March 2010 - Boost-users - lists.preview.boost.org

Re: [Boost-users] setup question
by Steven Watanabe 19 Mar '10

19 Mar '10

AMDG Rick Boesch wrote: > Hello all -- looking for a small bit of guidance please. I'm building a 64bit Boost library from scratch. I'm fairly sure I did almost everything correctly, but would like to check one nagging detail. My system is Windows 7, Visual Studio 2005 and I'm building for 64 bit, specifically AMD64. > > I believe bjam.exe built correctly because my path was set up properly for the x64 toolset and bjam.exe was deposited in the auto-generated directory BOOST_ROOT\tools\jam\src\bin.ntx86_64. To do this, I ran "build vc8" on the command line. > > My question concerns the use of bjam along with the "using" instruction in the ...\V2\user-config.jam file. I set the "using" instruction to be > > using msvc : 8.0 : cl : <setup-amd64>vcvarsall ; > > and my command line is > > bjam address-model=64 --build-dir="C:\Program Files\boost\boost_1_42_0\boost-build" --build-type=complete stage > > I built two different ways: > > 1) In a virgin cmd window -- which is not set up for 64 bit tools by default -- I ran vcvarsall manually on the command line (with no args, I set it up to default to setting up 64 bit), which sets up my 64 bit environment and which I have echo to the screen "AMD64 env" for confirmation. I confirmed I was then defaulting to the correct "cl", etc., and then executed in BOOST_ROOT the bjam command line above. I have yet to test the output, but it seemed to work. > > 2) In another virgin cmd window, I cd'd to BOOST_ROOT and executed ONLY the bjam command line above. According to the documentation for the "using" instruction, the batch file vcvarsall (which is in my default PATH) is supposed to be called before any tools are used, but I never saw "AMD64 env" echoed to the screen, nor to a .log file which I piped to in order to make sure I didn't miss anything. Yet, bjam created an output tree containing various "address-model-64" directories. When I checked my PATH after the build, it did not contain the paths that vcvarsall is supposed to set up. > The environment setup the Boost.Build does is local to each command. > Question: In scenario (2) above, vcvarsall clearly did not run in my cmd session, but did bjam run it in some internally generated shell? If no, why not (did I do something incorrectly)? If yes, is this why I didn't see my echo confirmation? You can run bjam -a -n <other options> to see the commands that Boost.Build actually runs. > So I'm wondering if I have a 64 bit build or not -- and I prefer scenario (2) to scenario (1)! > In Christ, Steven Watanabe P.S. The timestamp on this message was 2004.

1 0

custom allocators Re:pool_alloc
by B Hart 19 Mar '10

19 Mar '10

I am not understanding the use of pool with std::set. i.e. the release_memory() function referenced in last example shown on: http://www.boost.org/doc/libs/1_42_0/libs/pool/doc/interfaces.html seems to do nothing. I call GetProcessMemoryInfo() and don't see any change after the set has gone out of scope. maybe someone could relate what happens to the allocated memory when set goes out of scope and release_memory() isn't called versus what happens when set goes out of scope and release_memory() is called. I want to do something like: 1.std::set<unsigned int, std::less<unsigned int>, boost::pool_allocator<unsigned int>> v; 2. for(int i=0; i<(K); i++) { v.insert(i); } 3. use v 4. release_memory() on v //dump all set elements in one action. Note, without custom allocator this is very slow for large sets. 5. repeat 1-4 with a new set

6 19

[interprocesss] managed_mapped_file mutex question
by Andy Wiese 19 Mar '10

19 Mar '10

This is a somewhat different take on the question I asked earlier about interprocess mutexes in allocators. I am using a default interprocess::managed_mapped_file to hold a index for a multiprocess property store. There is one object named "index" in the mapped file, which is an interprocess::map. The map holds an interprocess::set of fixed-length strings. (I had an allocation problem with sets of string objects, but that's another topic) Many processes will both read and write this index map. Do the interprocess:: containers provide the needed synchronization, or do I need other synchronization mechanisms? If I am reading the docs correctly, it seems to say that I don't need additional synchronization if I am not adding/removing named objects. However, my tests break with multiple processes. I don't know if I just need my own synchronization or if a deeper problem is implied. Following is the simple test I have been using. Works with one process, breaks with two concurrent. (os-x 10.5, boost trunk) Thanks in advance for any advice! Andy ------------------------------------------------------- namespace bip = boost::interprocess; typedef struct { char t[64]; } thing_t; struct ThingLess { bool operator()(thing_t const & lhs, thing_t const & rhs) const { return (strcmp(lhs.t, rhs.t) < 0); } }; typedef bip::managed_mapped_file shared_store_t; typedef shared_store_t::segment_manager segment_manager_t; typedef bip::allocator<void, segment_manager_t> void_allocator_t; typedef bip::allocator<char, segment_manager_t> char_allocator_t; typedef bip::basic_string<char, std::char_traits<char>, char_allocator_t> char_string_t; typedef bip::allocator<char_string_t, segment_manager_t> char_string_allocator_t; typedef bip::allocator<thing_t, segment_manager_t> thing_allocator_t; typedef bip::set<thing_t, ThingLess, thing_allocator_t> thing_set_t; typedef std::pair<const char_string_t, thing_set_t> map_value_t; typedef bip::allocator<map_value_t, segment_manager_t> map_value_allocator_t; typedef bip::map< char_string_t, thing_set_t, std::less<char_string_t>, map_value_allocator_t> index_map_t; int main(int argc, char* argv[]) { //fork(); // forking here will usually cause a bus error int pid = getpid(); // find or create the mapped file std::auto_ptr<shared_store_t> store(new shared_store_t(bip::open_or_create, "/tmp/testindex", 1024*1024)); void_allocator_t allocator(store->get_segment_manager()); // find or create the index object index_map_t* index = store->get_segment_manager()- >find<index_map_t>("index").first; if (!index) { index = store->construct<index_map_t>("index") (std::less<char_string_t>(), allocator); assert(index); } // add pairs to the map, find them, remove them. for (int j=0; j<1000; ++j) { for (int i=0; i<100; ++i) { try { std::ostringstream ks; ks << "key" << i; std::ostringstream ts; ts << "thing" << i; // look for thing in map char_string_t key(ks.str().c_str(), allocator); index_map_t::iterator foundIt = index->find(key); if (foundIt == index->end()) { // add thing to map thing_t thing; strcpy(thing.t,ts.str().c_str()); thing_set_t thingSet(ThingLess(), allocator); thingSet.insert(thing); map_value_t entry(key,thingSet); index->insert(entry); std::cout << pid << " Added " << key << std::endl; } // look again... foundIt = index->find(key); if (foundIt != index->end()) { if (ts.str() == foundIt->second.begin()->t) { std::cout << pid << " FOUND " << key << ":"<< foundIt- >second.begin()->t << std::endl; } else { std::cout << pid << " FAILED? " << key << ":"<< foundIt- >second.begin()->t << std::endl; } } if (j%2) { // erase thing from map, every other pass just for some variety index->erase(key); std::cout << pid << " Erased " << key << std::endl; } } catch( bip::bad_alloc & ba ) { std::cout << "bad_alloc at " << j<<","<<i<< " "<< ba.what() << std::endl; } catch(std::exception & e) { std::cout << e.what() << std::endl; } catch(...) { std::cout << "some other exception" << std::endl; } } std::cout << pid << " --------- End Round " << j << " --------- " <<std::endl; } return 0; }

3 7

[proto] generation of adaptive precision floating-point geometric predicates
by Andrew Durward 19 Mar '10

19 Mar '10

Hi all, I'm exploring the possibility of using proto to transform an expression (at compile time) into a sequence of terms whose sum represents an increasingly accurate floating-point approximation of the result. The underlying idea is based on a paper by Shewchuk [1]. The central concept is that for any expression (a # b) where # is {+, -, *}, the result can be expressed as (x + y) where x is the floating-point approximation of (a # b), and y is an error term whose value can also be computed as a floating-point value. In toying with proto, I've been able to transform simple expressions successfully but I can't seem to figure out how to define a grammar that will handle arbitrary combinations of {+, -, *} recursively. Here's what I've got so far: struct sum_approx{/* */}; struct sum_error{/* */}; struct diff_approx{/* */}; struct diff_error{/* */}; struct prod_approx{/* */}; struct prod_error{/* */}; struct Transform : proto::or_< proto::when< proto::plus< proto::terminal< _ >, proto::terminal< _ > > , fusion::cons< proto::_make_function( sum_approx() , proto::_left , proto::_right ) , fusion::cons< proto::_make_function( sum_error() , proto::_left , proto::_right ) > >() > // ... similar substitutions for minus and multiplies > {}; So the expression (a + b) yields the sequence: sum_approx(a,b) sum_error(a,b) Here's where I'm stuck. In the case of the expression (a + b)*(c - d), the addition and subtraction should be replaced by the appropriate sequences before the multiplication gets expanded to yield the final sequence: prod_approx( sum_approx(a,b), diff_approx(c,d) ) prod_error( sum_approx(a,b), diff_approx(c,d) ) prod_approx( sum_approx(a,b), diff_error(c,d) ) prod_error( sum_approx(a,b), diff_error(c,d) ) prod_approx( sum_error(a,b), diff_approx(c,d) ) prod_error( sum_error(a,b), diff_approx(c,d) ) prod_approx( sum_error(a,b), diff_error(c,d) ) prod_error( sum_error(a,b), diff_error(c,d) ) Obviously I need a recursive transform here but I'm not sure how to pass it the sequences generated by both the left and right subtrees. struct Transform : proto::or_< // ... as before , proto::when< proto::multiplies< Transform, Transform > , /* WHAT_GOES_HERE */< Transform( proto::_left ) , Transform( proto::_right ) > > > {}; What exactly do I need here? A metafunction to generate the proper sequence? What's the proper syntax to invoke it? Can anybody point me in the right direction? Thanks for any help, andrew [1] Shewchuk, Jonathon. Robust Adaptive Floating-Point Geometric Predicates. http://www.cs.cmu.edu/afs/cs/project/quake/public/papers/robust-predicates.p s

1 0

Building boost-1.42.0 wuth VS 2005 - precompiled header error
by john blair 19 Mar '10

19 Mar '10

I am trying to build boost-1.42.0 with visual studio 2005 and getting the following error: assoc_laguerre.cpp libs\math\build\..\src\tr1\assoc_laguerre.cpp(6) : error C2858: command-line opt ion 'program database name (/Fdc:\toolchain\src\boost-1.42.0\build-win32-vc80-de bug\boost\bin.v2\libs\math\build\msvc-8.0\dbg\dbg-str-dtbs\lnk-sttc\thrd-mlt\lib boost_math_tr1-vc80-mt-gd-1_42.pdb)' inconsistent with precompiled header, which used '/Fdc:\toolchain\src\boost-1.42.0\boost_1_42_0\vc80.pdb' libs\math\build\..\src\tr1\assoc_laguerre.cpp(6) : error C2859: c:\toolchain\src \boost-1.42.0\build-win32-vc80-debug\boost\bin.v2\libs\math\build\msvc-8.0\dbg\d bg-str-dtbs\lnk-sttc\thrd-mlt\libboost_math_tr1-vc80-mt-gd-1_42.pdb is not the p db file that was used when this precompiled header was created, recreate the pre compiled header. ...failed updating 1 target... Anybody knows how to resolve this?

2 1

Link issue with condition_variable
by Bradley W. Settlemyer 19 Mar '10

19 Mar '10

Hello, I'm using boost for the first time and I'm struggling to link my program. I've added libboost_pthread-mt to the link link (which got most of the symbols), but I still can't seem to get the symbols for the condition variables. The issue could be with the code, as I find the lock-mutex-condition_variable interactions to be very confusing, but it does appear to be a link issue. Consider the following error message emitted during link command: mpicxx <objects> -D_FILE_OFFSET_BITS=64 -L/lib64 -lfuse -lrt -ldl -lboost_thread-mt -lpthread -o aio_benchmark src/aio/thread_pool_adaptive_writer.o: In function `aio::ThreadPoolWriter::Write(aio::File*, std::vector<int, std::allocator<int> >, void*, unsigned long)': /home/bws/sandbox/aiofs/src/aio/thread_pool_writer.cc:59: undefined reference to `boost::condition_variable::notify_all()' /home/bws/sandbox/aiofs/src/aio/thread_pool_writer.cc:63: undefined reference to `boost::condition_variable::wait(boost::unique_lock<boost::mutex>&)' src/aio/thread_pool_writer.o: In function `aio::ThreadPoolWriter::threadWait()': /home/bws/sandbox/aiofs/src/aio/thread_pool_writer.cc:76: undefined reference to `boost::condition_variable::wait(boost::unique_lock<boost::mutex>&)' /home/bws/sandbox/aiofs/src/aio/thread_pool_writer.cc:110: undefined reference to `boost::condition_variable::notify_one()' Is condition variable in a different library? I am using the boost stuff distributed with Fedora 11: boost-1.37.0-7.fc11.x86_64 boost-devel-1.37.0-7.fc11.x86_64 Cheers, Brad

2 2

Re: [Boost-users] program option library segfaults for exception
by Neetu Garg 19 Mar '10

19 Mar '10

I did little more digging into the issue If I compile and link following sample on Linux dynamically using following command g++ -c -I. -I/usr/local/include/boost-1_38 -I/usr/include testprogramoptions.cpp -o testprogramoptions.o g++ testprogramoptions.o -static-libgcc -L/u/neetug/Projects/performance/perfstat_cluster/lib -lboost_program_options-gcc41-mt -lpthread -o testprogramoptions [root@perf-tools-7 src]# ./testprogramoptions --o ERROR unknown option o ERROR Error while parsing CLI options Exiting Here "--o" is unknown option ======================================================================= If I compile and link following sample on Linux statically using following command g++ -c -I. -I/usr/local/include/boost-1_38 -I/usr/include testprogramoptions.cpp -o testprogramoptions.o g++ testprogramoptions.o -static-libgcc -Wl,-Bstatic -L/u/neetug/Projects/performance/perfstat_cluster/lib -lboost_program_options-gcc41-mt -lpthread -o testprogramoptions [root@perf-tools-7 src]# ./testprogramoptions --o Aborted Here "--o" is unknown option ======================================================================== I am not sure why the behavior is different with different linking option why I am not getting exception for static linking. Really appreciate any help in this issue. Here in the test program #include <iostream> #include <boost/program_options.hpp> namespace boostOptions = boost::program_options; using namespace std; int main(int argCount, char* argList[]) { try { boostOptions::options_description commonOpt("Options (common)"); commonOpt.add_options() ("help,h", "Produce help message.") ("version,v", "Prints perfstat version, and supported ONTAP versions.") ; boostOptions::options_description runModeOpt("Options ('run' mode)"); runModeOpt.add_options() ("clusterinfo", boostOptions::value<string>(), "cluster IP.") ("sample-interval", boostOptions::value<int>(), "Specifies statit/sysstat frequency") ; boostOptions::positional_options_description posOption; posOption.add("clusterinfo", -1); boostOptions::options_description allOptDesc("Allowed options :"); allOptDesc.add(commonOpt).add(runModeOpt); boostOptions::variables_map varMap; boostOptions::store(boostOptions::command_line_parser(argCount, argList). options(allOptDesc).positional(posOption).run(), varMap); boostOptions::notify(varMap); } catch(std::exception &e) { cout << "ERROR " << e.what() << "\n"; cout << "ERROR " << "Error while parsing CLI options Exiting" << "\n"; exit(0); } catch(...) { cout << "ERROR 1 " << "\n"; cout << "ERROR 1 " << "Error while parsing CLI options Exiting" << "\n"; exit(0); } } Regards, NG ----- Original Message ---- From: Neetu Garg <garg_neetu(a)yahoo.com> To: boost-users(a)lists.boost.org Sent: Wed, March 17, 2010 1:17:07 PM Subject: program option library segfaults for exception Sorry for the duplicate post, I think my message posted under different conversation. Hi, I have application that is using boost program option library. If I provide unknown option via CLI it should throw exception. The library is statically linked on windows and Linux environment. It works fine on windows platform I do get exception for unknown CLI option. On Linux platform I get "segfault" in place of exception for unknown CLI option. It is s a multi-threaded application and it also uses boost thread library. Application does throw exception on Linux also if I link dynamically. I am using following flags to link the library statically: LDFLAGS= -fexceptions -static-libgcc -Wl,-static -L/usr/lib -lboost_program_options-gcc41-mt -lboost_filesystem-gcc41-mt -lboost_thread-gcc41-mt -lboost_system-gcc41-mt -lnet7ssh -lbotan -ltar -lz -lrt -lpthread I am using following version of gcc: # g++ --version g++ (GCC) 4.1.1 Copyright (C) 2006 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Here is the core dump for the segfault: (gdb) backtrace #0 __gnu_internal::get_global () at /usr/software/src/gcc-4.1.1/libstdc++-v3/libsupc++/eh_globals.cc:58 #1 0x082c0f67 in __cxa_get_globals () at /usr/software/src/gcc-4.1.1/libstdc++-v3/libsupc++/eh_globals.cc:71 #2 0x082c0b3f in __cxa_allocate_exception (thrown_size=32) at /usr/software/src/gcc-4.1.1/libstdc++-v3/libsupc++/eh_alloc.cc:154 #3 0x080a71f8 in boost::program_options::detail::cmdline::finish_option () #4 0x080a77eb in boost::program_options::detail::cmdline::run () #5 0x0806f3aa in boost::program_options::basic_command_line_parser<char>::run () #6 0x080660e9 in CLIHandler::collectCLIOptions () #7 0x0808a0d1 in perfstat::runPerfstat () #8 0x0808af3d in main () Current language: auto; currently c++ Appreciate your help Regards, NG

1 0

[Proto] Turnign grammar into primitive transform : _default is being naughty with me :(
by joel falcou 19 Mar '10

19 Mar '10

I'm polisihng some generic recursive descent grammar to evaluate various sort of expression that only differ on how nodes and leaf are treated. I was trying, once the grammar was set up using combo of transform to make a single primitive transform to speed up compile-time. Everythign worked until I hit a brick wall while trying to put the last non-generic part (the call to proto::_default) inside my own transform. Here is link to the working.cpp and not_working.cpp stand alone source file for inspection. test were done on 1.42 release. http://www.lri.fr/~falcou/tmp/working.cpp http://www.lri.fr/~falcou/tmp/not_working.cpp The error is : http://www.lri.fr/~falcou/tmp/errors.txt thanks for any pointer on this issue which, I think, is a stupid mistake on my side. -- ___________________________________________ Joel Falcou - Assistant Professor PARALL Team - LRI - Universite Paris Sud XI Tel : (+33)1 69 15 66 35

2 2

boost::asio destructor segfault?
by Sean McAllister 19 Mar '10

19 Mar '10

I've got a class that's using ASIO to run a simple server listening for command messages from a remote interface. It's got some private members to help with this (class abridged for shortness): class ClassifierContext { private: io_service ioService_; t_socket_shptr socket_; }; It runs it's event loop in it's own thread (started with boost::thread) // Hosts thread to run IO event loop void ClassifierContext::io_loop() { try { connect(); } catch (std::exception &e) { RERROR("Error connecting to site server -- %s", e.what()); } // Work object prevents run() from stopping when it runs out of work asio::io_service::work work(ioService_); ioService_.run(); } And it periodically sends data back along a single socket associated with the ioService and also has an asynchronous read pending to listen for command messages (I can post this code if needed, it's just a couple async_sends and an async_read) The problem I'm seeing comes when the destructor for the class is run: ClassifierContext::~ClassifierContext() { ioService_.stop(); } Periodically (every 20 times or so) I get a segfault when the destructor for ioService_ is called. A representative stack trace looks like this: #0 0x0921bfd6 in ?? () #1 0x08125877 in boost::asio::detail::reactor_op_queue<int>::destroy_operations ( this=0x8ec5f40) at opt/linux/include/boost/asio/detail/reactor_op_queue.hpp:268 #2 0x081274ad in boost::asio::detail::epoll_reactor<false>::shutdown_service ( this=0x8ec5ed8) at opt/linux/include/boost/asio/detail/epoll_reactor.hpp:119 #3 0x08127b3b in ~service_registry (this=0x8ec5958) at opt/linux/include/boost/asio/detail/service_registry.hpp:75 #4 0x08127c29 in ~io_service (this=0x8492360) at opt/linux/include/boost/asio/impl/io_service.ipp:62 #5 0x08117110 in ~ClassifierContext (this=0x8492360) at ClassifierContext.cpp:91 #6 0xb781cbb9 in exit () from /lib/tls/i686/cmov/libc.so.6 #7 0xb780477d in __libc_start_main () from /lib/tls/i686/cmov/libc.so.6 #8 0x081079c1 in _start () at ../sysdeps/i386/elf/start.S:119 Has anyone seen a segfault like this before? It's intermittent so that tells me there's some kind of race condition. Any idea how I can work around it?

2 8

Compile under 32bits but not under 64bits
by Juanma Font 19 Mar '10

19 Mar '10

Hello ¡¡¡ Probed with boost 1.38 and 1.40 on linux ubuntu karmic Example: check64.cpp --------------- #include "check64.hpp" int main() { point2<int> png_size1(3,4); point2<int> png_size2; wxMessageBox( wxT("Reading png dimensions prueba.png "), wxT("Hello") ); // ok, compile into 32bits and 64bits png_size2=png_size1; // ok, compile into 32bits and 64bits png_read_dimensions( "prueba.png" ); //ERROR, THIS ONLY COMPILE WITH 32bits png_size1 = png_read_dimensions( "prueba.png" ); return 0; } ---------------------------------------- The error when try to compile with 64bits is: error: no match for ‘operator=’ in ‘png_size1 = boost::gil::png_read_dimensions(const char*)()’ Anybody knows what is happening ? The head file, check64.hpp is -------------------------- check64.hpp ------------- #ifndef CHECK64_HPP_INCLUDED #define CHECK64_HPP_INCLUDED #include <boost/gil/gil_all.hpp> #include <boost/gil/extension/io/png_io.hpp> #include <wx/wx.h> using namespace boost; using namespace boost::gil; using namespace std; #endif // CHECK64_HPP_INCLUDED --------------------------end of code -------------- Thank's a lot of

2 2