August 2013 - Boost-users - lists.preview.boost.org

[Spirit/Qi] Malformed JSON parser
by Szymon Gatner 05 Aug '13

05 Aug '13

Hi, I am trying to add a support for degenerate / corner cases for my JSON parser. First I needed support for malformed null values for object members like that: obj : { "bad_member: , "correct_member: null } I would like the parser to generate same result for both members in that case (null value). I eventually came up with custom semantic action that does the job: struct MakeMember { typedef void result_type; template <class Arg1, class Ctx> result_type operator()(Arg1& arg1, Ctx& ctx, bool& b) const { using namespace boost::fusion; auto m0 = boost::fusion::at_c<0>(arg1); auto m1 = boost::fusion::at_c<1>(arg1); ctx.attributes.car.first = m0; ctx.attributes.car.second = m1 ? *m1 : NullT(); } }; but I have a feeling that is not the best way to do it. I would rally appreciate an advice on that. Secondly I want to add support for empty array members: obj : { "empty_array" : [] } but this time I am totally lost. My array parser looks like this: array = '[' >> value % ',' >> ']'; which is only god for 1 or more element list. I tried array = '[' >> -(value % ',') >> ']'; and altho it compiled it still didn't reckognize "[]' as Array object. Full source code of parser / grammar attched. Regards, Szymon

1 0

Trying out Build on Windows, have questions
by Daryle Walker 05 Aug '13

05 Aug '13

After using a bunch of IDEs to compile my trial Boost code on Windows, I'm trying out Boost's included build system. I've been using the directions at http://www.boost.org/doc/libs/1_54_0/more/getting_started/index.html. I'm using a 32-bit Windows 8 Pro laptop. I have the MinGW-Builds version of GCC 4.8.1 installed. I downloaded a copy of Clang 3.3, built with the aforementioned GCC. I also have Visual Studio Express 2012 (v11.0) installed, the Desktop version along with the Metro and Web versions. The copy of Clang is installed into the same directories as GCC, and the "bin" sub-directory is in my PATH. I'm using a Subversion copy of the Boost libraries. Per the Getting Started instructions, I changed to the Boost Subversion directory and ran "bootstrap." I'm using a "user-config.jam" file of: //===== using gcc : 4.8 : : <cxxflags>-std=c++11 ; using msvc : 11.0 : "C:\Program Files\Microsoft Visual Studio 11.0\VC\bin\cl" ; using clang : 3.3 : : <cxxflags>-std=c++11 ; //===== Then I ran ".\b2", and the GCC version of the Boost libraries were built. Then I ran ".\b2 toolset=msvc" for the Visual Studio version. When I tried ".\b2 toolset=clang", it failed miserably since it couldn't find any of the standard headers. Now I've heard that Clang can build for Windows, but actually using it is broken (no matter how it's built). Is there a way to adjust the clang entry in my "user-config.jam" so it gets all the stuff it needs from the GCC installation? I got the sample regex program to build with GCC: //===== C:\Users\Daryle\Documents\Projects>g++ -I ..\Repositories\boost-trunk example2.c pp -o example2 -L ..\Repositories\boost-trunk\stage\lib -lboost_regex-mgw48-mt-1 _55 //===== but not with MSVC: //===== C:\Users\Daryle\Documents\Projects>cl /EHsc /I ..\Repositories\boost-trunk examp le2.cpp /link /LIBPATH:C:\Users\Daryle\Documents\Repositories\boost-trunk\stage \lib Microsoft (R) C/C++ Optimizing Compiler Version 17.00.51106.1 for x86 Copyright (C) Microsoft Corporation. All rights reserved. example2.cpp Microsoft (R) Incremental Linker Version 11.00.51106.1 Copyright (C) Microsoft Corporation. All rights reserved. /out:example2.exe /LIBPATH:C:\Users\Daryle\Documents\Repositories\boost-trunk\stage\lib example2.obj LINK : fatal error LNK1104: cannot open file 'libboost_regex-vc110-mt-s-1_55.lib ' //===== When I looked at the library-storage directory, there were no variants with "-s-". Should I run "b2," which is now in my PATH, again to add the required variants? What options do I need to add to the command line? Daryle W.

2 1

boost logging to file
by Jarrett Chisholm 05 Aug '13

05 Aug '13

Thanks to everyone who's replied to my previous boost logging question - still having some issues I'm afraid :S I've been struggling with boost log for a while now - I got their simple example writing to a log file (http://boost-log.sourceforge.net/libs/log/example/doc/tutorial_file.cpp) However, when I try to copy that code into my own 'Logger' class, I can't get it to write to the log file. I can see the file `default.log` get created, but there is nothing in it. I'm on debian 7 64bit. Everything compiles fine - compile line is: g++ -o build/Logger.o -c -std=c++11 -Wall -g -O0 -DBOOST_LOG_DYN_LINK -DDEBUG src/Logger.cpp g++ -o build/logtest build/Logger.o -lboost_log -lboost_log_setup -lboost_date_time -lboost_thread -lboost_wave -lboost_regex -lboost_program_options Here's my code: `Logger.cpp` #include "Logger.h" namespace logging = boost::log; namespace sinks = boost::log::sinks; namespace src = boost::log::sources; namespace expr = boost::log::expressions; namespace attrs = boost::log::attributes; namespace keywords = boost::log::keywords; namespace dhlogging { Logger::Logger(std::string fileName) { initialize(fileName); } Logger::Logger(Logger const&) { } Logger::~Logger() { } Logger* Logger::logger_ = nullptr; Logger* Logger::getInstance(std::string logFile) { if ( Logger::logger_ == nullptr ) { logging::add_file_log( logFile ); logging::core::get()->set_filter ( logging::trivial::severity >= logging::trivial::info ); logging::add_common_attributes(); Logger::logger_ = new Logger(logFile); } return Logger::logger_; } void Logger::initialize(std::string fileName) { BOOST_LOG(log_) << "Hello, World!"; BOOST_LOG_SEV(log_, info) << "Hello, World2!"; } void Logger::logInfo(std::string message) { BOOST_LOG_SEV(log_, info) << message; } void Logger::logDebug(std::string message) { BOOST_LOG_SEV(log_, debug) << message; } void Logger::logWarn(std::string message) { BOOST_LOG_SEV(log_, warning) << message; } void Logger::logError(std::string message) { BOOST_LOG_SEV(log_, error) << message; } void Logger::logFatal(std::string message) { BOOST_LOG_SEV(log_, fatal) << message; } } int main(int, char*[]) { logging::add_common_attributes(); using namespace logging::trivial; dhlogging::Logger::getInstance()->logInfo("himom"); return 0; } `Logger.h` #ifndef LOGGER_H_ #define LOGGER_H_ #include <boost/log/core.hpp> #include <boost/log/trivial.hpp> #include <boost/log/expressions.hpp> #include <boost/log/sinks/text_file_backend.hpp> #include <boost/log/utility/setup/file.hpp> #include <boost/log/utility/setup/common_attributes.hpp> #include <boost/log/sources/severity_logger.hpp> #include <boost/log/sources/record_ostream.hpp> namespace logging = boost::log; namespace sinks = boost::log::sinks; namespace src = boost::log::sources; namespace expr = boost::log::expressions; namespace attrs = boost::log::attributes; namespace keywords = boost::log::keywords; using namespace logging::trivial; namespace dhlogging { class Logger { public: static Logger* getInstance(std::string logFile = "default.log"); void logInfo(std::string message); void logDebug(std::string message); void logWarn(std::string message); void logError(std::string message); void logFatal(std::string message); private: Logger(std::string fileName); Logger(Logger const&); Logger& operator=(Logger const&); virtual ~Logger(); void initialize(std::string fileName); src::severity_logger< severity_level > log_; static Logger* logger_; // singleton instance }; } #endif /* LOGGER_H_ */ Anyone have any ideas? Jarrett

2 4

Boost Iterator: Limiting an existing iterator
by אנטולי קרס נר 03 Aug '13

03 Aug '13

Hello, I'm writing a C++ class which has an STL container as a data member, and I'd like to write an iterator class which allows to iterate through the container. Which container is used, is implementation detail. I want the iterator to be a ForwardIterator, even if a random-access container is used, e.g. std::vector. What's the quick and simple way to define an iterator adapter class which uses the "base" iterator's functions, but only the ones supported by a forward iterator? The reason I need it is error detection. If the user gets a random-access iterator and treats it as such, the compiler won't complain, but in future versions of my code the user's code may break because random access shouldn't really be supported, e.g. if I change from vector to list. Thanks in advance, fr33domlover Hmmm... I guess using boost::iterator_adapter should help

1 0

[asio] using libgamin
by Alexander Nagel 03 Aug '13

03 Aug '13

Hello, one can use "signal_set" to connect to a SIGHUP signal and then reload a configuration file. I am thinking about a different approach. What about using libgamin to get notified if the config file changes and then reload automatically, if the config file is correct and valid. Is it possible to implement that behavior with asio? I read through the reference but I couldn't find something. I am not yet familiar with all aspects of asio. Perhaps I have overread it. Does anybody have a hint or an idea, please? regards, Alexander -- I would prefer to have encrypted communication. If you also prefer to encrypt emails, please do. Alexander Nagel Email: alexander(a)acwn.de Homepage: http://www.acwn.de/

1 0

Boost Regex link errors with clang on Mac OS
by Rodrigo Madera 02 Aug '13

02 Aug '13

Linking with Boost Regex delivers this message: Undefined symbols for architecture x86_64: "boost::regex_error::regex_error(std::string const&, boost::regex_constants::error_type, long)", referenced from: boost::re_detail::basic_regex_creator<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::fixup_recursions(boost::re_detail::re_syntax_base*) in tools_utils.o boost::re_detail::basic_regex_creator<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::create_startmaps(boost::re_detail::re_syntax_base*) in tools_utils.o boost::re_detail::basic_regex_creator<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::create_startmap(boost::re_detail::re_syntax_base*, unsigned char*, unsigned int*, unsigned char) in tools_utils.o boost::re_detail::basic_regex_parser<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >::fail(boost::regex_constants::error_type, long, std::string, long) in tools_utils.o "boost::re_detail::lookup_default_collate_name(std::string const&)", referenced from: boost::re_detail::cpp_regex_traits_implementation<char>::lookup_ collatename(char const*, char const*) const in tools_utils.o ld: symbol(s) not found for architecture x86_64 According to the Stack Overflow link [1], the problem lies in a clang bug. But I updated to the latest version I could find and it's still there. Apparently the BOOST_REGEX_NO_EXTERNAL_TEMPLATES define does work on preventing the hack on the template keyword, on boost/regex/v4/instances.hpp . Anyone else had this bug? Thanks for any help letting me compile without ugly hacks, Rodrigo I'm using Mac OS X 10.8.4, and compiling with C++11. > c++ --version Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) Target: x86_64-apple-darwin12.4.0 Thread model: posix References: [1] http://stackoverflow .com/questions/10879284/how-to-dynamically-build-and-link-boostregex -with-hidden-inlines

1 0

[multi-index] retrieve unique keys of a non-unique index
by Rory McCarthy 02 Aug '13

02 Aug '13

Hi everyone (my apologies if this shows as a report), In the following example is there a more efficient way to retrieve the unique keys of a non-unique index? #include <boost/multi_index_container.hpp> #include <boost/multi_index/member.hpp> #include <boost/multi_index/hashed_index.hpp> #include <boost/tuple/tuple.hpp> #include <algorithm> #include <iostream> #include <iterator> #include <string> using boost::multi_index_container; using namespace boost::multi_index; struct name {}; struct value {}; struct guid {}; struct TaggedNodeEntry { int name; int value; std::string guid; TaggedNodeEntry(int name_, int value_, std::string guid_) : name(name_), value(value_), guid(guid_) {} friend std::ostream& operator<<(std::ostream& os, const TaggedNodeEntry& e) { os << e.name << " " << e.value << " " << e.guid; return os; } }; typedef multi_index_container< TaggedNodeEntry, indexed_by< hashed_non_unique< tag<name>, BOOST_MULTI_INDEX_MEMBER(TaggedNodeEntry, int, name) >, hashed_non_unique< tag<value>, BOOST_MULTI_INDEX_MEMBER(TaggedNodeEntry, int, value) >, hashed_non_unique< tag<guid>, BOOST_MULTI_INDEX_MEMBER(TaggedNodeEntry, std::string, guid) > > > TaggedNodes; int main() { TaggedNodes taggedNodes; taggedNodes.insert(TaggedNodeEntry(1, 1, "one-one")); taggedNodes.insert(TaggedNodeEntry(1, 2, "one-two")); taggedNodes.insert(TaggedNodeEntry(1, 3, "one-three")); taggedNodes.insert(TaggedNodeEntry(2, 1, "two-one")); taggedNodes.insert(TaggedNodeEntry(2, 2, "two-two")); taggedNodes.insert(TaggedNodeEntry(3, 6, "three-six")); std::cout << std::endl << "Unique names :-" << std::endl; boost::multi_index::index<TaggedNodes, name>::type::iterator ic0 = get<name>(taggedNodes).begin(); boost::multi_index::index<TaggedNodes, name>::type::iterator ic1 = get<name>(taggedNodes).end(); while (ic0 != ic1) { std::cout << ic0->name << std::endl; ic0 = get<name>(taggedNodes).equal_range(ic0->name).second; } return 0; }

3 4

[Wave] Making wave driver work with 0 arguments passed as __VA_ARGS__
by Szymon Kurek 02 Aug '13

02 Aug '13

Hello everyone, I'm porting some g++ code to MSVC, and I'm trying to use the Boost.Wave driver executable as a preprocessor instead of Visual Studio's. Boost 1.54.0. I'm generally successful, but I'm having trouble with macros like these: #define LOG(msg, ...) whatever Usage examples: LOG("Hello world!"); LOG("Hello and print out some variables", foo, bar); Both g++ and MSVC have custom extensions that make these examples work, but Wave is pretty strict and follows the standard to the letter, so gives the error message "not enough arguments for macro LOG" for the first example. I know it's possible to rewrite each macro like this in a standard-compliant way, but it's a lot of work and gets complicated if I want recursive macro invocation. Is it possible to enable this behaviour in wave somehow? Do I need to compile Wave Driver with a specific option? I searched through the docs and tried looking at the code, but I can't find anything. Regards, Szymon Kurek

2 1

[Interprocess] Page default
by Oodini 02 Aug '13

02 Aug '13

Hello, I use Boost.Interprocess on Windows 2003 Server. I am experiencing performances troubles. I am trying to understand what happens, and arrived on Microsoft web pages. I seems that on Windows, shared memory is implemented only through mapped files, and that only a portion of the shared memory is physically in the RAM (this limited number of pages is called "working set"). The Windows API provides some functions for this working set : http://msdn.microsoft.com/en-us/library/windows/desktop/cc441804%28v=vs.85%… I would have liked Interprocess wraps the functions GetProcessMemoryInfo() and SetProcessWorkingSetSize(), so that the full shared memory is put in RAM. But for now, I fear it is impossible to have with Interprocess performances similar to the ones provided by the alternative : data in the process memory.

2 2

[boost::gil] Type conversion in Generic Image Library
by Nestor Morales Hernández 01 Aug '13

01 Aug '13

Hello, I'm a newbie in this library and I feel a little bit lost. I have the following problem: I currently have an image of the type ` boost::gil::rgb8c_view_t`. I need to pass it to a function (from a code I don't mantain) with this prototype: void function(const boost::gil::rgb8c_view_t& input, const int index, const boost::gil::rgb8c_view_t::view_t &output) I have created an output image by using the following syntax: boost::gil::rgb8c_view_t::view_t output(input._dynamic_cast<boost::gil::rgb8c_view_t::view_t>()); At this point, the compiler accepts the image output as input for the function. However, I need the image back to the `boost::gil::rgb8c_view_t` type. The question is: - Is it a correct way to allocate the image output? - How do I transform it back to the type boost::gil::rgb8c_view_t? Thank you in advance, Néstor

3 4