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

[interprocess] Mutex Query
by Rick Way 20 May '09

20 May '09

I want to make sure that when my program exits ungracefully (CTRL-C, seg fault, etc) it releases a Boost.Interprocess named_upgradable_mutex, if it's holding it. I've created a signal handler to catch these ungraceful exits. Is there any way to tell if my process currently holds an exclusive or sharable lock on the mutex? I've tried blindly calling mutex.unlock() and mutex.unlock_sharable() but that seems to permanently lock the mutex when the shareable lock is set. RickW

2 1

[test] Ignoring SIGCHLD for signal codes other than CLD_EXITED
by Gevorg Voskanyan 20 May '09

20 May '09

I could not find a response to the following message: http://lists.boost.org/boost-users/2008/09/40618.php , which proposes to modify ignore_sigchild() implementation to account for a proposed BOOST_TEST_IGNORE_SIGCHLD preprocessor switch. I am seconding this request. Has it ever been considered? Or was it the message simply went unnoticed? Thanks, Gevorg

3 10

failure of VertexListGraphConcept check
by Christian Bähnisch 20 May '09

20 May '09

Hello, Trying to write an BGL adaption of a BGL foreign graph-class I followed the conversion guide: "http://www.boost.org/doc/libs/1_39_0/libs/graph/doc//leda_conversion.html" Using "boost::function_requires<boost::VertexListGraphConcept<MyGraph> >();" in order to check my implemetation gcc complains about: "boost/graph/graph_concepts.hpp:199: error: no matching function for call to 'vertices(const MyGraph&)'" I further noticed that the test-case "stanford_graph_cc.cpp" found in "BOOST/libs/graph/test" also fails with this error. Any help would be much appreciated! greetings, Chris

1 0

Re: [Boost-users] [Regex] Source-only link error in MSVC
by dtexg＠hush.com 20 May '09

20 May '09

On Tue, 19 May 2009 03:42:57 -0500 John Maddock <john(a)johnmaddock.co.uk> wrote: >> [...] >> error LNK2019: unresolved external symbol >> [...] >> (?do_assign@?$basic_regex@DU?$regex_traits@DV?$ >> w32_regex_traits@D@boost@@@boost@@@boost@@ >> AAEAAV12@PBD0I@Z) referenced >> [...] >> (?assign@?$basic_regex@DU?$regex_traits@DV?$ >> w32_regex_traits@D@boost@@@boost@@@boost@@ >> QAEAAV12@PBD0I@Z) >> [...] >> I wrote the following code: >> >> #define BOOST_REGEX_SOURCE >> #include <boost/regex.hpp> >> >> int main() >> { >> boost::regex r("a"); >> return 0; >> } >> >> Everything compiles, but linking fails with the above error. >> [...] >> >> In a post from 2006, John wrote, "if you are doing something, >> umm, strange, and you can't track down the cause of the error, >> you can always just add the regex source directly to your >> application," but I've not seen this successfully demonstrated >> in Visual C++. I'm just looking for an absolutely parsimonious >> way to add Boost Regex to a VC++ project. > >Well, you need to do what I said, not what you did ;-) > >Take all the .cpp files in libs/regex/src and add them to your >console project's source, build and voila, it should all work. > >BTW use BOOST_REGEX_NO_LIB not BOOST_REGEX_SOURCE in your source >files, as the latter is for private regex use only and most >certainly does not do what you think it does. > >In point of fact I do this all the time, so it definitely does >work. > >HTH, John. My embarrassment is outstripped only by my gratitude. That did help completely, so I just want to confirm it for indexing and posterity: I changed the #define BOOST_REGEX_SOURCE to BOOST_REGEX_NO_LIB (or BOOST_ALL_NO_LIB) instead, then right- clicked on Source Files, chose Add -> Existing Item, and added all the .cpp files in libs/regex/source. When I built, it compiled all those .cpp files (where obviously it wasn't before), then linked successfully and everything just worked and everyone lived happily ever after. Sadly, I never would have figured that out. Thank you very much! /Eric -- Wanna lose weight? Weight Loss Programs that work. Click here. http://tagline.hushmail.com/fc/BLSrjkqbR3flkqN6HzqHIZfbT5vIrA1g60gAsnOhshGi…

2 1

[zip_iterator] Which header files?
by Robert Jones 20 May '09

20 May '09

The iterators documentation seems a bit thin! Does it say anywhere which header files are required to use zip_itrerators? Also, is the use of zip_itertor and others still very much current practise, or are there other techniques which tend to be used instead? Thanks, Rob.

1 0

[boost 1.39][Serialization] BOOST_CLASS_EXPORT_GUID macro failed to compile with C2888 error under MSVS.
by Vitaly Dolya 19 May '09

19 May '09

Hi, All! Since I had no answer on my previous post, I'll try to clarify the problem here. So, I use a macro BOOST_CLASS_EXPORT_GUID from Boost.Serialization library. And everything was just fine until boost 1.39 came up to stage recently. Now the same very of my code, I've been using for so long (starting with boost 1.35 I guess), failed to compile under MS Visual Studio 2008 SP1. Compiler gives me an error error C2888: 'boost::archive::detail::`anonymous-namespace'::init_guid<CSomeClass>' : symbol cannot be defined within namespace 'anonymous-namespace'. The error occurs when export.hpp file (contains BOOST_CLASS_EXPORT_GUID) is included to stdafx.h (to precompiled headers) and the macro is used in any other cpp-file. I guess the problem is that 'init_guid' structure declaration (declared inside anonymous namespace) and definition appear in different translation units. So, it looks like there is a bug in boost 1.39 to me. Am I right? And what should I do? Should I file a bug? Thanks in advance.

2 6

using bind with operator new
by Archie14 19 May '09

19 May '09

I am trying to use bind to generate ptr_vector list of objects from another vector #include <boost/shared_ptr.hpp> #include <boost/bind.hpp> class a { }; class b { public: b(a* val) : _val(val) { } private: boost::shared_ptr<a> _val; }; void test() { boost::ptr_vector<a> lsta; boost::ptr_vector<b> lstb; typedef boost::ptr_vector<b> listofb; // create lstb by applying new to each element of lsta and pushing it to lstb. Fails. I will appreciate any help with this. std::for_each (lsta.begin(), lsta.end, boost::bind((&listofb::push_back), lstb, boost::bind((&operator new), b, _1) ); }

2 1

boost::interprocess::managed_shared_memory uses C: - Need to change this
by Moorthi 19 May '09

19 May '09

I am using boost::interprocess::managed_shared_memory to share the data between processes. But the shared memory internally creates memory mapped file in C: dirve. if anyone knows how to change the configuration to use some other drive instead of C: as my production environment doesnt have much space in C: drive. -- View this message in context: http://www.nabble.com/boost%3A%3Ainterprocess%3A%3Amanaged_shared_memory-us… Sent from the Boost - Users mailing list archive at Nabble.com.

2 2

directory_iterator ++operator never return
by gtsml owevwr 19 May '09

19 May '09

Hi, I'm getting a list of the file in a directory like that: fs::directory_iterator end ; for( fs::directory_iterator iter(myfilepath) ; iter != end ; ++iter ) {...} And at some point in this for loop, the ++operator never return, see callstack (single-thread build). I found a related thread here: http://marc.info/?l=boost-users&m=118835374908173&w=2 I compiled boost in threaded and non-threaded mode and still have the same pb in both mode... This bug reproduce with iphone os21 and os3.0, boost 1.38.0 (multi/single thread). Anybody? Thanks, Frank #0 0x31d41270 in swtch_pri #1 0x31dd19cc in sched_yield #2 0x000ac1d4 in boost::detail::yield at yield_k.hpp:110 #3 0x000adb54 in scoped_lock [inlined] at detail:45 #4 0x000adb54 in boost::detail::spinlock::lock at detail:32 #5 0x000adb54 in atomic_exchange_and_add [inlined] at detail:61 #6 0x000adb54 in boost::detail::sp_counted_base::release at detail:100 #7 0x000adb54 in boost::shared_ptr<boost::filesystem::detail::dir_itr_imp<boost::filesystem::basic_path<std::string, boost::filesystem::path_traits> > >::reset at detail:216 #8 0x000adb54 in boost::filesystem::basic_directory_iterator<boost::filesystem::basic_path<std::string, boost::filesystem::path_traits> >::increment #9 0x0009cc44 in boost::iterator_core_access::increment<boost::filesystem::basic_directory_iterator<boost::filesystem::basic_path<std::string, boost::filesystem::path_traits> > > at iterator_facade.hpp:522 #10 0x0009cc64 in boost::iterator_facade<boost::filesystem::basic_directory_iterator<boost::filesystem::basic_path<std::string, boost::filesystem::path_traits> >, boost::filesystem::basic_directory_entry<boost::filesystem::basic_path<std::string, boost::filesystem::path_traits> >, boost::single_pass_traversal_tag, boost::filesystem::basic_directory_entry<boost::filesystem::basic_path<std::string, boost::filesystem::path_traits> >&, int>::operator++ at iterator_facade.hpp:664

2 7

building boost 1.39 on iphone os3.0
by gtsml owevwr 19 May '09

19 May '09

Hi, I understand a lot of effort is being spent on making boost easy to compile and so on but as far as my experience goes, that's always a bit painful... So here's what I'm going to do as a background task: I'm going to play it dumb (eg. not spend 30 minutes on the web everytime I hit something) and tell you all about it :-) My goal here is to compile boost 1.39 on the iphone os3.0. Note that I already compiled 1.38 on os21 and os30 so I have gathered some experience already, but as I said, I'm playing it dumb. So: - getting boost from the web, uncompress and so on - boostrap, so far so good - add my user-config with: ################################################################# # Compiler configuration using darwin : 4.0.1~iphone : /Developer/Platforms/iPhoneOS.platform/Developer/usr/bin/gcc-4.2 -arch arm : <striper> : <architecture>arm <target-os>iphone <macosx-version>iphone-3.0 ; using darwin : 4.0.1~iphonesim : /Developer/Platforms/iPhoneSimulator.platform/Developer/usr/bin/gcc-4.2 -arch i386 : <striper> : <architecture>x86 <target-os>iphone <macosx-version>iphonesim-3.0 ; ################################################################# - and type ./bjam BUT, here we go, first problem, my user-config.jam seems to be ignored. the compiler used is "g++": "/Developer/Platforms/iPhoneOS.platform/Developer/usr/bin/gcc-4.2" still the doc says: On startup, Boost.Build searches and reads two configuration files: site-config.jam and user-config.jam. On top of that, the project-config.jam file says: # Compiler configuration. This definition will be used unless # you already have defined some toolsets in your user-config.jam # file. if ! darwin in [ feature.values <toolset> ] { using darwin ; } Anybody get any idea? Frank

2 4