April 2009 - Boost-users - lists.preview.boost.org

[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

Why doesn't boost use empty throws() list?
by Joseph Garvin 21 Apr '09

21 Apr '09

For example, all the smart pointer headers have: // never throws Next to their definitions, rather than an actual "throws()". I'd guess this has to do with compiler compatibility, but there are already plenty of macro hacks for enabling/disabling details like this on a compiler by compiler basis. Just curious.

4 3

[Graph] How to get the key given a value in property_map
by Phani Deepak Parasuramuni 21 Apr '09

21 Apr '09

hi, i want to name the vertices and add the edges between vertices by referring to their names dynamicaly. i 'll have something like this. typedef boost::property<boost::vertex_name_t, string> VName; typedef boost::adjacency_list<boost::vecS, boost::vecS, boost::directedS, VName > Graph; typedef boost::property_map<Graph, boost::vertex_name_t>::type VertexNames; Graph g; VertexNames names = get(vertex_name, g); boost::add_vertex(VName("me") , g); boost::add_vertex(VName("myNeighbour") , g); i want to be able to do something like this. boost::add_edge( names.getKey("me"), names.getKey("myNeighbour") ,g); // but there is no any such kind of property_map i looked around and found that there are ways to get the name of vertex if i have a vertex_descriptor but how can i get a vertex_descriptor given the vertex_name there are only three ops defined on property_map :- get,put,iterator; is there any other way to get the key from value other than iterating through all the pairs in property_map & matching each pair.second with vertex_name? pardon me if it is a basic thing, i just cant able to get at it. Any help would be greatly appreciated. Thanks, Phani.

2 1

Blocking receive in message queue
by Johan Mazel 21 Apr '09

21 Apr '09

Hi I'm trying to use a message with a blocked (or timeout) receiver through the receive (or timed_receive) method(s). But the problem is that I can open the message once and then, I always get the "No such file or directory" message. It looks like the method can open the message queue on reading mode only once since each time that I launch it after, it ends right after the constructor of the message_queue object. I forced the program which send data on the message queue to be running through an infinite loop and I don't use a destructor on the message object inside the sending program. So I assume that the message_queue is still there. So, my question is : is there a real limitation which force me to open the message queue only one time in open_only mode ? Thanks in advance for the help. Johan Mazel

1 0