October 2011 - Boost-users - lists.preview.boost.org

Common parameter on command line and in a config file with sections
by Gruhn, Dan 01 Nov '11

01 Nov '11

Greetings, I would like to have a parameter "expire-days" that has a default and can be set via the command line --expire-days|-e <days> and a config file [foo] expire-days = <days> It would be most elegant to be able to specify things like this (notice the automatic storage to the expireDays variable): namespace bpo = boost::program_options; int expireDays = 30; bpo::options_description desc("Allowed options"); desc.add_options() ("expire-days,e", bpo::value<int>(&expireDays)->default_value(expireDays), "Number of days before files expire") ; bpo::options_description configFileDesc; configFileDesc.add_options() ("foo.expire-days", bpo::value<int>(&expireDays), "") ; bpo::variables_map varMap; try { // Attempt to parse the configuration file. std::ifstream cfg("/tmp/all.cfg"); if (cfg.is_open()) { std::cout << "Opened /tmp/all.cfg" << std::endl; static const bool allowUnregistered = true; bpo::store(bpo::parse_config_file(cfg , configFileDesc, allowUnregistered), varMap); } // Parse the command line bpo::store(bpo::parse_command_line(argc, argv, desc), varMap); bpo::notify(varMap); } std::cout << "Expire days: " << expireDays << std::endl; My code may be strange but I am trying to find the right order and combination of adding options, parsing, storing, and notifying that gives the command line priority over the config file which in turn has priority over the 30 day default. Oddly enough, it doesn't seem to make much difference which order I use, the config file seems to have precedence This is not what I expected so perhaps I'm doing something wrong but I can't find it. I know that I can accomplish this via extra code on my part, but I'm trying to come up with a general scheme that would work for multiple programs with multiple variables using a common config file and the simpler the better. The documentation mentions using sections in config files but doesn't give much detail. Can anyone help me here or at least describe for me which of add_options, parse_*, store, and notify does the following: 1) Sets the default value 2) Marks that a value has been set from the command line/config file. 3) Actually does the store to the given variable. 4) What does notify do? I've tried (and continue to try) looking through the code but am having problems find the right place examine. I'll take pointers in that area as well. This does not seem like an exotic use case. Has anyone else accomplished this or something like it? As always, any help of any kind will be greatly appreciated. Dan

3 2

How to build the bjam tool on Windows?
by Congmin min 31 Oct '11

31 Oct '11

Hello. I just downloaded the boost library. I want to build bjam tool on my Windows machine. I need to get boost work for my eclipse IDE. But why I can't find the jam directory under boost/tools? How can I do that? Thanks.

3 2

Re: [Boost-users] [Tokenizer] Compile error with lots of missing binary operator statments
by Collin Day 31 Oct '11

31 Oct '11

Previously, I had posted what I thought was a bug. I did not think to do this originally, but I just tried a simple Hello World program that included boost/tokenizer.hpp. It compiled fine. I then looked at where the sample code was pulling headers from, and it appears that the download included old boost headers it was including instead of 1.47.0. I completley missed this. I apologize - and take back the error - there is none :) Thanks!

1 0

boost random uniform real infinite loop
by Peter Man 31 Oct '11

31 Oct '11

Hi, Recently, I've encountered a problem with the generate_uniform_real function (the is_integral version) in boost/random/uniform_real_distribution.hpp. A bug in my code meant that I was trying to generate a uniform random number where the min and max bounds were exactly equal. However, this meant that when on entering the function stated earlier, an infinite loop is never exited from! Is this a bug in boost::random, or have I missed something? Thanks Peter

2 1

sending large serialized objects over network using MPI
by Jonathan Bishop 31 Oct '11

31 Oct '11

HI, I am involved in a project which requires the transport of large C++ instances over MPI communication channels. I am using boost serialization to pack and unpack the objects into a buffer. The whole buffer is sent in one shot. Of course for large objects this can be problematic. I have two questions... 1) is there any way to tell beforehand how large the buffer needs to be? In other words, can you run serialization in a counting only mode? 2) is there an archive implemented which streams across MPI communication channels (not sure this is the right terminology.)? Seems like it might be useful. This might be a better question for the boost MPI team, though. Thanks, Jon

2 1

[Signals] trackable detection for user defined callabcks
by Szymon Gatner 30 Oct '11

30 Oct '11

Hi, Signals docs say: "Warning: User-defined function objects and function objects from other libraries (e.g., Boost.Function or Boost.Lambda) do not implement the required interfaces for trackable object detection, and will silently ignore any bound trackable objects. Future versions of the Boost libraries will address this limitation." My questions: 1) What is the required interface for user-defined callbacks? 2) Is it still true that it will be addressed in future versions of Boost.Function? Regards, Simon -- Szymon Gatner The Lordz Games Studio www.thelordzgamesstudio.com

3 7

[boost] [C++ Now! 2012] Call for Submissions
by Brian Wood 30 Oct '11

30 Oct '11

Hartmut Kaiser: > The C++ Now! Conference is dedicated to discussion and education > about C++, an open and free language and standard. It costs a lot to buy a copy of the standard now! Hope conference goes well and maybe I can attend someday. -- Brian Wood Ebenezer Enterprises http://webEbenezer.net

1 0

Re: [Boost-users] sending large serialized objects over network
by Brian Wood 30 Oct '11

30 Oct '11

Jonathan Bishop: > HI, > I am involved in a project which requires the transport of large C++ > instances over MPI communication channels. I am using boost serialization to > pack and unpack the objects into a buffer. The whole buffer is sent in one > shot. > > Of course for large objects this can be problematic. I have two questions... > > 1) is there any way to tell beforehand how large the buffer needs to be? In > other words, can you run serialization in a counting only mode? I'm not sure about the Boost Serialization library, but the C++ Middleware Writer counts the number of bytes in a message before any is marshalled. Here's an example void remote_messages::Marshal (SendCompressedBuffer& buf , marshalling_integer const& abt1 , marshalling_integer const& abt2 , cmw_user_input const& abt3 ) { Counter cntr(msg_length_max); abt1.CalculateMarshallingSize(cntr); abt2.CalculateMarshallingSize(cntr); abt3.CalculateMarshallingSize(cntr); cntr.Add(sizeof(msg_id_direct)); buf.Receive(&msg_id_direct, sizeof(msg_id_direct)); abt1.Marshal(buf, false); abt2.Marshal(buf, false); abt3.Marshal(buf, false); } -- Brian Wood Ebenezer Enterprises http://webEbenezer.net <http://webebenezer.net/>

1 0

split() and MSVS .NET 2003
by Larry 30 Oct '11

30 Oct '11

I have found using Boost 1.47 with MSVS .Net 2003 that if I include split() that it tends to kill the Intellisense function in the editor for a current method or function. I usually have something like typedef vector<string> Splits; Splits sv; string input_string; split(sv,input_string,is_any_of("|")); I was curious if anyone else had run across this and better yet if there was a way around it. I think I may have seen it in earlier Boost versions but my memory is fuzzy about that. There was a time when this behavior did not exist but I am not certain when and what MSVS changes may have been applied since it did work. The code compiles OK and works correctly. Larry

3 3

Build error with boost-1.45.0 in ptree_implementation.hpp
by Sunil Thomas 29 Oct '11

29 Oct '11

Hi all! I am getting the following error with boost-1.45.0 using MS VS 2010 Profession, Win64 (Windows Vista 64 bit). Appreciate if anyone knows a fix. This error was seen with older versions like 1.41 and so on; hence, it must've been fixed by now. <path\to>\boost/property_tree/detail/ptree_implementation.hpp(30) : error C2440: 'specialization' : cannot convert from 'const std::string std::_Pair_base<_Ty1,_Ty2>::* ' to 'const std::basic_string<_Elem,_Traits,_Ax> std::pair<_Ty1,_Ty2>::* ' with [ _Ty1=const std::string, _Ty2=boost::property_tree::basic_ptree<std::string,std::string> ] and [ _Elem=char, _Traits=std::char_traits<char>, _Ax=std::allocator<char>, _Ty1=const std::string, _Ty2=boost::property_tree::basic_ptree<std::string,std::string> ] Standard conversion from pointer-to-member of base to pointer-to-member of derived is not applied for template arguments <path\to>\boost/property_tree/detail/ptree_implementation.hpp(78) : see reference to class template instantiation 'boost::property_tree::basic_ptree<Key,Data>::subs' being compiled with [ Key=std::string, Data=std::string ] <path\to>\boost/foreach.hpp(265) : see reference to class template instantiation 'boost::property_tree::basic_ptree<Key,Data>::const_iterator' being compiled with [ Key=std::string, Data=std::string ] libs\graph\src\graphml.cpp(36) : see reference to class template instantiation 'boost::foreach_detail_::auto_any<T>' being compiled with [ T=boost::property_tree::basic_ptree<std::string,std::string>::const_iterator ] Appreciate any help! Thanks, --Sunil.

1 1