November 2008 - Boost - lists.preview.boost.org

[thread?] code for binding/unbinding current thread to a processor
by k-oli＠gmx.de 18 Nov '08

18 Nov '08

Hello, I've uploaded code to boost vault which allows to bind/unbind current thread to a processor. boost::this_thread::bind_to_processor( unsigned int n); // thread gets bound to processor n boost::this_thread::bind_to_any_processor(); // thread can be scheduled to all available processors platforms: Windows, Linux, Solaris, HP/UX, AIX, IRIX best regards, Oliver -- Der GMX SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen! Ideal für Modem und ISDN: http://www.gmx.net/de/go/smartsurfer

1 0

How do I change template-depth for the qcc toolset so that the accumulator tests pass
by Niklas Angare 18 Nov '08

18 Nov '08

The accumulators/test/weighted_tail_variate_means test fails with "template instantiation depth exceeds maximum of 100" with the qcc toolset. I've tried to increase the depth by editing tools/build/v2/tools/qcc.jam so that it's like gcc.jam which has a default depth of 128 but the compiler is still called with -Wc,-ftemplate-depth-100. How do I change the default depth? I've tried the attached patch which makes that section of the file equal to gcc.jam but it doesn't work. I also tried leaving the $(1) which I have no idea what it does but the result was the same. Regards Niklas Angare

2 3

BOOST_FOREACH not compiling with Multindex containers, Intel Compiler version 11
by lipik 17 Nov '08

17 Nov '08

Hi, I use boost.Multindex in my code, to simulate std::map. The code works fine on VC++ 2003 and Intel Compiler 10, but after upgrading to ICL 11 (latest), I can't use BOOST_FOREACH macros like so: BOOST_FOREACH (const parameters_map::value_type &pv, params) { } The compiler throws up the following error: .\interaction_request_handler.cpp(192): error: operand types are incompatible ("boost::foreach_detail_::rvalue_probe<const parameters_map>" and "const parameters_map") BOOST_FOREACH (const parameters_map::value_type &pv, params) { everything else compiles fine. I am guessing that some automatic conversion is not being triggered in this compiler. Any hints as to what I can try tweaking to get this to compile? Else I will just have to expand all FOREACH macros manually... I am using Boost 1.35, and would rather not be forced to upgrade just yet (since I also use built components, and I remember it was quite time-consuming, building the required components with Intel manually since bjam support for it is spotty). Thanks

2 2

[serialization] Warnings with MSVC 7.1
by Chard 17 Nov '08

17 Nov '08

I've been working on a cut-over of a code-base from Boost 1.34.1 to 1.37 and have found I now get a lot of compiler warnings in the serialization code. The warnings (emitted via BOOST_STATIC_WARNING) appear to be where non-polymorphic objects are serialized via shared_ptr. The boost serialization code seems to want them to be polymorphic. Are these benign warnings that can be disabled? Apologies if this has been raised earlier; I had a good look.. The following example code triggers the warnings (MSVC 7.1): ====== #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/serialization/serialization.hpp> #include <boost/serialization/shared_ptr.hpp> using namespace boost; struct X { template <typename T> void serialize(T &ar, const unsigned int version) {} }; typedef boost::shared_ptr<X> X_P; void test_ser() { X_P x_p; std::ifstream ifstr("test.txt"); boost::archive::text_iarchive ar(ifstr); ar & x_p; }

2 3

[threadpool] new version 16: pining of processors (processor affinity)
by k-oli＠gmx.de 17 Nov '08

17 Nov '08

Hello, I've uploaded a new version (16) to boost vault. In the new version as much worker-threads are created as boost::thread::hardware_concurrency() returns if no poolsize is specified in the pool constructor . Each worker-thread will be pined to one processor. Currentyl supported platforms for processor pining: Windows, Linux, HP/UX, AIX, IRIX, Solaris. regards, Oliver

4 4

gcc 3.6 compile for nios2
by Shlomo Kut-SM 16 Nov '08

16 Nov '08

Hi, I am trying cross compile boost for the nios2 processor. I am running into trouble when bjam generates the -pthread compiler option which is rejected by gcc for nios. I would like it to generate -D_REENTRANT instead. How can I do this? Here is what I am doing now: 1. cd boost_1_35_0 2. echo "using gcc : nios : nios2-linux-g++ ;"> tools/build/v2/user-config.jam 3. ../bjam clean 4. ../bjam -d2 --toolset=gcc-nios '-sBUILD=release static multi/single' link=static --prefix= --layout=system --with-thread --with-signals --with-filesystem install Please can you help? Shlomo Best Regards, Shlomo Kut eMail: shlomo_kut(a)softmodule.com Mobile: 972-54-771-8308

1 0

[iostreams] How to overload the flush on a outpou stream
by vicente.botet 16 Nov '08

16 Nov '08

Hi, I want to to take a specific action when I flush an output stream. Can Boost.Iostreams be of any use? Thanks is advanve, Vicente

1 0

[unordered] patch to remove warning on Visual C++
by Christopher Lux 16 Nov '08

16 Nov '08

Hi, appended is a small patch to fix some annoying warnings using Visual C++ 8.0 and 9.0. It would be nice if this could be looked at and applied to the trunk. The warnings were: 1>brick_pool.cpp 1>E:\devel\working_copies\schism_x64_vc90_sp1\externals\inc\boost\boost/unordered/unordered_map.hpp(388) : warning C4396: 'boost::operator ==' : the inline specifier cannot be used when a friend declaration refers to a specialization of a function template 1>E:\devel\working_copies\schism_x64_vc90_sp1\externals\inc\boost\boost/unordered/unordered_map.hpp(389) : warning C4396: 'boost::operator !=' : the inline specifier cannot be used when a friend declaration refers to a specialization of a function template Regards -chris -- Christopher Lux | | Bauhaus-Universität Weimar | faculty of media - virtual reality systems group

2 4

[Multiarray] & [Range] MSVC/gcc warning: returning address of local variable or temporary
by John Reid 15 Nov '08

15 Nov '08

The following code generates a worrying warning under MSVC: #include <boost/range/iterator_range.hpp> #include <boost/multi_array.hpp> typedef boost::multi_array< double, 1 > my_array; typedef my_array::const_iterator it; void test(const my_array & a) { boost::iterator_range< it > it_range(a.begin(), a.begin() + 1); it_range[0]; } C:\Dev\ThirdParty\boost\boost_1_37_0\boost/range/iterator_range.hpp(356) : warning C4172: returning address of local variable or temporary C:\Dev\ThirdParty\boost\boost_1_37_0\boost/range/iterator_range.hpp(354) : while compiling class template member function 'const double &boost::iterator_range<IteratorT>::operator [](__w64 int) const' with [ IteratorT=it ] iterator_range.cpp(8) : see reference to class template instantiation 'boost::iterator_range<IteratorT>' being compiled with [ IteratorT=it ] I'm using msvc 8.0 express and boost 1.37.0 release. I've seen similar warnings from gcc 4.3 on similar code. Is this a problem with boost.Multiarray, boost.range or the compiler? My guess is that Multiarray has defined its iterators incorrectly. Can anyone confirm this/provide a fix? Thanks, John.

6 11

Boost conformance to standard c++0x libraries statement
by vicente.botet 15 Nov '08

15 Nov '08

Hi, There are a lot of Boost libraries already included in the C++ Standards Committee's Library Technical Report (TR1). More Boost libraries are proposed for the upcoming TR2. As all these libraries do not conform completly to the standard for different raisons: the compliler do not support some feautres of the C++0x, the emulation seems to complex or any other justification. I think that it will help the users to have a clear statement, library by library, of which features the library conform to or not. This statement should contain the library feature(class, typedef, constants, functions, ...), dependencies on other C++0x features, whether it conforms on C++0x compilers supporting the needed compiler features, if it use some workaround on other compilers and finaly for the non conformant features why the library do not conforms. It will be also useful to state which features on the library are not on the standard. I have prepared as an example a table for the Boost.Thread library (limited tothread and this thread API). Anthony the informations on these tables could be wrong. Attached boost_thread_conformance.qbk and boost_thread_conformance.html Any comments and ideas are welcome. HTH, Vicente

1 0