- Boost - lists.boost.org

release manager for 1.52?
by Oliver Kowalke 28 Aug '12

28 Aug '12

Who is the release manager for 1.52?

2 1

[atomic] lockfree failure on msvc 64-bit
by Christoph Macheiner 27 Aug '12

27 Aug '12

As indicated by http://www.boost.org/development/tests/trunk/developer/output/cm_x64-boost-…. In boost/atomic/detail/interlocked.hpp, BOOST_ATOMIC_LLONG_LOCK_FREE must be defined to 2 if _WIN64 is defined (cas64strong.hpp is included afterwards). Because the library has not yet been released, I didn't bother to create a ticket and because the change is so trivial I didn't create a patch as well (if either or both is required or wanted, please let me know). Question: If I recall correctly, long long can be made lockfree also in 32-bit. If I am right, is there a reason why it hasn't been done? Thanks. --Christoph.

1 0

memory corruption in async_read_until using asio::ssl
by Jim Page 27 Aug '12

27 Aug '12

Afternoon All I am seeing an occasional memory corruption bug, using valgrind, in our web server application that is using boost::asio::ssl, and it seems to be related to calls to async_read_until, where several chunks of data are required before the terminating regex is found. What appears to be happening is that engine::perform() causes a write (via ssl3_read_bytes()) on a deleted memory region, the buffer into which decrypted data is written. This region was delete'd via invocation of the detail::io_op::handler, and in turn via read_until_delim_op::operator(), due to prepare() being called on the streambuf, and a std::vector<char> is delete'd in order to create a larger one. It seems like one of the chain of handlers is using a pointer which is becomes invalid when the streambuff is expanded after failing to find the regex terminator after one or more reads, and prepare() is called before the next async_read_some() on the underlying stream. I have included the valgrind output at the end - sorry for the large size. I am pretty certain that I have ruled out the obvious (multiple simultaneous attempts to read using the same streambuf), and I am only using one thread for the io_service. This seems a pretty common potential failure mode, so perhaps I am barking up the wrong tree - I'd appreciate anyone's opinion before I start a potentially tricky debug effort requiring modification to the boost libraries on our production servers. I'm using boost 1.49. There's a lot of optimization on the application which is why some calls are optimized out of the valgrind dump. The machines are running Centos 5.8, x86_64. Compiler is gcc 4.4.6. Thanks for looking Jim Page ==29870== Thread 22: ==29870== Invalid write of size 1 ==29870== at 0x4A088F8: memcpy (mc_replace_strmem.c:587) ==29870== by 0x3400821EDE: ssl3_read_bytes (in /lib64/libssl.so.0.9.8e) ==29870== by 0x340081E750: ??? (in /lib64/libssl.so.0.9.8e) ==29870== by 0x66A75B: boost::asio::ssl::detail::engine::perform(int (boost::asio::ssl::detail::engine::*)(void*, unsigned long), void*, unsigned long, boost::system::error_code&, unsigned long*) (engine.ipp:219) ==29870== by 0x7E3326: boost::asio::ssl::detail::io_op<boost::asio::basic_stream_socket<boost::asio ::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> >, boost::asio::ssl::detail::read_op<boost::asio::mutable_buffers_1>, boost::asio::detail::rea d_until_expr_op<boost::asio::ssl::stream<boost::asio::basic_stream_socket<bo ost::asio::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> > >, std::allocator<char>, boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >, boost::_bi::bind_t<void, boost::_mfi::mf2<void, HttpConnection, boost::system::error_code const&, char const*>, boost::_bi::list3<boost::_bi::value<boost::shared_ptr<HttpConnection> >, boost::arg<1> (*)(), boost::_bi::value<char const*> > > > >::operator()(boost::system::error_code, unsigned long, int) (engine.ipp:151) ==29870== by 0x7E45C2: boost::asio::detail::wait_handler<boost::asio::ssl::detail::io_op<boost::asi o::basic_stream_socket<boost::asio::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> >, boost::asio::ssl::detail::read_op<boost::asio::mutable_buffers_1>, boost::asio::detail::read_until_expr_op<boost::asio::ssl::stream<boost::asio ::basic_stream_socket<boost::asio::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> > >, std::allocator<char>, boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >, boost::_bi::bind_t<void, boost::_mfi::mf2<void, HttpConnection, boost::system::error_code const&, char const*>, boost::_bi::list3<boost::_bi::value<boost::shared_ptr<HttpConnection> >, boost::arg<1> (*)(), boost::_bi::value<char const*> > > > > >::do_complete(boost::asio::detail::task_io_service*, boost::asio::detail::task_io_service_operation*, boost::system::error_code const&, unsigned long) (bind_handler.hpp:46) ==29870== by 0x619783: boost::asio::detail::task_io_service::run(boost::system::error_code&) (task_io_service_operation.hpp:37) ==29870== by 0x79F4FE: BayeuxServer::run() (io_service.ipp:59) ==29870== by 0x364800DB8E: ??? (in /usr/lib64/libboost_thread.so.1.49.0) ==29870== by 0x3AE300677C: start_thread (in /lib64/libpthread-2.5.so) ==29870== by 0x3AE20D325C: clone (in /lib64/libc-2.5.so) ==29870== Address 0x1b508610 is 0 bytes inside a block of size 1,024 free'd ==29870== at 0x4A05A33: operator delete(void*) (vg_replace_malloc.c:346) ==29870== by 0x658E54: std::vector<char, std::allocator<char> >::_M_fill_insert(__gnu_cxx::__normal_iterator<char*, std::vector<char, std::allocator<char> > >, unsigned long, char const&) (new_allocator.h:95) ==29870== by 0x68B805: boost::asio::basic_streambuf<std::allocator<char> >::reserve(unsigned long) (stl_vector.h:851) ==29870== by 0x7E2EF2: boost::asio::detail::read_until_expr_op<boost::asio::ssl::stream<boost::asio ::basic_stream_socket<boost::asio::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> > >, std::allocator<char>, boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >, boost::_bi::bind_t<void, boost::_mfi::mf2<void, HttpConnection, boost::system::error_code const&, char const*>, boost::_bi::list3<boost::_bi::value<boost::shared_ptr<HttpConnection> >, boost::arg<1> (*)(), boost::_bi::value<char const*> > > >::operator()(boost::system::error_code const&, unsigned long, int) (basic_streambuf.hpp:207) ==29870== by 0x7E32DE: boost::asio::ssl::detail::io_op<boost::asio::basic_stream_socket<boost::asio ::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> >, boost::asio::ssl::detail::read_op<boost::asio::mutable_buffers_1>, boost::asio::detail::read_until_expr_op<boost::asio::ssl::stream<boost::asio ::basic_stream_socket<boost::asio::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> > >, std::allocator<char>, boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >, boost::_bi::bind_t<void, boost::_mfi::mf2<void, HttpConnection, boost::system::error_code const&, char const*>, boost::_bi::list3<boost::_bi::value<boost::shared_ptr<HttpConnection> >, boost::arg<1> (*)(), boost::_bi::value<char const*> > > > >::operator()(boost::system::error_code, unsigned long, int) (read_op.hpp:59) ==29870== by 0x7E493C: boost::asio::detail::reactive_socket_recv_op<boost::asio::mutable_buffers_1, boost::asio::ssl::detail::io_op<boost::asio::basic_stream_socket<boost::asio ::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> >, boost::asio::ssl::detail::read_op<boost::asio::mutable_buffers_1>, boost::asio::detail::read_until_expr_op<boost::asio::ssl::stream<boost::asio ::basic_stream_socket<boost::asio::ip::tcp, boost::asio::stream_socket_service<boost::asio::ip::tcp> > >, std::allocator<char>, boost::basic_regex<char, boost::regex_traits<char, boost::cpp_regex_traits<char> > >, boost::_bi::bind_t<void, boost::_mfi::mf2<void, HttpConnection, boost::system::error_code const&, char const*>, boost::_bi::list3<boost::_bi::value<boost::shared_ptr<HttpConnection> >, boost::arg<1> (*)(), boost::_bi::value<char const*> > > > > >::do_complete(boost::asio::detail::task_io_service*, boost::asio::detail::task_io_service_operation*, boost::system::error_code const&, unsigned long) (bind_handler.hpp:118) ==29870== by 0x613778: boost::asio::detail::epoll_reactor::descriptor_state::do_complete(boost::asi o::detail::task_io_service*, boost::asio::detail::task_io_service_operation*, boost::system::error_code const&, unsigned long) (task_io_service_operation.hpp:37) ==29870== by 0x619783: boost::asio::detail::task_io_service::run(boost::system::error_code&) (task_io_service_operation.hpp:37) ==29870== by 0x79F4FE: BayeuxServer::run() (io_service.ipp:59) ==29870== by 0x364800DB8E: ??? (in /usr/lib64/libboost_thread.so.1.49.0) ==29870== by 0x3AE300677C: start_thread (in /lib64/libpthread-2.5.so) ==29870== by 0x3AE20D325C: clone (in /lib64/libc-2.5.so)

1 0

Boost.Process 0.5 released
by Boris Schaeling 27 Aug '12

27 Aug '12

Boost.Process 0.5 has been released! Documentation and library can be found at: <http://www.highscore.de/boost/process0.5/> Why version 0.5? We have no official process management library in Boost yet. But we had countless drafts since 2006. In order not to get completely lost between all those drafts I gave them version numbers. For example, 0.1 was the very first draft in 2006. 0.31 is the draft documented at <http://www.highscore.de/cpp/process/>. And 0.4 was the draft which was rejected in a formal review in spring 2011. Version 0.5 is the latest attempt to create a process management library. What's new? Version 0.5 has a very different design than all other drafts before. The design was proposed by Jeff Flinn who introduced it briefly at BoostCon 2011 when we gave a presentation on Boost.Process. The goals of this new version are: * A lightweight implementation (no overly complicated code anymore). * A fully extensible library (no need for arbitrary "extension points"). * Support for system-level cross-platform concepts only (no Java-like constructs). Other goals like making the library easy to use were of course also targeted. But these were goals before. The goals above are important as those things have been criticized in earlier drafts and were as far as I see the main reasons why version 0.4 had been rejected. What now? Please have a look at the documentation and think about whether this library could fit into Boost. If you have more time, please download and play around with the new version. Concentrate on the bigger picture: Does the design make sense? Is it easy to understand? Can you do what you want to do? Does it look and feel like modern C++? This is not a formal review. But it would help to get some feedback to see a tendency. After six years, two Google Summer of Code projects, countless drafts and still a lot of interest from many people in a process management library in Boost, we might be able to get somewhere this time. (If you are on Windows: There is a bug in Boost.Iostreams which has been fixed in the release candidate 1 of Boost 1.51.0. See the documentation if you run into a problem when closing pipes.) Anything missing? We had a closed beta test in the previous two weeks with about ten people who had contacted me because of their interest in Boost.Process. I got already some good feedback and have to change a few details. These are only details though and shouldn't stop anyone from using this new version now. If everyone is happy with the overall design and architecture and it's only details we have to discuss, it's already a huge step forward. Last but not least: Boost.Process 0.5 wouldn't exist without the help of Intra2net AG (<http://www.intra2net.com/>). Intra2net AG sponsored a project to support the development of Boost.Process. It was this sponsorship which made it possible to create this new version. Thanks to them! Boris

9 46

[rational] Over/underflow checked update ready for comments
by Dan Searles 26 Aug '12

26 Aug '12

>> Quick overview of the update: the original usage and >> code of the template class remains mostly as before, ie >> boost::rational<int> RationalVar; should work as >> in the original version. Adding 'true' as a second >> template parameter, however, allows an exception to >> be raised if an operation with a class variable results >> in a value that can't accurately be represented >I haven't looked at anything else, but I dislike using true/false as the template argument. I'd prefer meaningful names. _____ >Rob Stewart robert.stewart(a)sig.com >Software Engineer using std::disclaimer; >Dev Tools & Components >Susquehanna International Group, LLP http://www.sig.com Replaced the bool template parameter with an enum: enum RationalChecktype {RationalNoChecking, RationalCheckforOverflow};

1 0

[contract] "Duplicated globals" problem
by Andrzej Krzemienski 26 Aug '12

26 Aug '12

Hi Lorenzo, I think that installing broken contract handlers will not work correctly when a program is linking with DLLs on Windows. I do not know how dynamic linking works on Linux. Here is a sketch of a sample program that fails to work as expected. The program consists of two projects: one builds a DLL, the other builds an executable that links with the DLL: DLL: ==================================== # include <contract.hpp> # include "testdll.h" CONTRACT_FUNCTION ( int (give1) () postcondition(auto result = return, result == 1) ) { return 2; // obviously wrong } DLL_MODE void test_contracts() { int i = give1(); } ==================================== EXE: ==================================== # include <contract.hpp> # include "../testdll/testdll.h" void PostconditionHandler( contract::from const& context ) { std::cerr << "ERROR!!!" << std::endl; } int main() { contract::set_postcondition_broken(&PostconditionHandler); test_contracts(); } ==================================== I expect that the program would use the broken postcondition handler I installed, but it still uses the default one. The cause is that Contract is using some global object to store the installed handler, but when a program is linked with a DLL both the program and the DLL store their own copy of the global. The global is duplicated. THe one in the EXE registers my callback, but the global in the DLL still keeps the default handler. The only way to solve this issue is to -- at least in the cases when users' programs link with DLLs -- ship Contract with a small DLL that stores the unique copy of the globals. This is how Boost.Log handles this problem. Quoting the documentation ( http://boost-log.sourceforge.net/libs/log/doc/html/log/installation.html) *"If your application consists of more than one module (e.g. an exe and one or several dll's) that use Boost.Log, the library must be built as a shared object. If you have a single executable, you may build the library as a static library." * Regards, &rzej* *

2 1

complex acos and atanh
by Stephen Montgomery-Smith 26 Aug '12

26 Aug '12

Hey people, I was trying out your complex arc-trig functions, to compare them with some that I have written: http://people.freebsd.org/~stephen/catrig.c. I am finding that your acos function is getting some of them wrong. For example, your program evaluates acos(1.00000002785941 + I*5.72464869028403e-200) as 0 - I*0.000236048349018331 whereas it should be 2.42520172707401e-196 - I*0.000236048349018331. Also your complex atanh function does quite badly in some edge cases. I presume it is because Hull, Fairgrieve and Tang did not write a paper on that function. If you email Fairgrieve, he may give you a copy of the paper that was never published. He sent me a copy, but doesn't want it widely distributed. I devised my own algorithm at http://people.freebsd.org/~stephen/catrig.c, which actually turns out to be very similar to the unpublished algorithm of Hull, Fairgrieve and Tang. Anyway, I would be happy to work with whomever is currently working on these algorithms. I think someone must be, because I see that someone fixed a bug in asin since 1.48.0. Best regards, Stephen

2 2

[locale] Custom Gettext paths.
by Jookia 26 Aug '12

26 Aug '12

Hello! I'm extremely happy about Boost.Locale, but I've found a few things lacking. So at first I wrote a some wrappers to get around a few flaws with my usage of Boost.Locale. One of these flaws was how it's hardcoded to use the Gettext directory hierarchy. I'd rather store my stuff in 'lang/en_US.mo' rather than 'lang/en_US/LC_MESSAGES/my_app.mo'. The patch adds a 'path format' feature, which allows you to format the directory structure when finding Gettext catalogs, to achieve the effect above. All you really have to do is run: gen.add_path_format("{1}/{2}.mo"); // Use a smaller hierarchy. to achieve the result that I prefer, or gen.add_path_format("{1}/{2}/{3}/{4}.mo"); // Use a Gettext hierarchy. to achieve the result that Boost.Locale uses right now. Ripped straight from Doxygen comments: {1} A path to search for catalogs in. {2} The locale's name. {3} The locale's category. {4} The Gettext domain. I'm new to submitting patches and things, and I'm not sure how I could make this feature better or even acceptable if it isn't already, but I'm sure it'd help others out there. There are a few drawbacks I've found with my code: - To be backwards compatible I've had to add some hackery to add the Gettext format if the path formats vector is empty. - Without consulting documentation, the format is unreadable. - I haven't tested it on the Win32 backend. I'm 90% sure the code would work, it's only 4 lines of code added and it's identical to the code added in the other backends, but it's still untested code. - It's not tested properly. By properly, I mean the tests past with adding a format, and falling back to the default format, but I haven't tried having a 'compact' format. I'm not sure if this would indicate a fault in the formatting code though, as that'd be the only difference code-wise, unless there's more code that has the Gettext hierarchy hardcoded, that didn't fail when I didn't fix the tests. The patch itself is based off of SVN r80097, and despite the diff being for r80104 I haven't tested it in that.. I did the work internally in Git but I made the patch using SVN, in case that matters. I've attached the patch, but I'm not sure how well that's going to go over considering I'm new to mailing lists, and I'm using.. Thunderbird.. Yeah. In case the attachment fails, it's available at http://sprunge.us/VSiX . I'd also like to know if it's appropriate to put '[patch]' in the title for any future patches, and if it should go after '[locale]' or at the end of the title. Thanks, Jookia.

2 13

boost::filesystem -- using non-POSIX locale crashes on Solaris
by JonathonS 26 Aug '12

26 Aug '12

Hi, I am currently using boost::filesystem::path on Solaris, and I noticed that if I set the locale to anything except "C" or "POSIX" such as "en_US.UTF-8", that it crashes i.e. a run-time error gets thrown. I found a link describing the issue here: https://svn.boost.org/trac/boost/ticket/5928 My question is though, whether it is possible to use boost::filesystem without locales? Meaning, if I know my string is a UTF-8 string, is it possible to avoid the locale processing within the boost filesystem? Thanks, J

3 2

[Review] Boost.Contract Review Starts Today(*)
by Dave Abrahams 26 Aug '12

26 Aug '12

The review period for Lorenzo Caminiti's Boost.Contract library runs from today through August 31st. I'm pleased to serve as review manager. ,----[ Important Notes ] | - Please begin your review's subject line with "[Review:Algorithms]" | - Please do not post your review or discussion as a reply to this | message. | - Please feel free to note any individual issues in the issue | tracker at https://sourceforge.net/apps/trac/contractpp/report/3 | (you'll need to log in at Sourceforge to add a ticket) and make | reference to them from your review. | - Please see below for *further* important notes `---- THE LIBRARY Boost.Contract implements Contract Programming (a.k.a., Design by Contracts) for C++. In addition, the library implements virtual specifiers, concept checking, and named parameters. Documentation: http://contractpp.sourceforge.net Source: http://sourceforge.net/projects/contractpp/files/latest/download All N1962 requirements and Eiffel features are supported by the library, among others: 1) Support for preconditions, postconditions, class invariants, block invariants, and loop variants. 2) Subcontract derived classes (with support for pure virtual functions and multiple inheritance). 3) Access expression old values and function return value in postconditions. 4) Optional compilation and checking of preconditions, postconditions, class invariants, block invariants, and loop variants. 5) Customizable actions on contract assertion failure (terminate by default but it can throw, exit, etc). This library is implemented for C++03 and it does not require C++11. Obviously, if the library is accepted, all macros will be prefixed by BOOST_... and other symbols will go under the boost::... namespace. Potential Interactions with Existing Boost Libraries ==================================================== This submission potentially interacts with the following existing Boost libraries: * parameter - http://boost.org/libs/parameter * static_assert - http://www.boost.org/libs/static_assert * MPL - http://www.boost.org/libs/mpl * Concept - http://www.boost.org/libs/concept_check Reviewers should consider these potential interactions. What to include in Review Comments ================================== - Keep your criticism constructive. Constructive criticism usually comes with specific suggestions for improvement. - If you identify problems along the way, please note if they are minor, serious, or showstoppers. - Please *do* vote on whether the library should be accepted. Everyone who takes the time to look at the library is qualified to vote, and every vote is valuable. - Other questions you might want to answer in your review: * What is your evaluation of the design? * What is your evaluation of the implementation? * What is your evaluation of the documentation? * What is your evaluation of the potential usefulness of the library? * Did you try to use the library? With what compiler? Did you have any problems? * How much effort did you put into your evaluation? A glance? A quick reading? In-depth study? * Are you knowledgeable about the problem domain? (*) Well, OK, the review period officially started yesterday. Sorry for the delayed announcement -- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost

2 2