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

Re: [Boost-users] [Chrono] Memory leak in Chrono?
by Don 13 Feb '13

13 Feb '13

> Subject: Re: [Boost-users] [Chrono] Memory leak in Chrono? > From: "Vicente J. Botet Escriba" <vicente.botet(a)wanadoo.fr> > Date: 2/12/2013 1:17 PM > > Le 12/02/13 18:13, gast128 a écrit : >> we use Boost.Chrono too. MFC applications incorrectly report memory leaks, which can easily be solved. > Hi, > > if you have some patches to propose to avoids this incorrect reports I would try to include them. Vincente, I am using MFC, but I don't believe this is a MFC false-leak report. I added a destructor for duration_units_default_initializer_t, like this (near the end of chrono/io/duration_units.hpp): ~duration_units_default_initializer_t() { if (duration_units_default_holder<CharT>::initialized_) { delete[] duration_units_default_holder<CharT>::n_d_valid_units_; duration_units_default_holder<CharT>::n_d_valid_units_ = 0; delete[] duration_units_default_holder<CharT>::valid_units_; duration_units_default_holder<CharT>::valid_units_ = 0; duration_units_default_holder<CharT>::initialized_ = false; } } This solves the memory leak, and seems to work fine for my simple test program. But n_d_valid_units_, valid_units_, and initialized_ are static variables, so I assume you want to only create these variables once. When run with my test program and step through it in the debugger, only one instantiation of duration_units_default_initializer_t occurs. (That is, once for each char type, single-byte and wide.) This occurs when the duration_units_default_initializer is instantiated (at the very end of duration_units.hpp). Since this is a global variable, this occurs during application startup and the destructor runs during application shuttdown. But presumably this can occur in some other situations, otherwise why would you have all the logic of making valid_units_ etc. class-static variables. I don't think I understand enough about what is going on to propose a robust solution. Thanks for your help, Don

2 1

boost::thread include order error
by Rob Patro 13 Feb '13

13 Feb '13

Hi, I noticed that when I don't include <boost/thread/thread.hpp> before the standard library includes, I get the following errors: In file included from /fs/szattic-asmg6/rob/dist/include/boost/thread/lock_types.hpp:14:0, from /fs/szattic-asmg6/rob/dist/include/boost/thread/pthread/thread_data.hpp:12, from /fs/szattic-asmg6/rob/dist/include/boost/thread/thread.hpp:17, from /fs/szattic-asmg7/rob/countseq/include/iterative_optimizer.hpp:26, from /fs/szattic-asmg7/rob/countseq/src/sailfish.cpp:33: /fs/szattic-asmg6/rob/dist/include/boost/thread/lockable_traits.hpp:90:37: error: expected identifier before ‘*’ token /fs/szattic-asmg6/rob/dist/include/boost/thread/lockable_traits.hpp:114:37: error: expected identifier before ‘*’ token However, when I move the boost thread include to the top of the set of included headers (above the standard library), the program compiles and runs correctly. What is the reason for this frustrating requirement? Is it possible to eliminate this behavior, so that <boost/thread/thread.hpp> can be included anywhere (or at least after the standard library) ? Thanks, Rob

2 1

[Preprocessor] How to work around empty sequence?
by Adam Badura 13 Feb '13

13 Feb '13

I was using sequences happily until I realized that quite often I will have an empty collection. Yet empty sequences are not supported. So I switched to lists. But they are syntactically so much worse! You have to add that BOOST_PP_NIL. You have to take care of proper nesting of elements. Much less readable and editable in my opinion. This made me consider going back to sequences after all and requiring user to provide some kind of special symbol for empty sequence or end eery sequence with a fixed item so that I will now it is end and not a real item. In my opinion the extra effort would still be less than with the lists while overall "user experience" would be better. (At the expense of complicating my own code - but I'm ready to accept that.) Does anyone of you ever did something like that? Guidelines? Ideas? How do sequence function deal with empty sequences? If I would pass to BOOST_PP_SEQ_ENUM a result of BOOST_PP_SEQ_POP_FRONT on a single-element sequence would it work fine (enumerate nothing)? Or it is not guaranteed?

2 1

some questions about boost.log & boost.process
by niXman 13 Feb '13

13 Feb '13

For my project I needed a logger and the ability to run/manage processes. Since the project will use boost, I do not want to drag into it external libraries. I googled boost.process[1] and boost.log[2][3], and I have three questions: 1) whether these libraries are suitable for use in real projects? have anyone used them? 2) do I understand correctly that the boost.process library does not have open CVS? 3) do I understand correctly that for the boost.log library I should use bleeding-edge branch? Thanks. -- Regards, niXman ___________________________________________________ Dual-target(32 & 64-bit) MinGW compilers for 32 and 64-bit Windows: http://sourceforge.net/projects/mingwbuilds/ ___________________________________________________ Another online IDE: http://liveworkspace.org/

4 8

How to search multi_index container using composite key with date ranges
by Jeff 13 Feb '13

13 Feb '13

Hi, I am having syntax issues and wish to search a bmi container by the composite index defined below. I have read about boost::lambdas and understand how they work. I simply have gotten stuck in the syntax to perform the get on the index and how to create the tuple(s) containing the proper objects. I want to search such that searchKey1="key1" and searchKey2="key2" and ValidFrom is <= '2013-03-01' and ValidTo >= '2013-02-08'. Can anyone help me please? Many Thanks!! Jeff R. bmi=boost::multi_index bg=boost::gregorian typedef bmi::multi_index_container<TestRecord, bmi::indexed_by< bmi::ordered_non_unique< bmi::tag<idx_test >, bmi::composite_key<TestRecord, bmi::member<TestRecord, std::string, &TestRecord::searchKey1>, bmi::member<TestRecord, std::string, &TestRecord::searchKey2>, bmi::member<TestRecord, bg::date, &TestRecord::ValidFrom>, bmi::member<TestRecord, bg::date, &TestRecord::ValidTo> > > > > TestCache;

2 3

[python] boost-python headers - are they python version specific?
by Steve Lorimer 12 Feb '13

12 Feb '13

I have python 2.7.3 and 3.2.1 deployed in my environment. I want to build 2 variants of boost python - one for each python version. My question is - will the boost headers be different; or can I just use the same header location for the boost headers, and link against a different boost-python library depending on which python version I want? TIA Steve

2 3

unix tools
by gast128 12 Feb '13

12 Feb '13

Hello all, since Boost has marvelous cross platform libraries, wouldn't it be an idea to rewrite all the standard commandline unix tools (e.g. cp, grep, ps, ls, diff) in c++ with Boost? Libraries like program options, filesystem and regex already fullfill a lot of the functionality. Only thing left is a person with a good knowledge of the syntax of the cmd line tools and ofc a lot of time:). I mainly use Windows and i really miss these tools from time to time. And yes there is Cygwin but I would prefer just the tools without changing my environment.

6 8

[Chrono] Memory leak in Chrono?
by Don 12 Feb '13

12 Feb '13

Boost 1.53 Chrono version 2 VC 2008 Win7-64 If I #include <boost/chrono.hpp> and run my program, on program exit the VC debugger reports about 40 memory leaks. I don't have to actually use chrono, just include the header. A little trial-and-error with #include statements tracked this down to boost\chrono\io\duration_units.hpp. It looks like there are a couple of "new" allocations in the constructor for duration_units_default_initializer_t that are never deleted. As a test I added a destructor that did delete[] on them, and that fixed the memory leaks. However, there is some logic with the initialized_ variable in the constructor that leads me to suspect that this is not a viable solution for real code. :-) Can anyone suggest a better solution for this? Or failing that, give me any ideas about what this code is actually doing and/or how I might proceed? There is a good chance that someone will say "that's not a real memory leak, it's allocated once and never deleted, but it goes away when your program exits. Don't worry about it". The problem with that answer is that when you have 40 or 50 memory leaks that are "normal", everyone stops paying any attention to the memory leak report and you start adding (unintentionally) real memory leaks in your code. Thanks for your help, Don

4 4

[Preprocessor] Comma in sequence element
by Adam Badura 12 Feb '13

12 Feb '13

I have following code: #include <boost/preprocessor/seq/enum.hpp> #include <boost/preprocessor/seq/transform.hpp> template< int I, int J > struct Base { }; #define SEQ ( 1 ) ( 2 ) ( 3 ) ( 4 ) #define OP( s, data, elem ) Base< elem, elem > struct Derived : BOOST_PP_SEQ_ENUM( BOOST_PP_SEQ_TRANSFORM( OP, nil, SEQ ) ) { }; The Derived has bases classes generated from a sequence (SEQ). The bases classes themselves are instantiations of Base template type. I expected the code to be preprocessed into something like (just the important part): struct Derived : Base< 1, 1 >, Base< 2, 2 >, Base< 3, 3 >, Base< 4, 4 > { }; but instead it is preprocessed into: struct Derived : Base< 1, Base< 2, Base< 3, Base< 4 { }; This obviously is caused by Base being template with to parameters and thus having a comma. Adding parenthesis around Base in OP solves the issue by generating expected code fully. But it still does not compile because parenthesis are not allowed in that place (base classes list). Is there any other workaround? How to deal with it?

3 2

udp::socket::socket() vs udp::socket::open()
by Chris Stankevitz 11 Feb '13

11 Feb '13

Hello, The UDP async server example opens the udp socket like so: a) udp::socket::socket(io_service, udp::endpoint(udp::v4(), 13)) Boost's comment regarding this invocation: * This constructor creates a datagram socket and automatically opens it bound * to the specified endpoint on the local machine. The protocol used is the * protocol associated with the given endpoint. Please consider the udp::socket constructor that takes only an io_service: b) explicit basic_datagram_socket(boost::asio::io_service& io_service) Commented as: * This constructor creates a datagram socket without opening it. The open() * function must be called before data can be sent or received on the socket. And finally, udp::socket::open: (c) void open(const protocol_type & protocol = protocol_type()); Please note that (a) takes an endpoint while (c) takes a protocol. === Question: How can I effectively call (a) using the two-part invocation (b) then (c)? Answer: 1) Just call udp::socket::open(endpoint(udp::v4(), 13)); 2) This is impossible because (a) takes a endpoint while (c) wants a protocol 3) Why would you want to do that you moron? 4) Why don't you write your own udp::socket::open that takes an endpoint if you are so smart? Try giving back to the community instead of just taking. You are a leech and not worthy of reading this mailing list let alone posting to it. 5) none of the above Note: (1) fails with error: no matching function for call to 'boost::asio::basic_datagram_socket<boost::asio::ip::udp>::open(boost::asio::ip::udp::endpoint)' Thank you, Chris

1 0