- Boost-users - lists.boost.org

Re: [Boost-users] unwanted exported symbols in a DLL which statically links to Boost.Library
by John Dlugosz 22 Apr '09

22 Apr '09

> Date: Wed, 22 Apr 2009 15:22:31 +0200 > From: "Luca Cappa" <luca.cappa(a)sequoia.it> > Subject: [Boost-users] unwanted exported symbols in a DLL which > statically links to Boost.Library > To: boost-users(a)lists.boost.org > Message-ID: <op.ussh3tr5r4jy9l(a)vortex.homenet.telecomitalia.it> > Hello, > > I am creating a DLL using the VisualC++ 8sp1, which statically links to > Boost.Library, and also to another static library, which links to > Boost.Library statically too. > > In the export section of the built DLL I see, among the few public > functions which it should export, hundreds of exported functions like > the > ones below. I would really like to export the symbols the DLL defines > as > exported, and not anything else which is unexpectly exported, like the > ones shown below: > I'm not familiar with Boost's libs, but I am good with Windows, and DLL stuff in particular. So maybe I can offer some advice, which may be of help when combined with some information from others. I would agree that if the library containing those functions is static, it should not export anything. It's difficult to cause anything to be exported (nearly impossible for C++ decorated names, IME) other than by decorating the declaration of that function or variable, or the class containing it. On VC++, the decoration appears at the beginning and takes the form __declspec(dllexport) To unravel the macros and conditional compilation, pass one of the files through the preprocessor only. Do this in Visual Studio by opening the properties window for just the one CPP file (not the project itself). Under C/C++, Command Line, add the "/E" option, without the quotes, where you can type Additional Options. Then right-click on the filename in the solution explorer and choose to compile just that file. It won't actually compile, but it will produce another text file (I forget where it puts it). Change it back before Visual Studio gets confused. If "dllimport" is in that text somewhere, it is indeed doing that on purpose. TradeStation Group, Inc. is a publicly-traded holding company (NASDAQ GS: TRAD) of three operating subsidiaries, TradeStation Securities, Inc. (Member NYSE, FINRA, SIPC and NFA), TradeStation Technologies, Inc., a trading software and subscription company, and TradeStation Europe Limited, a United Kingdom, FSA-authorized introducing brokerage firm. None of these companies provides trading or investment advice, recommendations or endorsements of any kind. The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer.

1 0

error while loading shared libraries
by ChristinaDRS 22 Apr '09

22 Apr '09

Greetings all, I first created a class for a library like this: library.h: class Library { public: Library() { }; ~Library() { }; void testLib(); }; library.cpp: #include "library.h" #include <boost/thread/mutex.hpp> #include <boost/thread.hpp> #include <iostream> using namespace std; void Library::testLib() { boost::thread* lt; boost::mutex mu; boost::mutex::scoped_lock lock(mu); cout << "test!" << endl; } I then created the static library like so: g++ -I ~/boost_1_38_0/ -c library.cpp ar rcs libtest.a library.o and then I created a main.cpp like so: #include "library.h" #include <iostream> using namespace std; int main() { Library l; l.testLib(); } now when I try to compile main.cpp like this: $ g++ -I ~/boost_1_38_0/ -I. main.cpp -o main -L/usr/local/lib/ -L. -ltest -lboost_thread-gcc41- now when I run main: $ ./main ./main: error while loading shared libraries: libboost_thread-gcc41-mt-1_38.so.1.38.0: cannot open shared object file: No such file or directory I get that error. Can anyone explain why? -- View this message in context: http://www.nabble.com/error-while-loading-shared-libraries-tp23175557p23175… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

unwanted exported symbols in a DLL which statically links to Boost.Library
by Luca Cappa 22 Apr '09

22 Apr '09

Hello, I am creating a DLL using the VisualC++ 8sp1, which statically links to Boost.Library, and also to another static library, which links to Boost.Library statically too. In the export section of the built DLL I see, among the few public functions which it should export, hundreds of exported functions like the ones below. I would really like to export the symbols the DLL defines as exported, and not anything else which is unexpectly exported, like the ones shown below: boost::archive::detail::oserializer<class boost::archive::xml_oarchive,struct sqy::Vector2<float> >::oserializer<class boost::archive::xml_oarchive,struct sqy::Vector2<float> >(void) boost::archive::detail::oserializer<class boost::archive::xml_oarchive,struct sqy::Vector2<float> >::oserializer<class boost::archive::xml_oarchive,struct sqy::Vector2<float> >(void) boost::archive::detail::oserializer<class boost::archive::xml_oarchive,struct sqy::Vector2<double> >::oserializer<class boost::archive::xml_oarchive,struct sqy::Vector2<double> >(void) In my static library, which statically link to the Boost.Serialization library, I have the following code which explicitly instantiate some template type on a set of a priori known datatypes: template MYLIB_API struct sqy::Vector2<int>; template MYLIB_API struct sqy::Vector2<float>; template MYLIB_API struct sqy::Vector2<double>; template MYLIB_API struct sqy::Vector2<size_t>; #define INSTANTIATE_VECTOR2_SERIALIZATION(x)\ template void MYLIB_API Vector2<x>::serialize<boost::archive::xml_oarchive> (\ boost::archive::xml_oarchive&, const unsigned int);\ template void MYLIB_API Vector2<x>::serialize<boost::archive::xml_iarchive> (\ boost::archive::xml_iarchive&, const unsigned int); INSTANTIATE_VECTOR2_SERIALIZATION(int); INSTANTIATE_VECTOR2_SERIALIZATION(float); INSTANTIATE_VECTOR2_SERIALIZATION(double); INSTANTIATE_VECTOR2_SERIALIZATION(size_t); Notice that, when the DLL is created, MYLIB_API is defined like # define MYLIB_API so i wonder where those symbols above comes from? Anyone could help me to remove from the DLL exporting section those unwanted symbols? Many thanks in advance, Luca

2 1

[multi_array] warning #279 in icc 11.0
by Anton Daneyko 22 Apr '09

22 Apr '09

Hello, I get "warning #279: controlling expression is constant" (full text in attachment). Is there a way to disable that? I tried the following in my source, but it seem to have no effect. #pragma warning ( push ) #pragma warning ( disable: 279 ) #include "boost/multi_array.hpp" #pragma warning( pop ) My icc version: $icc --version icc (ICC) 11.0 20081105 Regards, Anton Daneyko.

2 1

ptime on Win32?
by Del 22 Apr '09

22 Apr '09

Hi, I've been trying to use the boost local_time() function in a win32 program using mingw. After I inserted the call into my program, the program started to crash. So, I wrote a very simple test program: #include <iostream> #include <boost/date_time/posix_time/posix_time.hpp> using namespace boost::posix_time; int WinMain(HINSTANCE,HINSTANCE,LPSTR,int) { ptime time = microsec_clock::local_time();; std::cout << "The current time is " << time << std::endl; return 0; } It compiles and links fine, but when I run the program, I end up with this error message: This application has requested the Runtime to terminate it in an unusual way. Please contact the application's support team for more information. Any ideas?

3 3

Modifying Dijkstra's Shortest Paths algorithm in Boost::Graph to not relax certain edges
by Artyom 22 Apr '09

22 Apr '09

Hi, I am using Boost::Graph and DSP which comes with it to work out shortest paths around my graph. The problem is there is a certain exceptional situation when certain vertices can become unreachable along certain paths. I.e. although there is a valid distance value for edge B ? C, the same edge can become invalid if B is reached along a specific path. If this happens, I want DSP to reject that edge. I attempted to create a list of visitors. EdgeRelaxed records how vertices in the DSP queue were reached. EdgeExamined checks for the exceptional situation and if it's detected, it temporarily changes the distance value of the edge in question to infinity. EdgeNotRelaxed then restores the distance value to it's original. Unfortunately this approach doesn't seem to work, I think the reason for this is because dijkstras_shortest_paths() function gets information about distances from the weight_map which is passed to it. Is there any way I can modify the values inside the weight_map in between DSP stages to invalidate certain paths which will result in them being rejected for relaxation? Or perhaps there is an easier and tidier way of achieving the same result? Thanks in advance, Arty

4 4

Why must a Single Pass Range be default constructible?
by athor 22 Apr '09

22 Apr '09

A Range does not require copy semantics according to the boost docs. http://www.boost.org/doc/libs/1_38_0/libs/range/doc/range.html The SinglePassRangeConcept (and all range concepts using it) requires a range to be default constructible. (There is a line "T a;" in the concept The default constructible requirement means an abstract base class can never implement any Range Concept. If a range does not require any copy semantics, I don't see why it must be default constructible? Are there any good reasons I don't know about? A change from "T a;" to "const T& a" in concepts.hpp would do the trick... Example: struct AbstractArraySlice { int* begin() const = 0; int* end() const = 0; }; // FAILS: BOOST_CONCEPT_ASSERT((SinglePassRangeConcept<AbstractArraySlice>)) Footnote: I often use abstract base classes to verify my algorithms do not use more functionality than necessary and it annoys me that I can't use the range concept check: struct TestSum { const AbstractArraySlice& r; void Test() { CalculateSum(r); } }; template<class T> T CalculateSum(const T& range) <... function requires SinglePassRangeConcept ...> -- View this message in context: http://www.nabble.com/Why-must-a-Single-Pass-Range-be-default-constructible… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

Scoped_array with constructors?
by Space Ship Traveller 21 Apr '09

21 Apr '09

Hi, I want to use scoped_ptr with std::vector or scoped_array, but the problem is that the type I want to allocate has a non-trivial constructor. What is the standard way to make this work? My current best guess is to use std::vector with boost::scoped_ptr in the following way: TextureUnitsController::TextureUnitsController () { GLint textureUnitCount = 1; glGetIntegerv(GL_MAX_TEXTURE_UNITS, &textureUnitCount); m_textureUnits.resize(textureUnitCount); for (unsigned i = 0; i < textureUnitCount; i += 1) { m_textureUnits[i].reset(new TextureUnit(GL_TEXTURE0 + i)); } std::cerr << "Found " << textureUnitCount << " texture units..." << std::endl; } But this doesn't work because scoped_ptr also has a non-trivial constructor, so resize fails. Thanks Samuel

5 5

How to build boost for iPhone?
by Manfred Schwind 21 Apr '09

21 Apr '09

Hi, I tried to build boost for iPhone on Mac OS X. If I try the following ... bjam --prefix=/somePath toolset=darwin target-os=iphone macosx- version=iphone-2.0 define=_LITTLE_ENDIAN link=static install ... the libraries will build fine, but they are i386 code, not arm, so there're unusable. But if I add architecture=arm ... bjam --prefix=/somePath toolset=darwin architecture=arm target- os=iphone macosx-version=iphone-2.0 define=_LITTLE_ENDIAN link=static install ... the whole thing just does not build. It fails very early with the following error: g++-4.0: installation problem, cannot exec '/usr/bin/arm-apple-darwin9- g++-4.0.1': No such file or directory Any ideas/hints how to build the boost libraries for the iPhone? (By the way, /usr/bin/arm-apple-darwin9-g++-4.0.1 is existing on my Mac.) Thanks in advance, Mani -- http://mani.de - friendly software iVolume - listen to music freehand Sahara - sand in your pocket Watchdog - baffle the curious LittleSecrets - the encrypted notepad

2 1

Re: [Boost-users] boost variant compile time visitation: verification
by Hicham Mouline 21 Apr '09

21 Apr '09

Following my prev message, I wrote the following test on a win platform, compiled with VS2005 /O2 in release mode #include <iostream> #include <boost/variant/variant.hpp> #include <windows.h> typedef boost::variant<int, double, char> VT; struct print_type : public boost::static_visitor<const char*> { const char* operator()( int ) { return "int"; } const char* operator()( double ) { return "double"; } const char* operator()( char ) { return "char"; } }; int main() { const VT v('c'); LARGE_INTEGER usec0, usec1; const char* volatile res; QueryPerformanceCounter( &usec0 ); for(size_t s=0; s<2000000; ++s) res = v.apply_visitor( print_type() ); QueryPerformanceCounter( &usec1 ); std::cout<< (usec1.QuadPart - usec0.QuadPart) <<std::endl; QueryPerformanceCounter( &usec0 ); for(size_t s=0; s<2000000; ++s) res = "char"; QueryPerformanceCounter( &usec1 ); std::cout<< (usec1.QuadPart - usec0.QuadPart) <<std::endl; } There is no statistically significant difference between the 2 cases. Can I deduce that at least on this platform, for this code, the visitation is done at compile-time? Rds,

3 3