right ones. In particular, under mingw, the flag -lboost_unit_test_framework Will match boost_unit_test_framework.lib (the import library for the dll) Or libboost_unit_test_framework.a But NOT libboost_unit_test_framework.lib Which is what bjam produces. Here is an approach that WONT work: Using -llibboost_unit_test_framework Or -llibboost_unit_test_framework.lib Fails because the leading 'lib' is automatically stripped out by ld. As a solution I just renamed all lib*.lib files to lib*.a, but something seems basically wrong with the situation. I don't know WHY bjam used a '.lib' suffix, since I built the libraries using gcc. The static libraries I generate from gcc should be linkable from gcc... I use mingw + msys to build boost as follows: 1) I installed MinGW Automated Installer, MingW GCC 4.3.0, and then MSYS via its installer. - this creates C:\MingW and C:\msys\1.0. The gcc-4.3.0 files are extraced into C:\mingw 2) I open the MSYS shell by running msys.bat 3) I fetch boost/trunk into my home directory using svn. 4) I run tools/jam/build_dist.sh to create bjam 5) I copy bjam.exe to /mingw/bin 6) from within the msys shell, I run bjam --toolset=gcc --layout=system --build-type=complete --prefix=/mingw When I am done, under /mingw/lib I have files with names like: boost_*.dll boost_*.lib libboost_*.lib In particular, the static libraries have a '.lib' suffix. I have googled and searched the boost site, and I can find issues related to mingw/etc but nobody else seems to have _this_ problem. Is there something I am just doing wrong here? BTW: --toolset=mingw failed to find the mingw toolset, but when I build with gcc is seems to know what to do. Thanks, --John Femiani

1 0

[build] Library suffix on mingw/win32
by John Femiani 05 Jul '08

05 Jul '08

I have used boost for some time on windows with msvc, but I am trying to install & use it with mingw-g++. I'm no gcc expert, but it seems that ld requires specific extensions for libraries, and bjam is not generating the right ones. In particular, under mingw, the flag -lboost_unit_test_framework Will match boost_unit_test_framework.lib (the import library for the dll) Or libboost_unit_test_framework.a (which does not exist after I run bjam!) But NOT libboost_unit_test_framework.lib, Which is what bjam produces. As a solution I just renamed all lib*.lib files to lib*.a, but something seems basically wrong with the situation. I don't know WHY bjam used a '.lib' suffix, since I built the libraries using gcc. The static libraries I generate from gcc should be linkable from gcc... I use mingw + msys to build boost as follows: 1) I installed MinGW Automated Installer, mingw gcc 4.3.0, and then msys via its installer. - this creates C:\MinGW and C:\msys\1.0. The gcc-4.3.0 files are extraced into C:\mingw 2) I open the MSYS shell by running msys.bat 3) I fetch boost/trunk into a temp directory using svn. 4) I run tools/jam/build_dist.sh to create bjam 5) I copy bjam.exe to /mingw/bin 6) from within the msys shell, I run bjam --toolset=gcc --layout=system --build-type=complete --prefix=/mingw When I am done, under /mingw/lib I have files with names like: boost_*.dll boost_*.lib libboost_*.lib In particular, the static libraries have a '.lib' suffix. I have googled and searched the boost site, and I can find issues related to mingw/etc but nobody else seems to have _this_ problem. Is there something I am just doing wrong here? Thanks, --John Femiani

1 0