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

[string_ref] degree of conformance to standard proposal
by Peter A. Bigot 24 Dec '13

24 Dec '13

(Background: I need a string_view with features in N3762 that aren't in N3442 (viz., pos parameters to the string methods). I'm trying to determine how much effort I should put into a set of patches to update the Boost implementation to the latest draft proposal.) The default constructor for Boost.string_ref produces an object which returns a null pointer from data(). N3442 and its successor N3762 both require the return value from data() to be non-null (http://isocpp.org/files/papers/N3762.html#h5o-5). GCC's libstdc++ folks solved this by adding a constexpr unit length string to serve as the base referenced object (http://gcc.gnu.org/ml/libstdc++/2013-11/msg00104.html). But for Boost I don't see how this can be changed without introducing a link dependency. In a cases like this, how closely does Boost want to track the standard proposals? Similarly, the standard proposal has settled on *string_view instead of *string_ref. Would Boost rename the library/class, or just introduce an alias? Peter

2 3

Re: [boost] [git] Effects of git submodule update
by Beman Dawes 24 Dec '13

24 Dec '13

On Mon, Dec 23, 2013 at 2:17 PM, Robert Ramey <robertmacleranramey(a)gmail.com > wrote: > > > On Monday, December 23, 2013 8:47:21 AM UTC-8, Beman Dawes wrote: > <snip ..> > > Comments welcome, >> > > Beman. > > This is beyond awesome !!!! (I can't believe I used the word "awesome") > > I've been dabbled with git on my own machine (now a new machine) for a few > months > and it's been OK but I never really "got" the branching/merging, etc..... > Mainly this > is because I didn't really have to yet. I've also been lurking around > waiting for > the smoke to clear with boost modularization - a cowardly position I know, > but expedient in my particular circumstances right now. The work you've > done > is getting this organized AND well explained is incredible. I'm aware > that the dust has to settle some. But this looks to me like it's going to > support a whole new trajectory for boost. It makes me very, optimistic > about future of boost and C++ in general. > > Personally, I don't think that it's too much to say that boost saved C++. > > And I don't think it's too much to say that you're saving boost right now. > Robert, thank you very much for the kind words. Knowing that work is appreciated is a strong motivation to continue! But do remember this is a long process with a lot of people contributing, both now and in the past. Dave Abrahams and Daniel Pfeifer deserve special mention for several years of work on the git and modular boost conversion implementation. The release managers and boost.build team are working on infrastructure. The testers will soon be doing their part. And mailing list participants are providing the ideas and feedback that drive the whole process. Thanks to all, --Beman

1 0

[Proposal] Boost Community Maintenance Team Goals and Guidelines
by Marshall Clow 24 Dec '13

24 Dec '13

Start here: https://dl.dropboxusercontent.com/u/9897665/community-maintenance.html I have contacted the authors of the libraries listed in Beman’s proposals. Some have responded affirmatively; the rest have not responded. So, I thought I would put down some thoughts about how the Boost Community Maintenance Team (Boost.CMT) should work. Goals (in priority order): 1. Keep things building on all the supported compilers (and as many others as possible) 2. Get the test matrix “all green” and keep it that way. 3. Fix the accumulated bugs; drive the bug count to zero, and keep it there. 4. Add new features. How I see this working: 0) We’ll set up a separate Boost.CMT mailing list. (the main developer list has too much traffic already) 1) People would pick a problem and post a note to the list that they were working on it. 2) {Optional] discussion ensues about the way to solve the problem. 3) A potential solution is implemented (and tested!) 4) A pull request is posted to the mailing list (with tests) 5) Discussion ensues. Repeat steps 3/4/5 as needed. 6) The pull request is merged to ‘develop’ 7) After test cycles, the change is merged to ‘master' Notes: 1) All of this should happen on the Boost.CMT ML. 2) Discussions _about_ the CMT should happen on the main Boost ML. 3) The person who lands the patch is responsible for checking the test results and merging to “master” (and closing out the bug report, if any). 4) The goal of the CMT will be the same as the Boy Scouts - “Leave the code cleaner than when you found it” Comments? -- Marshall Marshall Clow Idio Software <mailto:mclow.lists@gmail.com> A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait). -- Yu Suzuki

2 2

Re: [boost] [atomic] Probable documentation issue: how do we know atomic is safe and/or efficient?
by Pete Bartlett 24 Dec '13

24 Dec '13

> On 24 Dec 2013, at 13:43, "John Maddock" <john(a)johnmaddock.co.uk> wrote: > > This message has no content.

1 0

Ominous Warning — should I be worried?
by John M. Dlugosz 23 Dec '13

23 Dec '13

I'm compiling with VS10, and using Boost 1.49.0 “reference member is initialized to a temporary that doesn't persist after the constructor exits” That does not sound good, and browsing through the mailing list archives I see discussions of regressions in 2010 that cite C4413 and use of emplace. I can't read the source code because it's just a ball of preprocessor stuff. Is this a real problem or a warning that can be ignored? I would think that if Boost.Container was broken on VS10, it would be mentioned somewhere where it's easy to spot. Thanks, —John E:\boost_1_49\boost/preprocessor/iteration/detail/local.hpp(37): warning C4413: 'boost::container::container_detail::advanced_insert_aux_non_movable_emplace1arg<A,Iterator,P0>::m_p0' : reference member is initialized to a temporary that doesn't persist after the constructor exits with [ A=std::allocator<std::tr1::shared_ptr<MyTypeA>>, Iterator=std::tr1::shared_ptr<MyTypeA> *, P0=std::tr1::shared_ptr<MyTypeA> ] E:\boost_1_49\boost/preprocessor/iteration/detail/local.hpp(37) : see declaration of 'boost::container::container_detail::advanced_insert_aux_non_movable_emplace1arg<A,Iterator,P0>::m_p0' with [ A=std::allocator<std::tr1::shared_ptr<MyTypeA>>, Iterator=std::tr1::shared_ptr<MyTypeA> *, P0=std::tr1::shared_ptr<MyTypeA> ] E:\boost_1_49\boost/preprocessor/iteration/detail/local.hpp(37) : while compiling class template member function 'boost::container::container_detail::advanced_insert_aux_non_movable_emplace1arg<A,Iterator,P0>::advanced_insert_aux_non_movable_emplace1arg(A &,P0 &&)' with [ A=std::allocator<std::tr1::shared_ptr<MyTypeA>>, Iterator=std::tr1::shared_ptr<MyTypeA> *, P0=std::tr1::shared_ptr<MyTypeA> ] E:\boost_1_49\boost/preprocessor/iteration/detail/local.hpp(37) : see reference to class template instantiation 'boost::container::container_detail::advanced_insert_aux_non_movable_emplace1arg<A,Iterator,P0>' being compiled with [ A=std::allocator<std::tr1::shared_ptr<MyTypeA>>, Iterator=std::tr1::shared_ptr<MyTypeA> *, P0=std::tr1::shared_ptr<MyTypeA> ] E:\boost_1_49\boost/preprocessor/iteration/detail/local.hpp(37) : see reference to class template instantiation 'boost::container::container_detail::advanced_insert_aux_emplace1arg<A,Iterator,P0>' being compiled with [ A=std::allocator<std::tr1::shared_ptr<MyTypeA>>, Iterator=std::tr1::shared_ptr<MyTypeA> *, P0=std::tr1::shared_ptr<MyTypeA> ] mysourcefile.cpp(172) : see reference to function template instantiation 'void boost::container::vector<T>::emplace_back<std::tr1::shared_ptr<_Ty>>(P0 &&)' being compiled with [ T=std::tr1::shared_ptr<MyTypeA>, _Ty=MyTypeA, P0=std::tr1::shared_ptr<MyTypeA> ]

2 1

Beginner
by prabhdeep singh 23 Dec '13

23 Dec '13

I want to start contributing to the open source community, but I do not where to begin. Your community's cause seems great! So I would love to start but I am absolutely clueless as to where to begin. I have a background in C++ and algorithmic operations. Any help would be appreciated!

2 1

[git] Effects of git submodule update
by Beman Dawes 23 Dec '13

23 Dec '13

There has been a lot of concern expressed about updating submodules, so I've updated the applicable section of "Getting Started with Modular Boost Library Maintenance" to include reassurance that uncommitted changes will not be overwritten. See https://svn.boost.org/trac/boost/wiki/StartModMaint A table has been added describing the effects of git submodule update. Actual tests were run to determine the effects, using git 1.8.2.msysgit.0. See https://svn.boost.org/trac/boost/wiki/StartModMaint#Effectsofgitsubmoduleup… In looking at the table, a couple of things stand out: * git submodule update, like git checkout, will not normally overwrite uncommitted changes, regardless of whether the working copy is detached or on a branch. See the table for the specific error message. From the error message it appears checkout may actually be being called. * The case of an update applied to a submodule that is on branch master and at a commit behind the super-project's commit for that submodule just does a pull without a detach. That is the only case where a" checkout --detach" isn't done before a pull. I'm unsure whether this is a feature or a bug. Comments welcome, --Beman

1 0

generating documentation in modular boost
by Peter A. Bigot 23 Dec '13

23 Dec '13

I'm trying to update documentation in lib/utility but can't see how to regenerate the html files. Generalizing from the fact I can run the tests by invoking b2 within the module test subdirectory, I invoked b2 from within the doc subdirectory. This produced an error about an unknown feature "<format>" in Jamfile.v2. When I comment out those lines I do get xml files generated in a subdirectory of the superproject, but I expected html, matching what's been checked in. Is there documentation on building documentation that I haven't been able to find? Peter

3 3

Serializing a Graph Structure Using Boost
by Nouf M. 23 Dec '13

23 Dec '13

Hi, I would like to use boost to serialize my graph (it's in the boost adjacency list format)... I couldn't do it... I managed to serialize an array and some other structures as follows: int I[5] = {0,1,2,3,4}; boost::archive::text_iarchive ia(file), ia << I; but that doesn't work with graph data in the boost adjacency list format... when I write: ia << g //g is my graph I get the following compilation error: 'serialize' is not member of 'VertexProperties' 'VertexProperties' is a structure that I wrote as follows: struct VertexProperties { int id; int weight; }; I don't know how to add 'serialize' as a member function... I also tried "SavingArchieve" template but couldn't do it... Any idea's please? or references?

1 0

Access is denied for the "bootstrap" command
by Nouf M. 23 Dec '13

23 Dec '13

Hi, In order to use the serialization files... I need to build them... I have to go to the BGL directory from the command prompt and enter the "bootstrap" command... It works on some machines... But on my machine it gives me: "Building Boost.Build engine Access is denied. Failed to build Boost.Build engine. Please consult bootstrap.log for further diagnostics" My machine is an HP Workstation and I'm using Windows 7 professional edition... Why do you think I'm getting Access is denied?

1 0