September 2009 - Boost-users

[mpl] template parameters and mpl::if_/mpl::eval_if
by Edward Diener 16 Sep '09

16 Sep '09

I know I can try this out but I will ask here anyway just to make sure my test is not being incorrectly done. Can mpl::if_/mpl::eval_if be used to pick the type for which a default template parameter is set based on some metaprogramming boolean test of a previous template parameter type ? I know that C++ allows a template parameter default to be set based on a previous template parameter, but I don't know if mpl::if_/mpl::eval_if can be used in this context.

2 2

gdb crash when loading lib with (big) Spirit grammar
by Christoph Duelli 16 Sep '09

16 Sep '09

I have a quite big Spirit (1.38) grammar inside a lib. (Running OpenSuSE 11.1 Linux, gcc 4.3.2, gdb 6.8.50; both 32 bit and 64 bit.) When trying to debug a programm linked against this lib, gdb crashes - when trying to read the lib's symbol table. (gdb) at 19281 Attaching to process 19281 Reading symbols from ....i_mepl_debug...done. Reading symbols from ....libMEPL_debug.so.6... Segmentation fault (core dumped) (Note that this is not my program crashing but gdb. And no, the program is not buggy, has not corrupted the stack or anything like that ;-) Debugging does work on older openSuSE (or SLES) releases, e.g. on openSuSE 10.3., SLES 10, SLES9. I have build and tried the latest gdb release, but it shows the same behaviour as the one part of the release. A stack trace of the gdb core showed that the crash seems to have occured in cplus_demangle_v3. I have searched google and found some postings that indicate the issue might stem from too long symbols being generated for Spirit (or Xpressive) constructs. (A quick glance on nm output shows, there are some symbols of some 1000 characters in size. Might be even longer ones.) Has anyone else stumbled upon this and maybe found out what the culprit is? A bug in gcc? gdb? name (de)mangling? Any ideas? Best regards Christoph

2 1

Re: [Boost-users] building boost 1_33_1 now 1_37_0
by Dan Smithers 16 Sep '09

16 Sep '09

>> >> How do I change the includes in jam? > > Sorry, I don't understand the above. Boost.Build does not pass any > "intel includes" nor any "gcc 2.95 includes". Can you provide the commands > lines and errors in question and explain how exactly the command lines > should be altered. I've now found that as long as I make sure that gcc 3.4.3 is ahead of 2.95 in the search path I can get past the first hurdle. The command line that comes out is "icpc" -c -xc++ -O3 -Ob2 -w1 -fPIC -DBOOST_ALL_NO_LIB=1 -DBOOST_MATH_TR1_DYN_LINK -DNDEBUG -I"." -c -o "bin.v2/libs/math/build/intel-linux-release/threading-multi/hypot.o" "libs/math/build/../src/tr1/hypot.cpp" /usr/lib/pthread.h(163): error: expected a ")" extern int pthread_create(pthread_2 *__restrict __thread, ^ I have written a short test program that includes pthread.h and that compiles. thanks dan

2 1

[Asio] Why does Boost Asio only allow 64 buffers to be sent/recei ved using scatter/gather?
by Stephens Patrick 16 Sep '09

16 Sep '09

Hi all, Apologies if this question should be directed at an Asio-specific list but I could not see one. I am curious why the Boost Asio library specifies a maximum of 64 buffers can be used for scatter/gather operations? I would like to gain a better understanding as to why this value was used. The documentation (http://www.boost.org/doc/libs/1_40_0/doc/html/boost_asio/overview/implement ation.html), and the code backs it up, seems to specify that the maximum allowed will be 'min(64, IOV_MAX)' for linux implementations. Was this value (64) chosen for some particular reason? What would the consequences be of using IOV_MAX if it was greater than 64? I see that the Windows NT, 2000, XP, 2003 and Vista implementation also has a limit of 64 so is this something limited by the OS or to keep it consistent across all the Asio implementations? I had a look around to see if there was a restriction under Windows but the only documentation I could find (http://msdn.microsoft.com/en-us/library/ms742203%28VS.85%29.aspx) mentions that it is restricted to a value of 16 under Windows 95, 98 and ME, which does tally with what the Asio documentation states in the link above. Thanks for any help, Pat. This email, including any attachment, is a confidential communication intended solely for the use of the individual or entity to whom it is addressed. It contains information which is private and may be proprietary or covered by legal professional privilege. If you have received this email in error, please notify the sender upon receipt, and immediately delete it from your system. Anything contained in this email that is not connected with the businesses of this company is neither endorsed by nor is the liability of this company. Whilst we have taken reasonable precautions to ensure that any attachment to this email has been swept for viruses, we cannot accept liability for any damage sustained as a result of software viruses, and would advise that you carry out your own virus checks before opening any attachment.

1 0

[serialization] can load_construct_data() access private members?
by Igor R 16 Sep '09

16 Sep '09

Hello, How to grant load_construct_data() the permission to access private members of the serializable class? friend class ::boost::serialization::access; // doesn't help template<class Archive> friend void ::boost::serialization::load_construct_data(Archive &ar, MyObj *t, const unsigned int file_version); // doesn't compile with MSVC9.0SP1 One of my classes can be created with a static function only: class Foo { public: shared_ptr<Foo> create(Arg a) { shared_ptr<Foo> self(new Foo(a)); self->init(); return self; } private: Foo(Arg a); void init(); }; I try to simulate this behavior in load_construct_data() function. Thanks.

1 1

[Date Time] Redundant warnings with w4 on MSVC
by Василий Старостин 16 Sep '09

16 Sep '09

Hi! When I try to compile my program by MSVC 9 using w4 option (maximum warnings), I got several warnings in this library headers. It prevents me from using w4 :(. How can I use Boost Date-Time library with MSVC w4 option and don't get lots of warnings? WBR, Vasily

2 1

Re: [Boost-users] Boost-users Digest, Vol 2119, Issue 2
by N Z 16 Sep '09

16 Sep '09

Thank you Volodya, adding /MD works. Neil > Message: 2 > Date: Tue, 15 Sep 2009 09:37:27 +0400 > From: Vladimir Prus <vladimir(a)codesourcery.com> > To: boost-users(a)lists.boost.org > Subject: Re: [Boost-users] link error on boost example using binary > library > Message-ID: <h8n96i$dhd$1(a)ger.gmane.org> > Content-Type: text/plain; charset=utf-8 > > N Z wrote: > >> Hi there, >> I am a beginner using boost. I am installing boost 1.40 on >> windows XP with sp3 and Visual Studio Standard 2008. I followed the >> "Getting Started on Windows"? section 5.2???"Or, Simplified Build From >> Source. to build my binaries". i.e. >> >> bootstrap >> .\bjam >> >> I then tried to run the example in section 6:Link Your Program to a >> Boost Library. >> I got expected results when link from the IDE, however, I can't link >> under command prompt, here's the output I got when compiling: >> ****************************************************************************** >> C:\Documents and Settings\Administrator\My Documents\Visual Studio >> 2008\Projects\boostBinaryLibExamp>cl /EHsc /I "c:\Program >> Files\boost\boost_1_40_0" boostBinaryLibExamp\boostBinaryLibExamp.cpp >> /link "c:\Program >> Files\boost\boost_1_40_0\stage\lib\libboost_regex-vc90-mt-1_40.lib" >> Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 15.00.21022.08 for 80x86 >> Copyright (C) Microsoft Corporation. ?All rights reserved. >> boostBinaryLibExamp.cpp >> Microsoft (R) Incremental Linker Version 9.00.21022.08 >> Copyright (C) Microsoft Corporation. ?All rights reserved. >> /out:boostBinaryLibExamp.exe >> "c:\Program Files\boost\boost_1_40_0\stage\lib\libboost_regex-vc90-mt-1_40.lib" >> boostBinaryLibExamp.obj >> LINK : fatal error LNK1104: cannot open file 'libboost_regex-vc90-mt-s-1_40.lib' >> ****************************************************************************** >> >> I searched in the lib directory, there's no such file there. But why I >> can link from IDE but not in command prompt? > > Because IDE uses dynamic runtime by default. Seems like command line compiler > wants static runtime. You should probably pass /MD to the compiler. Also, > make sure "c:\Program Files\boost\boost_1_40_0\stage\lib\" is in libraries > search patch. (/LIBPATH) > > - Volodya >

1 0

[units] - overriding output
by Noah Roberts 15 Sep '09

15 Sep '09

What is the correct way to override the << op for a unit so that it say prints "Pa" instead of "m^-1 kg s^-2"? I tried this and no worky: #include <boost/units/io.hpp> namespace boost { namespace units { template<> inline std::string name_string< boost::units::pressure_dimension, boost::units::si::system >(const boost::units::si::pressure&) { return "Pa"; } }} // boost::units Also, I note that the operator << in io.hpp has a comment saying "INTERNAL USE ONLY". How would you suggest then that users of the library provide for stream output without running afoul of this implementation we're not meant to use?

2 2

[serialization] how to serialize boost::function
by Igor R 15 Sep '09

15 Sep '09

Hello, Is it possible to serialize the following class: struct foo { void f1(int a) {} void f2(const std::string& s) {} void init() { using boost::bind; functors_.push_back(bind(&foo::f1, this, 1)); functors_.push_back(bind(&foo::f2, this, "test")); } std::vector<boost::function<void(void)> > functors_; }; That is, I'd like save the functors, in order to run them later, when the object is loaded from the archive. A straightforward attempt fails: struct foo { //...... friend class boost::serialization::access; template<class Archive> void serialize(Archive &ar, const unsigned int version) { ar & BOOST_SERIALIZATION_NVP(functors_); } }; 1>d:\dev\servision\projects\windows\thirdparty\boost\boost\serialization\access.hpp(109) : error C2039: 'serialize' : is not a member of 'boost::function<Signature>'

4 8

shared_ptr<T> callback
by Vjekoslav Brajkovic 15 Sep '09

15 Sep '09

Hello! :) I'm writing a write-through cache unit which returns const shared_ptr<T>. After the data has been modified and once share_ptr goes out of scope, I would like the shared_ptr destructor to call a handler which would commit the modified value to some other location -- a DHT for instance. Is there a way to supply a callback function to shared_ptr destructor? Any help would be greatly appreciated. Please let me know if the question is too ambiguous. Thanks! V

2 2