August 2009 - Boost-users - lists.boost.org

[parameter] How to create make_ functions with boost::parameter returning templated classes?
by jesseperla 03 Aug '09

03 Aug '09

I want to write a make_ function for a complicated template, but can't figure out how to determine the deduced types for the return type: Say my class is: template<class F> struct templated_class { F f_; double y_; templated_class(F f, double y) : f_(i), y_(y){} }; I tried variations on the following for the template argument of templated_class: BOOST_PARAMETER_NAME(y) BOOST_PARAMETER_NAME(f) BOOST_PARAMETER_FUNCTION( (templated_class<typename boost::parameter::binding<ArgumentPack, tag::f>::type>), make_templated_class, tag, (required (f, *) ) (optional (i, *, 2.0) ) ) { return templated_class<typename boost::parameter::binding<ArgumentPack, tag::f>::type>(f, y); } This doesn't work. The parameter::binding was my attempt to extract the type, but ArgumentPack doesn't exist. Is this the right general method? What template parameter can I use for the metafunction? Thanks, Jesse

1 0

Boost's Linear Algebra Solution for y=Ax
by Peverall Dubois 03 Aug '09

03 Aug '09

Does boost have one? Where A, y and x is a matrix and vectors respectively. Either y or x can be unknown. - G.V.

2 1

[Serialization: Newbie] Compilation Problem Text Archive
by Soumen 03 Aug '09

03 Aug '09

Hi, I'm facing some compilation problem when I use text archive. But the same code works fine when I use binary archive. And it does write something on disk with specified file name (I'm yet to finish the restore part. Otherwise, I could've said it works). Do I need to do something differently? The error I get is something like following: ------------------------------------------------------------------------------------------ /boost/1_35_0/include/boost/archive/detail/oserializer.hpp: In function ‘void boost::archive::save(Archive&, T&) [with Archive = boost::archive::text_oarchive, T = TableMgr*]’: /boost/1_35_0/include/boost/archive/detail/common_oarchive.hpp:64: instantiated from ‘void boost::archive::detail::common_oarchive<Archive>::save_override(T&, int) [with T = TableMgr*, Archive = boost::archive::text_oarchive]’ /boost/1_35_0/include/boost/archive/basic_text_oarchive.hpp:75: instantiated from ‘void boost::archive::basic_text_oarchive<Archive>::save_override(T&, int) [with T = TableMgr*, Archive = boost::archive::text_oarchive]’ /boost/1_35_0/include/boost/archive/detail/interface_oarchive.hpp:64: instantiated from ‘Archive& boost::archive::detail::interface_oarchive<Archive>::operator<<(T&) [with T = TableMgr*, Archive = boost::archive::text_oarchive]’ /remote/gsoumen/hiers/trunk/code/cmds/Cmds.cpp:1518: instantiated from here /boost/1_35_0/include/boost/archive/detail/oserializer.hpp:564: error: invalid application of ‘sizeof’ to incomplete type ‘boost::STATIC_ASSERTION_FAILURE<false>’ cc1plus: warnings being treated as errors ------------------------------------------------------------------------------------------ My code is something like following: #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> //I use binary archive hpp when I use binary archive mode #include <boost/serialization/serialization.hpp> #include <boost/serialization/utility.hpp> std::ofstream ofs(fileName.c_str(), ios_base::out); // with bin archive I also use ios_base::binary boost::archive::text_oarchive oarchive(ofs); //for bin archive, oarchive is of binary_oarchive type TableMgr *tblMgr = TableMgr::instance(); if (tblMgr) { oarchive << tblMgr; } //I've TableMgr::serialize() defined ... Regards, ~ Soumen -- View this message in context: http://www.nabble.com/-Serialization%3A-Newbie--Compilation-Problem-Text-Ar… Sent from the Boost - Users mailing list archive at Nabble.com.

2 6

any C++ portable library to create pdf documents?
by Frédéric Bron 03 Aug '09

03 Aug '09

I have looked in in the vault but could not find any library to create pdf documents. Does anybody knows a good one written in portable C++? Regards, Frédéric

3 2

[asio] AES encryption/decryption
by Igor R 03 Aug '09

03 Aug '09

Hello all, Does someone have an experience with the subj? In my application I need an AES-encrypted socket. I try to accomplish this by building a custom asio socket that performs all the needed encryption (with openssl) and delegates the i/o to its service. However, I encountered an issue that seems to be unsolvable with the current approach: from one hand, every read opeartion has to return the number of *decrypted* bytes, but from the other hand, bytes_readable command returns the size of the available *encrypted* data. Sometimes this causes severe problems: void connection::read() { asio::socket_base::bytes_readable command(true); socket_.lowest_layer().io_control(command); size_t canRead = command.get(); if (canRead) // this line creates completion condition that might never be met, because read_some() would return less than canRead asio::read(socket_, input_, asio::transfer_at_least(canRead), err); } I'd appreciate any idea - probably I'm going the wrong way, and there's some more appropriate design that would give better results? Thanks.

2 1

[Test] strange test behavior
by Emre Guney 02 Aug '09

02 Aug '09

Greetings, I have the following class to test a class named Scoring (implements a graph using Boost Graph Library). If the line marked with //!!! below is uncommented the scores yielded by _s.getNetwork().getVertexScore(*it) are garbage values (like 2.57117366e+29 ). If it is commented test fails but the values are initial (sensible) values. When I run the same function apart from the tests, the function behaves correctly and gives correct calculated scores. There should be a context problem but I could not figure it out. I would be grateful for any suggestions and ideas. class TestScoring { Scoring _s; public: TestScoring() { _s = Scoring("test_vertices_small.txt", "test_edges_small.txt"); } ~TestScoring() { } void test_calculatedScores() { float vertex_scores[] = { 0.75, 4.0/3, 1.5, 1.0/3, 0.005 }; std::size_t n_vertex = sizeof(vertex_scores)/sizeof(float); //!!! this call changes the behavior of the test //!!! _s.run(1,2); VertexIterator it, itEnd; for(boost::tie(it, itEnd) = _s.getNetwork().getVertexIterator(); it != itEnd; it++) { BOOST_CHECK_CLOSE(_s.getNetwork().getVertexScore(*it), vertex_scores[_s.getNetwork().getVertexIndex(*it)], 0.00001f); } } }; // TestScoring is called from the following suite class TestSuiteScoring: public test_suite { public: TestSuiteScoring(): test_suite("test_suite_scoring") { shared_ptr<TestScoring> instance(new TestScoring()); test_case *calculatedScores = BOOST_CLASS_TEST_CASE( &TestScoring::test_calculatedScores, instance); add(calculatedScores); } ~TestScoring() { //instance.release(); } }; Emre

2 1

Remove elements from accumulator
by Archie14 02 Aug '09

02 Aug '09

I am trying to use accumulator to calculate statistics in "rolling window", traversing the data. I searched accumulators library for a method to remove first N elements from the accumulator and did not find one. Question: is is possible to remove first N values from the accumulator?

2 1

non-POD C++ objects and Boost.Interprocess
by Selçuk Cevher 02 Aug '09

02 Aug '09

Hi, Is it possible to pass user-defined "non-POD" C++ class objects to message queues and retrieve them back on the receiver side using Boost.Interprocess ? Thanks.

2 1

Valgrind's DRD tool produces false positives when used with Boost.Thread?
by Seth Fowler 01 Aug '09

01 Aug '09

Hi all, I've been trying to debug some data races in an application that uses Boost.Thread. I tried running valgrind's DRD tool against my binary to see if it could locate the source of the problem. It produced more errors than I expected, so I progressively removed portions of my application unless I was left with almost nothing. I found that even with the most minimal program I could produce, the DRD tool still reported errors - in fact, it even reported errors with the examples included with the Boost distribution! I was able to reproduce this behavior on Mac OS X 10.5.7 with Boost 1.39, and on Ubuntu 9.04 with Boost 1.37. Here are steps to reproduce the problem: 1. Build the sample program thread.cpp included in the Boost distribution in libs/thread/example. For Boost 1.39, I used the following command: g++ -L<Boost lib directory> -lboost_thread-xgcc40-mt-1_39 thread.cpp - o thread 2. Run it using valgrind's DRD tool: valgrind --tool=drd --var-info=yes ./thread On Mac OS X 10.5.7, I get the following output: ==20729== drd, a thread error detector. ==20729== Copyright (C) 2006-2009, and GNU GPL'd, by Bart Van Assche. ==20729== Using Valgrind-3.5.0.SVN and LibVEX; rerun with -h for copyright info ==20729== Command: ./thread ==20729== --20729-- ./thread: --20729-- dSYM directory is missing; consider using --auto-run- dsymutil=yes setting alarm for 5 seconds... ==20729== Conflicting load by thread 1 at 0x00012018 size 4 ==20729== at 0x58B1: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x390C: boost::thread::join() (in ./thread) ==20729== by 0x9A13: main (in ./thread) ==20729== Allocation context: Data section of ./thread ==20729== Other segment start (thread 2) ==20729== at 0x396B3: pthread_mutex_lock (drd_pthread_intercepts.c: 580) ==20729== by 0x5993: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment end (thread 2) ==20729== at 0x39BC2: pthread_mutex_unlock (drd_pthread_intercepts.c:633) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment start (thread 2) ==20729== at 0x396B3: pthread_mutex_lock (drd_pthread_intercepts.c: 580) ==20729== by 0x58D2: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment end (thread 2) ==20729== at 0x39BC2: pthread_mutex_unlock (drd_pthread_intercepts.c:633) ==20729== by 0x5988: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== ==20729== Conflicting load by thread 1 at 0x00012014 size 4 ==20729== at 0x58B4: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x390C: boost::thread::join() (in ./thread) ==20729== by 0x9A13: main (in ./thread) ==20729== Allocation context: Data section of ./thread ==20729== Other segment start (thread 2) ==20729== at 0x396B3: pthread_mutex_lock (drd_pthread_intercepts.c: 580) ==20729== by 0x5993: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment end (thread 2) ==20729== at 0x39BC2: pthread_mutex_unlock (drd_pthread_intercepts.c:633) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment start (thread 2) ==20729== at 0x396B3: pthread_mutex_lock (drd_pthread_intercepts.c: 580) ==20729== by 0x58D2: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment end (thread 2) ==20729== at 0x39BC2: pthread_mutex_unlock (drd_pthread_intercepts.c:633) ==20729== by 0x5988: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== ==20729== Conflicting load by thread 1 at 0x00013b5c size 4 ==20729== at 0x993A: boost::detail::get_once_per_thread_epoch() (in ./thread) ==20729== by 0x58BA: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x390C: boost::thread::join() (in ./thread) ==20729== by 0x9A13: main (in ./thread) ==20729== Allocation context: Data section of ./thread ==20729== Other segment start (thread 2) ==20729== at 0x227FF0: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment end (thread 2) ==20729== at 0x396B3: pthread_mutex_lock (drd_pthread_intercepts.c: 580) ==20729== by 0x58D2: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== ==20729== Conflicting load by thread 1 at 0x00013b5c size 4 ==20729== at 0x9972: boost::detail::get_once_per_thread_epoch() (in ./thread) ==20729== by 0x58BA: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x390C: boost::thread::join() (in ./thread) ==20729== by 0x9A13: main (in ./thread) ==20729== Allocation context: Data section of ./thread ==20729== Other segment start (thread 2) ==20729== at 0x227FF0: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment end (thread 2) ==20729== at 0x396B3: pthread_mutex_lock (drd_pthread_intercepts.c: 580) ==20729== by 0x58D2: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== ==20729== Conflicting load by thread 1 at 0x00397680 size 4 ==20729== at 0x20038D: pthread_setspecific (in /usr/lib/ libSystem.B.dylib) ==20729== by 0x997F: boost::detail::get_once_per_thread_epoch() (in ./thread) ==20729== by 0x58BA: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x390C: boost::thread::join() (in ./thread) ==20729== by 0x9A13: main (in ./thread) ==20729== Allocation context: Data section of /usr/lib/libSystem.B.dylib ==20729== Other segment start (thread 2) ==20729== at 0x227FF0: thread_start (in /usr/lib/libSystem.B.dylib) ==20729== Other segment end (thread 2) ==20729== at 0x396B3: pthread_mutex_lock (drd_pthread_intercepts.c: 580) ==20729== by 0x58D2: void boost::call_once<void (*) ()>(boost::once_flag&, void (*)()) (in ./thread) ==20729== by 0x30CC: thread_proxy (in ./thread) ==20729== by 0x3707B: vgDrd_thread_wrapper (drd_pthread_intercepts.c:272) ==20729== by 0x228154: _pthread_start (in /usr/lib/libSystem.B.dylib) ==20729== by 0x228011: thread_start (in /usr/lib/libSystem.B.dylib) alarm sounded... ==20729== ==20729== ERROR SUMMARY: 5 errors from 5 contexts (suppressed: 725 from 8) The output on linux is similar. Does this indicate a problem with valgrind's DRD tool, Boost.Thread, or both? Does any workaround exist? Thanks in advance, - Seth Fowler

1 0

Boost.test applicability
by Joel Falcou 01 Aug '09

01 Aug '09

I have a alrge number of numeric functions (~300) overloaded on various types. What I want to do is having some automated test that check if those functions returns proper values and behave correctly in term of precision. I get a look at Boost.Test but i can't really see if it's applicable to my need and, if it is, how exactly am I supposed to generate that much test and cover all cases (random value testing, singular value test, nan/inf handling etc...). My first idea was to use Boost.preprcoessor file iteration to quickly generate BOOST_TEST_CASE functions but I'm not sure this is a scalable practice. Other point, I saw that there is check in Boost.test which check for an arbitrary precision result. However it uses raw value (like 0.001) where having a check based on ULP or RMS maybe better. Is there anyway to integrate those test using existing test framework ? Thanks in advance

3 4