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

boost::shared_ptr with operator[]
by Kraus Philipp 29 Sep '11

29 Sep '11

Hello, I'm using boost::shared_ptr, my class overloads the operator[]. How can I access to an element of a shared pointer? class myclass { public : operator[](std::size_t) } boost::shared_ptr<myclass> x( new myclass ); How can I get an access to the [] operator? I have tried it with (*x)[5] but I'll get a "segmentation fault" error Thanks Phil

4 3

Assignment of ptr_container<A>::iterator value to ptr_container<B>::iterator variable
by Rob Desbois 29 Sep '11

29 Sep '11

So I've got a few ptr_container instances floating around, and have just tracked down a bug caused by (my) naff code essentially akin to the following: ptr_vector<int> vi; ptr_vector<string> vs; bool eq = (vi.begin() == vs.begin()); After digging I discovered the operator in question in void_ptr_iterator.hpp: template< class VoidIterT, class T, class VoidIterU, class U > inline bool operator==( const void_ptr_iterator<VoidIterT,T>& l, const void_ptr_iterator<VoidIterU,U>& r ) What is the reason for implementing comparison between different iterator types? I have a few guesses but unsure of whether they're sensible ones or not! Of course disallowing comparison between differing pointer types won't save people from comparing iterators from different containers, but I think it's a little too easy to fall into this trap, and difficult to spot the error once written in. (The equality comparison was an unlikely case, and fortunately found it during thorough unit testing :-) ) Is there any sensible way to suppress this behaviour, or are there any comments on why I might be being a muppet? Cheers, Rob -- Rob Desbois Eml: rob.desbois(a)gmail.com

2 2

Re: [Boost-users] Compilation of boost on HP-UX 11iv3 using aCC
by Avi Bahra 29 Sep '11

29 Sep '11

>>All, >> >>I think HP develops/maintain compilers but do not test them against boost. >>Compilation of code in many .hpp files in boost fails with aCC >>compiler's latest version i.e. aCC: HP C/aC++ B3910B A.06.26 [Apr 12 2011]. >> >>Anybody doing serious development on c++ on HP-UX would be left with >>only choice : gcc ( which is too not officially supported on HP-UX). >>So, I wonder if one can do some serious work on HP-UX involving boost >>libraries...... Can you be a bit more specific, about what boost libs you are using and what your compilation failures you are getting ? Up to boost 1.39 I had used gcc. I have not tested this with later version of boost or gcc. As I have stated in my previous email, I have successfully used boost 1.45 on HP-UX. The only problems I have faced is with utf8_codecvt_facet.cpp, where I had to hack some code out. Apart from that it was ok. ( You will have to live with the mega long compilation times however. My own opinion is that HP-UX is crap as a development environment, I develop on linux, and then port to HP-UX) Because of the large number of warning I suppress the following in my jamfile. # suppress - warning #2236-D: controlling expression is constant # warning #2461-D: initial value of reference to non-const must be an lvalue # warning #2191-D: type qualifier is meaningless on cast type # warning #2815-D: type qualifier on return type is meaningless # warning #4232-D: conversion from "boost::python::converter::rvalue_from_python_stage1_data *" to a more strictly aligned type "boost::python::converter::rvalue_from_python_storage<boost::shared_ptr<CtsApi>> *" may cause misaligned access # warning #4189-D: warning #4189-D: a non-POD class type used in offsetof macro BOOST_STATIC_ASSERT(BOOST_PYTHON_OFFSETOF(rvalue_from_python_storage<T>,stage1) == 0); using acc : : : <compileflags>+W2236,2461,2191,2815,4232,4189 ; Hope that helps> Best regards, Ta, Avi

1 0

bjam: how to rename during install
by Avi Bahra 29 Sep '11

29 Sep '11

I am building a python extension : python-extension my_ext : [ glob src/*.cpp ] ; However during install I want to rename this to '_my_ext' It appears that there is no way for install to rename the target. I can not find any reference to renaming in the documentation. Can any one advice on an alternative approach to achieve the same effect ? Best regards, Ta, Avi

1 0

Base64 encoding
by John M. Dlugosz 28 Sep '11

28 Sep '11

I followed a recipe for using Boost libraries to perform a base64-encoding of some data, as explained here (http://stackoverflow.com/questions/7053538/how-do-i-encode-a-string-to-base…) However, it gives the wrong result. For example, I expect a string to end with "…8EAA==" and instead I get a total length two chars shorter, ending with "…8EAM". More use of google turns up that boost::archive::iterators::base64_from_binary only works for inputs that are a multiple of three bytes, and I can't find any proper documentation on how to use this. Is there function that wraps base64_from_binary that works on any input? Or is it not worth trying to use this Boost library at all? Thanks, —John

1 0

Re: [Boost-users] Boost-users Digest, Vol 2858, Issue 2
by Alex Perry 28 Sep '11

28 Sep '11

On Wed, 28 Sep 2011 15:54:59 Christopher Pisz wrote :- > This might be more of a windows question, but... > > Everytime I use a boost template as part of a class interface that is exported > from a dll, I get this compiler warning. > > warning > C3251: 'boost_enable_shared_from_this<T>::weak_this_':class > 'boost::weak_ptr<T> ' needs to have a dll-interface to be used by clients of > class 'boost::enable_shared_from_this<T>' > > How do I solve these types of errors? Sadly with great difficulty (at least that was my experience) - Generally its so painful that its just easier not to export template code across a dll boundary. As you say its more of a windows question rather than a boost one see http://www.codesynthesis.com/~boris/blog/2010/01/18/dll-export-cxx-template… or any other result from a google search of template dll export. However if you are in full control of your project and can ensure consistent use of heap (/MD), exception handling and can always be 100% certain that all the dlls are compiled with the same version of the vs compiler you can just ignore the warnings (#pragma warning is always useful : - ) ) and things will sort of work - till they just don't at some point when someone breaks the rules above - so I wouldn't recommend it for production code. Alex

1 0

C4251 needs to have dll-interface
by Christopher Pisz 28 Sep '11

28 Sep '11

This might be more of a windows question, but... Everytime I use a boost template as part of a class interface that is exported from a dll, I get this compiler warning. For example I just cut and paste the TcpServer example from boost::asio::ip::tcp into a test project dll. I get: warning C3251: 'boost_enable_shared_from_this<T>::weak_this_':class 'boost::weak_ptr<T> ' needs to have a dll-interface to be used by clients of class 'boost::enable_shared_from_this<T>' It comes from the line: class __declspec(dllexport) TcpConnection : public boost::enable_shared_from_this<TcpConnection> { public: typedef boost::shared_ptr<TcpConnection> SmartPtr; . . . }; How do I solve these types of errors? Currently, I just include the correspnding boost header in the client project and igore the warnings, but I'd really like to get rid of them. Should I be including the boost headers and library in the client project? Doesn't the client project get what it needs from the dll, when the dll includes the headers and library?

1 0

[boost-newbie] Compiler Error on OS X 10.7.1 (Lion)
by Rob Jones 28 Sep '11

28 Sep '11

Trying to use BOOST_FOREACH() in some code and I'm getting the following errors: /Developer/SDKs/MacOSX10.7.sdk/usr/include/c++/4.2.1/bits/ios_base.h:779: error: 'std::ios_base::ios_base(const std::ios_base&)' is private /Developer/SDKs/MacOSX10.7.sdk/usr/include/c++/4.2.1/streambuf:802: error: 'std::basic_streambuf<_CharT, _Traits>& std::basic_streambuf<_CharT, _Traits>::operator=(const std::basic_streambuf<_CharT, _Traits>&) [with _CharT = char, _Traits = std::char_traits<char>]' is private The compiler is: GCC 4.2 O/S is: OS X 1.7.1 Boost 1.47.0 was installed using MacPorts. Any help / pointers here would be useful. Thanks in advance. Rob

3 3

Problem with the filesystem::exists()
by Danilo da Silveira Figueira 27 Sep '11

27 Sep '11

Hi folks, I´m stuck with an problem, when i try pass an accentued string to exists method: When i wrote in the code "C:\\ó.txt" and the file is under ANSI codification the function exists return true. But when i ask the input to the user the function return false. Ps: wstring in the basic_path template dont fix this. If someone know how and have a time to help me i will be very grateful

3 4

How do I print a standard pair?
by Krzysztof Żelechowski 27 Sep '11

27 Sep '11

The expression (cout << make_pair (0, 1)) gives me an error. Does Boost offer me a way to print a (rvalue) pair, as within an expression?

5 5