- Boost-users - lists.boost.org

Boost::Spirit problem
by Xavier Pegenaute 20 May '10

20 May '10

Hi, I have a problem converting "const boost::phoenix::actor<boost::spirit::argument<0> >&" to int, anyone know how I can do it?, looking in the documentation seems it should work as it is, but it does not ..., any idea? PD: Find attached a simple example of the problem. Thanks a lot. Xavi.

2 1

[iostreams]Can I write data directly to the device of a stream?
by Comet 20 May '10

20 May '10

Hi, I assumed that stream wrote data directly to it's associated device, but after a few tests I found I was wrong. I wrote this code snippet: <--- Code ---> int main() { using namespace std; using namespace boost::iostreams; string src; src.assign(10 * 1024, 'a'); size_t size = src.size(); // ContainerDevice is from boost::iostreams::example typedef ContainerDevice<std::vector<char> > MyDevice; vector<char> container; string str; container.reserve(size); str.reserve(size); for (size_t i = 0; i < 1000000; ++i) { // Version 1 MyDevice device(container); stream<MyDevice> io(device); io.write(src.c_str(), src.size()); io.flush(); container.clear(); // Version 2 #if 0 str.insert(str.end(), src.begin(), src.end()); container.insert(container.end(), str.begin(), str.end()); str.clear(); container.clear(); #endif } return 0; } <--- Code-End ---> The first version(write data to the device by stream) is slower than the second(copy data, then copy again to device), so I guess stream do some unnecessary copy operations (eg: copy data to inner buffer, then copy again from buffer to device). I tried replacing stream<MyDevice> io(device); with stream<MyDevice> io(device, 0, 0); to set inner buffer size to 0, but the result was much more slower. I set a breakpoint at ContainerDevice::write(const char_type* data, std::streamsize n) then found n passed was always 1, which means every call wrote only 1 byte. If there is a way to write data directly to the device, my program will be very efficient. thanks.

1 0

[units] How to assign "ft" to an si::length?
by Terry Golubiewski 20 May '10

20 May '10

I am having trouble seeing how to use mixed units. The program below compiles and runs. See the changes I "wish" I could make as comments. terry #include <boost/units/systems/si.hpp> #include <boost/units/base_units/us/foot.hpp> #include <boost/units/quantity.hpp> #include <boost/units/io.hpp> #include <iostream> #include <iomanip> using namespace std; using namespace boost; using namespace boost::units; int main() { static const si::length m = si::meter; // static const si::length ft = ???; typedef quantity<si::length> Length; Length x = 2.0 * m; Length y = 3.0 * m; // instead of "m" use "ft"; cout << "x=" << x << " y=" << y << endl; Length z = x + y; cout << "x+y=" << z << endl; // display result in "ft" return 0; } // main

3 10

[thread] Getting the current thread object
by Postal Peon 20 May '10

20 May '10

Is there a call to get the current running thread object? (e.g. boost::this_thread::instance()). I think some older boost release had the ability to get the current thread via the default constructor (prior to 1.34). Arnold

2 1

Re: [Boost-users] Boost-users Digest, Vol 2362, Issue 3
by Edward Rankin 20 May '10

20 May '10

On Tue, May 18, 2010 at 2:39 AM, <boost-users-request(a)lists.boost.org>wrote: > > > Edward Rankin wrote: > > On Mon, May 17, 2010 at 6:53 PM, <boost-users-request(a)lists.boost.org > > <mailto:boost-users-request@lists.boost.org>> wrote: > > > > > > Hi, > > I have a large project and for what ever reason if I add > xpressive.hpp > > in my precompiled header, it completly kills intelligence. So I > > figured > > I'd move the stuff I want to use off to a lib. Works fine in debug > but > > if I build the release, this lib tries to link in the Microsoft > > debug libs. > > > > In my project: > > 1>mfcs90.lib(nolib.obj) : warning LNK4098: defaultlib 'mfc90d.lib' > > conflicts with use of other libs; use /NODEFAULTLIB:library > > 1>mfcs90.lib(nolib.obj) : warning LNK4098: defaultlib 'mfcs90d.lib' > > conflicts with use of other libs; use /NODEFAULTLIB:library > > and just this one in the test solution: > > 1>msvcrt.lib(cinitexe.obj) : warning LNK4098: defaultlib > 'msvcrtd.lib' > > conflicts with use of other libs; use /NODEFAULTLIB:library > I just realised that your issue might have nothing to do with boost. The warning messages above mention mfcs90.lib which boost definitely doesn't use. Something strange is going on, but it is unlikely to have anything to do with boost. > > > > I have been able to recreate the problem in a small solution that > > should > > not dependent on my directories other than setting the boost path: > > > > <http://lakeweb.net/downloads/RaVpath.zip> > > (20k) vc 2008 (9.0) boost 1_42 > > > > And I don't know, this may have nothing to do with boost, but I have > > other dlls and libs without this problem. It is coming from > > 'boost::filesystem' as best as I can determine. > > > > And maybe someone has seen this before and knows off the top of > > their head. > > > > Thanks, Dan. > > > > > > What's happening here is that the libs you built were built against > > the microsoft debug libs. You will need to build a separate set of > > libs which were built against the microsoft release libs. This means > > when you are compiling your actual program in release mode, you will > > need to point the linker to the new set of libs you built. > > Hi Edward, > Thanks. I do think I have all this in place. This may be my problem and > boost is the wrong place to ask. But I have: > > TestCon.cpp > #ifdef _DEBUG > #pragma comment( lib, > "C:\\cpp\\ReserveAnalyst_9\\RaVpath\\static_libd\\RaVPathd.lib" ) > #else > #pragma comment( lib, > "C:\\cpp\\ReserveAnalyst_9\\RaVpath\\static_lib\\RaVPath.lib" ) > #endif > > and in the static_lib 'should' be only release dependent. I see here I > forgot to set the path relative for the test case. Should have been: > "..\\static_lib\\RaVPath.lib" > > I'm not sure how VS chooses libs but I have checked the .vcproj directly > and only NDEBUG is used for static_lib. And I don't have this trouble > with any other lib or dll I build. > You seem to be doing everything right. You might want to try from scratch building this lib and see this error still occurs. It could be some obscure setting we have overlooked. > And thanks OvermindDL1, > I see I would qualify for the $99 license. I'll give this strong > consideration. (Visual Assist X) > I use this at work too, and it is pretty cool tool. > > Best, Dan. > > Regards, Edward

1 0

[BGL] VertexProperties as a Nested Type
by Hossein Haeri 19 May '10

19 May '10

Dear all, What is my chance to get VertexProperties back in adjacency_list<OutEdgeList, VertexList, Directed, VertexProperties, EdgeProperties, GraphProperties, EdgeList> as a nested type? So, for example: typedef adjacency_list<vecS, vecS, directedS, string> graph_type; where VertexProperties == string. What nested type can I use to get to string from graph_type? It would have made a lot of sense for example if the following piece of code worked: typedef graph_traits<graph_type>::vertex_type my_type; and, as a result, my_type == string. Is a similar thing possible? TIA, --Hossein

2 5

Re: [Boost-users] [BGL] Is adjacency_iterator not Equality Comparable?
by Hossein Haeri 19 May '10

19 May '10

Dear Jeremiah, > Could you please send a complete program that I can try to compile here? Please find a minimalist (but different) example in the attachments. I currently do know what's wrong but it's probably still better for my next question to attach the file. Consider the following line in my attached code: iterator pos = find(begin, end, "3"); I know that this line doesn't compile because the value type of vertex iterator is not necessarily the same as VertexProperties. But, then, is there a simple way to find the vertex iterator that points to "3"? (I mean something as succinct as the above line of code. I DO already know that it's possible to pass an ugly function object to help find(). But, that's far too ugly.) TIA, --Hossein

2 2

[gil]Building io extensions with Visual Studio
by Mariusz Kwiczala 19 May '10

19 May '10

Hello, Is it any script, bat to automate compilation of gil io extensions? I would like use gil and io extension with Visual Studio 2010. Could anyone help me write below bat J or give me some advise how to automate building of gil io extensions? Built.bat Rem build boost @echo off call "%VS100COMNTOOLS%vsvars32.bat" bjam variant=debug threading=multi link=shared bjam variant=release threading=multi link=shared rem build gil io extensions .. .. . Mariusz

1 0

Boost on xcode(mac osx 10.6.3)
by Rob Biermann 19 May '10

19 May '10

Hi guys, im trying to add boost in a C++ project on Mac OS X for a while, but im stuck with reference errors. Undefined symbols: "boost::system::get_generic_category()", referenced from: __static_initialization_and_destruction_0(int, int)in Main.o __static_initialization_and_destruction_0(int, int)in Main.o __static_initialization_and_destruction_0(int, int)in Main.o "boost::system::get_system_category()", referenced from: boost::asio::error::get_system_category() in Main.o __static_initialization_and_destruction_0(int, int)in Main.o __static_initialization_and_destruction_0(int, int)in Main.o ld: symbol(s) not found collect2: ld returned 1 exit status I builded librarys with bjam, i linked them in the LIBRARY_SEARCH_PATHS: LIBRARY_SEARCH_PATHS = /Users/*usr*/boost_1_43_0/bin.v2/libs/date_time/build/darwin-4.2.1/release/threading-multi /Users/*usr*/boost_1_43_0/bin.v2/libs/filesystem/build/darwin-4.2.1/release/link-static/threading-multi /Users/*usr*/boost_1_43_0/bin.v2/libs/system/build/darwin-4.2.1/release/link-static/threading-multi / users/*usr*/boost_1_43_0/stage/lib <<Not sure if this is right, but no documentation shows right options Also i made sure the include paths are correct: HEADER_SEARCH_PATHS = /users/*usr*/boost_1_43_0 I looked on the web for answeres, but lots of info was not on mac, and they only suggested to do things without any form of instructions. I got boost working on microsoft visual studio 2008 on windows 7, so i know it should work, but i am not sure if it is supposed to work on xcode. If anyone has more experience with this, can you please help me? :) Thanks in advance. EDIT: I'm trying to include asio.hpp in boost folder

3 2

Bug in darwin.jam of boost 1.43 prevents Mac OS X backwards compatibility
by Manfred Schwind 19 May '10

19 May '10

After debugging problem that my app did not launch on older Mac OS X versions, I figured out that darwin.jam of boost 1.43 has a nasty bug: These lines: case mac* : { flags darwin.compile OPTIONS <macosx-version-min>$(version-feature) : -miphoneos-version-min=$(version[2-]:J=.) ; flags darwin.link OPTIONS <macosx-version-min>$(version-feature) : -miphoneos-version-min=$(version[2-]:J=.) ; } Should be: case mac* : { flags darwin.compile OPTIONS <macosx-version-min>$(version-feature) : -mmacosx-version-min=$(version[2-]:J=.) ; flags darwin.link OPTIONS <macosx-version-min>$(version-feature) : -mmacosx-version-min=$(version[2-]:J=.) ; } ("macosx" instead of "iphoneos". Seems to be a copy&paste bug.) To get backwards compatible binaries it's essential to pass the appropriate -mmacosx-version-min option to the compiler. But because of this bug, this option was never passed, so my built boost libraries where not backwards compatible. It took me a whole day to figure this out. First I checked all my build settings of my app. The bjam build settings looked OK, so it was not obvious that the boost libraries were the cause of all the trouble. Just wanted to let you know, in case other people have the same problem. Mani -- http://mani.de - friendly software iVolume - listen to music hands-free LittleSecrets - the encrypted notepad Sahara - sand in your pocket Watchdog - baffle the curious

1 0