- Boost-users - lists.boost.org

Memory usage of adjacency_list
by Ed Linde 24 Oct '12

24 Oct '12

Hi, Is it possible to get a ballpark figure of the maximum memory required for an adjacency_list, given its settings? I need these statistics to report how much memory is being used initially when reading in a large graph and then after vertex removal operations and edge removal operations, how much reduction is got in this memory finally. Do I have to write something of my own to do this? Thanks, Ed -- View this message in context: http://boost.2283326.n4.nabble.com/Memory-usage-of-adjacency-list-tp4637051… Sent from the Boost - Users mailing list archive at Nabble.com.

2 4

Is it possible to put a std::unique_ptr into a mulit_index container?
by Josh Quigley 24 Oct '12

24 Oct '12

Hi all, Is it possible to put a std::unique_ptr into a mulit_index container? In particular, the following fails. typedef std::pair<int, std::unique_ptr<int> > ValueT; typedef boost::multi_index_container< ValueT, boost::multi_index::indexed_by< boost::multi_index::hashed_unique< BOOST_MULTI_INDEX_MEMBER(ValueT, int, first) > > > HashMap; HashMap map; auto a = std::make_pair(5, std::unique_ptr<int>(new int(5))); map.insert(std::move(b)); // This line fails to compile (Visual Studio 2012) I'm guessing the answer is 'No' and it's because the rvalue reference overloads have not been done, but I'd like confirmation (or even better, to be told I'm wrong!). Thanks, Josh.

2 1

[spirit] Parsing quoted string
by Giampiero Gabbiani 24 Oct '12

24 Oct '12

Hi together, I have to parse the following row: "a_token": true|false, the last comma is optional. Testing the following: bool value; test_parser(" \"orig_bool\": true,", qi::lit('"') >> *(ascii::alnum | qi::char_('_')) >> qi::lit('"') >> qi::lit(':') >> boost::spirit::qi::bool_[phx::ref(value)=_1] >> -(qi::lit(','))); it works, but with the following instead the program compile but hangs ... bool value; test_parser(" \"orig_bool\": true,", qi::lit('"') >> *(!qi::lit('"')) >> qi::lit('"') >> qi::lit(':') >> boost::spirit::qi::bool_[phx::ref(value)=_1] >> -(qi::lit(','))); the only difference is the way by means I get the content of the quoted string, actually instead of using *(ascii::alnum | qi::char_('_')) I'd like to use *(!qi::lit('"')) that - imho - should be better but it actually hangs. What/where am I actually missing/wrong? Many thanks in advance Giampiero

2 1

Re: [Boost-users] asio: how to handle disk IO, recommendations
by Avi Bahra 24 Oct '12

24 Oct '12

> Currently I has a single threaded server based on boost ASIO. > This works correctly most of the time. > However on systems where i know that the disk's are slow. > , the server grinds to a halt, when dealing with the blocking > disk IO. > > I could not find any examples or documentation in ASIO, in > how to get round these issues. It appears that ASIO does > not provide any support for async' disk IO. > > So I assume this is handled by use of threads ? > Are there any examples of using ASIO with threaded disk IO, > since this must be very common use case ? > > Is the strand functionality appropriate for handing disk IO, > in an async' manner. > > Any recommendations/advice on this issues would be > greatly appreciated. > > Best regards, > Ta, > Avi > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users > >Hi Avi, > >If your server performs short read/writ frequently and data to be transfer >to network depend on these operations, then multi-threading does not help a >lot here. I would suggest you to recheck your I/O strategies and make sure >they are optimized. e.g. buffering, cache, .. etc. If they are perfect but >performance still bother you, then consider to use memory mapped file or >libaio. > >Best, >Acer. In my case the ASIO server needs to only write to a local disk.(no reads) One write is relatively short, i.e to a log file, to record all requests, the other is to write periodically a in memory node tree. This can vary between 1-100MB. My thinking was that if I used a thread reserved for writing the node tree, it will not block the server from handling requests, on slow file systems. I will look into memory mapped files and libaio. Ta, Avi

2 1

serialization
by Larry E. Ramey 24 Oct '12

24 Oct '12

I've got a system where class A { A(); A(const string&); virtual ~A(); }; class C : public A { C(); C(const A&, float); virtual ~C(); }; BOOST_CLASS_EXPORT( A ) BOOST_CLASS_EXPORT( C ) main() { ... const A* a = new a("stuff"); const C* c = new c(*a, 52.54390f); { ofstream ofs("dump.bin"); boost::archive::text_oarchive oa(ofs); oa<<a<<c; } ... A* a2=NULL; C* c2=NULL; try { ifstream ifs("dump.bin"); boost::archive::text_iarchive ia(ifs); ia>>a2; ia>>c2; } catch(boost::archive::archive_exception& ae) { cout<<ae.what(); } } This works just fine. If I replace A* a2=NULL; C* c2=NULL; with A* a2=NULL; A* c2=NULL; Then it throws an unregistered_class exception. If I change the iarchive to take: const A* a = new A("I'm cool"); const A* c = new C(*a,54.2f); then I can unarchive it with a pair of A*. Is this by design? Seriously? I can't save derived pointers and unarchive base pointers? Larry

3 4

adding 'blackberry' to boost build
by Gottlob Frege 23 Oct '12

23 Oct '12

I work for RIM (ie BlackBerry), and am trying to make using boost as easy as possible to use for BB developers. So I would like to make building boost for BB as easy as possible. I would also like to make building boost work "as expected" for anyone who understands Boost.Build. It is the "as expected" part that I'm not sure of. ie should I just offer a blackberry-config.jam? should Boost.Build be changed to auot-detect the BB NDK? should blackberry be a toolset? a target-os? ...? how would you *expect* it hook into Boost.Build? (And note that I expect that developers *also* use the same boost for Windows, etc, so I don't want to replace/affect the simple build for the host machine.) - And finally, it would be nice if this was upstreamed back into boost build (if people think that makes sense), so that boost "just works" on BlackBerry. Here's what I have so far: from the command line, you run something like: b2 install --prefix=C:\boost\ --user-config=blackberry-config.jam --layout=system toolset=gcc target-os=qnxnto I'd prefer it if they didn't need to type so much. And the config file looks like basically the rest of this email: import os ; # these env vars are set by installing the BB NDK QNX_TARGET_ARCH = [ os.environ QNX_TARGET_ARCH ] ; QNX_TARGET = [ os.environ QNX_TARGET ] ; QNX_HOST = [ os.environ QNX_HOST ] ; if ! $(QNX_TARGET) || ! $(QNX_HOST) { ECHO error: ; ECHO error: BlackBerry build of Boost requires QNX_TARGET and QNX_HOST env variables. Did you run bbndk-env.bat? ; ECHO error: ; EXIT ; } if $(QNX_TARGET_ARCH) = x86 { dashVOpt = gcc_ntox86_cpp ; linkArchDir = x86 ; } else { dashVOpt = gcc_ntoarmv7le_cpp ; linkArchDir = armle-v7 ; } using qcc # Version : # C++-compile-command : $(QNX_HOST)/usr/bin/QCC -V$(dashVOpt) # Compiler options : <compileflags>"-D__QNX__" <compileflags>"-D__QNXNTO__" # Add -D__USE_ISOC99 for rounding control mechanism (boost/numeric/interval/hw_rounding.hpp) <compileflags>"-D__USE_ISOC99" <compileflags>"-I$(QNX_TARGET)/usr/include" # Prevent this from being output: "note: the mangling of 'va_list' has changed in GCC 4.4" <compileflags>"-Wno-psabi" # Add -fPIC to link without error to LibTestNativeLibTuTest.o (for statechart tests) <compileflags>"-fPIC" # Requested by BBM team <compileflags>"-fno-strict-aliasing" # Functions like sync_val_compare_and_swap and sync_fetch_and_add aren't available <compileflags>"-DBOOST_SP_NO_SYNC" <linkflags>"-L$(QNX_TARGET)/$(linkArchDir)/usr/lib -L$(QNX_TARGET)/$(linkArchDir)/lib" ; import option ; option.set keep-going : false ; ------------------------ Thanks in advance, Tony

2 2

building universal boost binaries (MacOS)
by Nick Zavaritsky 23 Oct '12

23 Oct '12

Hi! Boost.build docs says that to build 'fat' binaries on the Mac two options exist: architecture=combined and/or address-model=32_64. The first one builds a binary with PowerPC and x86 code. The second one builds a binary with 32 and 64bit code. The doc further states that architectire= and address-model= are independent: one can use the first one alone or the second one alone or both simultaneously yielding different results. (http://www.boost.org/boost-build2/doc/html/bbv2/reference/tools.html#bbv2.r…) I would like to use address-mode=x86_64 to build x86/x86_64 fat binaries of boost libraries. Unfortunately with boost 1.51 it doesn't work this way — in order to get fat binaries one has to specify both architecture=combined and address-mode=32_64. Since we don't want power pc support we strip binaries from ppc code using lipo. This is a very suboptimal solution since we compile ppc variant anyway and that takes time. Even worse is that the modern toolchain doesn't support PowerPC hence we have to compile Boost using older toolchain. Currently address-model=32_64 works only if architecture=combined was specified as well (tried with Xcode 4.5 and 3.2.5). Is this really a bug in Boost.build? If that is by design what is the method to build boost fat binaries with modern toolchain lacking ppc support?

1 0

Re: [Boost-users] [EXTERNAL] Re: [MPI] missing symbols using serialization's static library design
by Robert Ramey 23 Oct '12

23 Oct '12

"Barrett, Brian W" <bwbarre(a)sandia.gov> wrote in message news:<69A29AB53D57F54D81061A9E4E45B8FD239CC111(a)EXMB01.srn.sandia.gov>... > On 10/17/12 1:00 PM, "Robert Ramey" <ramey(a)rrsd.com> wrote: > > >Barrett, Brian W wrote: > >> > Unfortunately, I'm barely competent in using Boost.MPI and > Boost.Serialization. I'm definitely not in a position to be able to apply > a patch. > > What I'm getting from your responses and the general silence elsewhere is > that the right answer is to deal with the incompatibilities between > versions of Boost and explicitly instantiate all the internal MPI > archivers and hope for the best. It's a little painful, but I think we > can do that. Well, your organization could contract with me to address this for you. I've done special work for several companies which use the serialization library. This work has been contributed to boost and become part of the distributed package. So a) The organization gets what it needs for a reasonable cost. b) Any generally useful enhancements get included in the library So it helps everyone all around. Thanks in advance for considering this. Robert Ramey > > Thanks, > > Brian > > -- > Brian W. Barrett > Scalable System Software Group > Sandia National Laboratories

3 4

[system] symbol visibility and GCC vs. Clang
by James Walker 23 Oct '12

23 Oct '12

When trying to link a bunch of libraries on OS X, I got this linker warning: ld: warning: direct access in boost::system::system_error::~system_error() to global weak symbol vtable for boost::system::system_error means the weak symbol cannot be overridden at runtime. This was likely caused by different translation units being compiled with different visibility settings. As far as I can tell, everything is compiled using the flags -fvisibility=hidden -fvisibility-inlines-hidden. I'm wondering whether the problem is that some of the libraries were built using gcc and some were built using Clang. In boost/system/system_error.hpp, the declaration of boost::system::system_error is decorated with BOOST_SYMBOL_VISIBLE. In boost/config/compiler/gcc.hpp, BOOST_SYMBOL_VISIBLE is defined to be __attribute__((visibility("default"))), whereas boost/config/compiler/clang.hpp does not define BOOST_SYMBOL_VISIBLE. -- James W. Walker, Innoventive Software LLC <http://www.frameforge3d.com/>

2 2

Re: [Boost-users] Interesting article on stack-based TMP
by Thorsten Ottosen 23 Oct '12

23 Oct '12

On 23-10-2012 15:05, Larry Evans wrote: > On 10/22/12 07:26, Giovanni Piero Deretta wrote: > [snip] >> Now, to make this thread on-topic: is anybody working on a C++11 >> rewrite of Boost.MPL? >> > Anyone attempting such a rewrite and planning to use the > variadic template features might want to start from: > > http://svn.boost.org/svn/boost/sandbox/variadic_templates/boost/mpl/ > > However, using variadic templates may cause a slowdown in the > compile times. IIRC, a variadic template version of proto was > developed but was abandoned in favor of using: > > http://www.boost.org/doc/libs/1_51_0/libs/preprocessor/doc/index.html > > because of large compile times. Hm. This is really sad if variadic templates can't be used for the thing they were created for. Is this an inherent problem for large-scale use of variadic templates? -Thorsten

2 1