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

Modular boost directory structure
by Edward Diener 27 Aug '13

27 Aug '13

I am seeing at https://svn.boost.org/trac/boost/wiki/StartModDev that the modular Boost directory structure is: simple include boost simple twice.hpp test twice_test.cpp Jamfile.v2 doc // from the notes on the page index.html Is this the latest correct information ? I am asking because the original recommended structure was, as I understood it: simple boost simple twice.hpp libs simple test twice_test.cpp Jamfile.v2 doc html index.html I want to make sure I have this right before I modularize my sandbox libraries for git using the correct directory structure. BTW I have not been able to figure out how to add and move directories using git in a local copy. Instead I have just done so using the OS and git seems to figure it out but I have a feeling I have lost git history in doing so.

4 6

[Boost.Log] SimpleEventLog configuration
by Pieter 27 Aug '13

27 Aug '13

LS, I am currently configuring my logging configuration using a property tree. When I include a simple_event_log_backend sink, it does not seem possible (yet) to configure an event type mapping to that sink from the property tree. The source code does contain the following comment: "For now we use only the default level mapping. Will add support for configuration later." so I guess my first question would be whether such support is indeed in the works? My second question would be whether it is possible to add such a mapping after the sink was created and added to the core. In the documentation an example is given where the mapping is done before the sink is added to the core: typedef sinks::synchronous_sink< sinks::simple_event_log_backend > sink_t; boost::shared_ptr< sink_t > sink(new sink_t()); sinks::event_log::custom_event_type_mapping< severity_level > mapping("Severity"); // . define mapping here . sink->locked_backend()->set_event_type_mapper(mapping); // Add the sink to the core logging::core::get()->add_sink(sink); but in the case at hand adding the sink happens behind the scenes when using the configuration utilities. It seems to me that there is no easy way to manipulate the sink in any way once it is added to the core (unless as in the example the shared_ptr to the sink is kept to access the sink maybe). Is this correct? Many thanks, Pieter P.s. BTW, in the section "Simplified library initialization tools" of the documentation Table 1.15 seems to miss a 'Format' parameter that I assume is still needed?

2 1

doubts about python3 support
by Neal Becker 26 Aug '13

26 Aug '13

I'm building using: bjam -d+2 -j8 cxxflags='-march=native -Wno-unused-local-typedefs' -sHAVE_ICU=1 - sEXPAT_INCLUDE=/usr -sEXPAT_LIBPATH=/usr/lib64 --layout=tagged threading=single,multi link=shared stage In user-config.jam I added: using python : 3.3 : /usr/bin/python3 : /usr/include/python3.3m ; I wind up with stage/lib/libboost_python.so stage/lib/libboost_python3.so But AFAICT, they were build the same way. It's be nice if one was built against python2 and the other python3, but that doesn't appear to be true. AFAICT, both were built from the same sources (using /usr/include/python3.3m). What I'd like is to build both python2 and python3 from the same source tree, installing the final shared libs into the same destination dir. Right now I have 2 different source trees, and build switching the one line in user- config.jam.

2 3

MSVC doesn't support "not"?!
by Daryle Walker 26 Aug '13

26 Aug '13

Just checked the regression testing pages. Most of the testing stations that reported fails for Boost.Rational have switched to pass since I've uploaded updates. Besides the stations that haven't run new tests, and the "teeks" team that run Clang tests although it seemingly isn't actually installed, the newly-run MSVC stations report fails. The old errors are gone, but the new error is: ..\libs\rational\test\rational_test.cpp(1012) : error C2065: 'not' : undeclared identifier ..\libs\rational\test\rational_test.cpp(1012) : error C2146: syntax error : missing ')' before identifier 'dummy' ..\libs\rational\test\rational_test.cpp(1012) : error C2059: syntax error : ')' This happens with MSVC 10.0 and 11.0. Hasn't "not" been an alternate symbol for "!" since C++98? I'll change it to "!", but that doesn't excuse not fixing it for 15 years! What happened? (I think some of the MSVC authors lurk here.) Daryle W.

6 7

[config] BOOST_NO_SFINAE_EXPR and GCC 4.6.4
by Agustín K-ballo Bergé 26 Aug '13

26 Aug '13

According to this table http://gcc.gnu.org/projects/cxx0x.html, GCC supports "Solving the SFINAE problem for expressions" since 4.4. According to Boost.Config, it is since 4.5. The following test case works fine with 4.8.1, but fails miserably with 4.6.4: http://ideone.com/RzgXKM Errors include invalid conversion from X const* to X*, and use of non-member-ptr in member-ptr expression. GCC 4.7.3 (tested by zao at #boost) shows similar results: https://www.acc.umu.se/~zao/kbsfinae11.txt Should BOOST_NO_SFINAE_EXPR be adjusted according to the results of this test case? Regards, -- Agustín K-ballo Bergé.- http://talesofcpp.fusionfenix.com

3 7

[RPC] Boost.Asio to write *concurrency ready* code
by microcai 26 Aug '13

26 Aug '13

Hi, boost experts! This is my first mail. Although I am not boost newbee, I am indeed new to the boost community. == I was using Boost.Asio for a long time, for *Async Network I/O*, but up till recently, I made an amazing discover that Boost.Asio is NOT ONLY a Network library, but indeed a very *general* toolkit for building concurrent algorithm. I made the first discover when my friend show up an concurrent sum ``` #include <iostream> #include <vector> #include <algorithm> #include <numeric> #include <future> template <typename RAIter> int parallel_sum(RAIter beg, RAIter end) { typename RAIter::difference_type len = end-beg; if(len < 1000) return std::accumulate(beg, end, 0); RAIter mid = beg + len/2; auto handle = std::async(std::launch::async, parallel_sum<RAIter>, mid, end); int sum = parallel_sum(beg, mid); return sum + handle.get(); } int main() { std::vector<int> v(10000, 1); std::cout << "The sum is " << parallel_sum(v.begin(), v.end()) << '\n'; } ``` Here the actual calculations are done async and concurrently. But, the problem remains: ** The total threads that this algorithm spawn is not predictable and not configurable. ** And that's why we will not like such algorithm altogether. Days later, when I am review an old code that was doing an *async directory walk through* , I soon discover an general *concurrency ready* algorithm that the number of threads utilized can be fully controlled. The code utilized *any number of threads that is doing io_service::run()* Here is the good old demo that uses recursive algorithm : ``` // this is sequential recursing version void hanoi_sync(int n, char A, char B, char C) { if (n == 1) { printf("Move disk %d from %c to %c\n", n, A, C); } else { hanoi_sync (n-1, A, C, B); printf("Move disk %d from %c to %c\n", n, A, C); hanoi_sync (n-1, B, A, C); } } ``` But with Boost.Asio, we turn it into parallelism algorithm: ``` // concurrent hannoi tower solver void hanoi_async(boost::asio::io_ service& io, int n, char A, char B, char C) { if (n == 1) { printf("thread id %d, Move disk %d from %c to %c\n", boost::this_thread::get_id(), n, A, C); } else { io.post(boost::bind(&hannoi_async, boost::ref(io), n-1, A, C, B)); printf("thread id %d, Move disk %d from %c to %c\n", boost::this_thread::get_id(), n, A, C); io.post(boost::bind(&hannoi_async, boost::ref(io), n-1, B, A, C)); } } ``` Thanks to the amazing io_service.post, we can now turing any (if not mostly) *recursing code* into *parallelism code*. And the threads that used by such *parallelism code* is fully controlled. If one thread is calling io_service::run(), then it is simple an async version of the sequential recursing one, if more that one thread is doing io_service::run(), then you know that the job is now run in parallel. And yet there needs only one place to control how may threads is used. Want it to be faster? spawn more threads to call io_service::run()! The code will *automatically* use all threads that you have allocated to Boost.Asio. I am *not sure* if anyone else has made such discovery, anyway, comment is welcomed.

1 0

[type_triats] [move] is_copy_constructible<T> trait
by Antony Polukhin 26 Aug '13

26 Aug '13

Hi, I have an implementation of is_copy_constructible<T> trait. It works great with C++11, in C++03 it detects constructors assuming that if type is derived from boost::noncopyable or uses BOOST_MOVABLE_BUT_NOT_COPYABLE macro - it has no copy constructor. Questions are: * May I add that trait to trunk version of Boost.TypeTraits (with documentation and tests, as usual)? * May I or Boost.Move maintainer apply this patch to BOOST_MOVE_IMPL_NO_COPY_CTOR_OR_ASSIGN in boost/move/core.hpp : #ifdef BOOST_NO_CXX11_DELETED_FUNCTIONS #define BOOST_MOVE_IMPL_NO_COPY_CTOR_OR_ASSIGN(TYPE) \ private:\ TYPE(TYPE &);\ TYPE& operator=(TYPE &);\ + public: \ + typedef int boost_move_no_copy_constructor_or_assign; \ + private: \ // #else Why do I need this trait? Well, it allows to implement move_if_noexcept() that is required for circular_buffer and possibly for other containers. -- Best regards, Antony Polukhin

2 4

[git] Some questions on Git transition
by Daryle Walker 26 Aug '13

26 Aug '13

* I see we have a GitHub account at https://github.com/boostorg, with repositories for each library. If we add a new library, and it has been Git-based, are we going to mirror it in the Boost account? That'll make sure if something happens to the original, we still have a last-known-good to work with. Would having all the libraries (or copies/mirrors) under the same account make scripts easier? What about non-Git libraries, will they be copied over? Would a bridge be used? (There's at least bridges for Git connecting to either Subversion or Mercurial.) * We currently ship all the libraries merged together as a single archive. Some package managers even provide the Boost libraries as a single item. When we go Git, is that still going to be in place (with the release managers doing some sort of merging script), or are we and every user going to move to doing several dozen downloads instead? (And everyone has to figure out how to use more than one Boost library at once.) * Are we eventually going to use something like http://nvie.com/posts/a-successful-git-branching-model/? If so, are we going to use that to script releases? (At release-time, grab the latest version of each library from their master branches, apply the merging script, run tests, then ship.) Daryle W.

2 3

[release] How do you make a library update official?
by Daryle Walker 24 Aug '13

24 Aug '13

I just uploaded changes ([85417] and [85418]) to Boost.Rational that fix the two major problems showing up on the regression tests. How do I get them on the official distribution? I know we don't read off the SVN-trunk directly; there's some staging area first. (That's why the shippers occasionally declare a code freeze deadline.) I know we have a "release" branch, but I'm not sure that's still being used. I want to blaze ahead making tweaks to Boost.Rational, but not if you can't "make official" a non-HEAD change-set. Even if I miss the deadline for these other tweaks, I want the two I just already did to be in Boost 1.55. Daryle W.

8 9

[admin] [svn] Boost.Commit list is not working
by Michel Morin 23 Aug '13

23 Aug '13

It seems that Boost.Commit mailing list has suspended for ten days. Regards, Michel

1 0