- Boost - lists.preview.boost.org

[review] The review period of Boost.DoubleEnded has been extended with one week. Now ends 7th of October.
by Thorsten Ottosen 28 Sep '17

28 Sep '17

Dear users and members of Boost, People have been requesting extra time for the review of Boost.DoubleEnded. We therefore continue one week extra, so the final date will be October the 7th. kind regards Thorsten Original review announcement ------------------------------------------ I'm proud to announce that the formal review of Benedek Thaler's Boost.DoubleEnded library starts today, September 21, and runs through September 30. This library is the result of Benedek's participation in Google Summer of Code 2015. The library may be downloaded from https://github.com/erenon/double_ended and the documentation may be found here: http://erenon.hu/double_ended/ Anyone is welcome to post a review and/or take part in subsequent discussions (see below for review guidelines). Introduction ------------ This is a container library that provides two containers: A. boost::devector B. boost::batch_deque Both containers provides efficient push_front and push_back operations, hence the name DoubleEnded. The boost::devector class can be seen as an extension of std::vector with efficient support for push_front (so boost::devector provides contiguous storage). In addition, it provides a customizable small buffer optimization like boost::small_vector(*). It furthermore allows for customization of the growth factor such that the user can decide if a reallocation should allocate 1.5 or 2 or whatever more memory. And finally it provides new functionality that does not exist in normal vector implementations -- this functionality allow the class to excel in performance in certain situations like serialization or transfer of data from sources that require a preallocated buffer. The boost::batch_deque is similar to std::deque, but with two important twists. Like std::deque, it provides reference stability when calling push_front/push_back. Reference stability is crucial for programs that uses intrusive containers e.g. using Boost.Intrusive (**). boost::batch_deque has the following main advantages: 1. It allows the user to specify the segment size 2. It allows for external iteration over the segments Taken together, they provide a very efficient and flexible basis for intrusive containers. Access to the segments also increases performance for tasks such as serialization. Finally, boost::batch_deque may be seen as an efficient alternative to std::vector if you want to replace a percent-wise memory overhead with a constant memory overhead and have no need for contiguous storage. Review guidelines ----------------- Please provide in your review whatever information you think is valuable to understand your final choice of ACCEPT or REJECT including Boost.DoubleEnded as a Boost library. Please be explicit about your decision. Some other questions you might want to consider answering: - 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 which compiler(s)? 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? More information about the Boost Formal Review Process can be found here: http://www.boost.org/community/reviews.html Kind regards Thorsten, Review manager (*) boost::small_vector: https://tinyurl.com/ycslp4ot (**) http://www.boost.org/doc/libs/1_64_0/doc/html/intrusive.html -- Best regards, Thorsten Jørgen Ottosen, Ph.D. Director of Research +45 23308797 Dezide (www.dezide.com)

1 0

[admin] Turning on Issues tab on Github for Boost libraries
by Edward Diener 28 Sep '17

28 Sep '17

Can we not have some resolution from an administrator about turning on the Issues tab in Github for Boost libraries ? There are two 'admin' issues regarding this and, since it should be a quick job for an administrator to implement this, I would really prefer not to wait for a long time for this to get done. Just now I had a "bogus" PR for iostreams, which I rejected, and the explanation from the person who created the PR was that the PR was "created" because there was no Issue tab for the library and he was only trying to create an Issue to be looked at. There is absolutely no logical reason that I can think of why the Issue tab on Github cannot be turned on for Boost libraries, to benefit end-users.

3 3

Re: [boost] [ANN] Boost.UI - a new C++ GUI library
by Kolya Kosenko 28 Sep '17

28 Sep '17

On 09/28/2017 03:02 PM, Stefan Seefeld via Boost wrote: >> - uses native system-provided widgets > > what widgets do you consider "system-provided" ? What backends do you > bind to on the different platforms, and can those bindings be changed ? Native system-provided widgets means widgets from toolkits GTK+, Qt, Windows Controls, Cocoa etc. >> - compatible with other Boost libraries > > What does that mean ? Boost.Geometry and Boost.Polygon could be used with Boost.UI coordinates, Boost.Chrono - boost::ui::on_timeout() function, Boost.Date_Time library are used in date and time pickers etc. > I'd be interested in some high-level documentation explaining the basic > programming and execution model of the API. You can ask for implementation details here. Documentation explains only API. > For example, how are events > handled and how are they bound to particular actions ? Library subscribes to event handlers in native toolkit. When user press button, native event handler eventually calls Boost.UI event handler. > Are programs > inherently multi-threaded or single-threaded (e.g., does the main event > loop run in its own thread) ? There is main GUI thread and you should call GUI functions from other your own threads using boost::ui::call_async() function. Boost.UI library (like all other C++ GUI libraries) aren't thread safe because native toolkits are not thread safe. > Is your library mostly just a (thin) layer over existing > platform-specific GUI libraries, or does it reimplement a lot of the > logic itself ? Boost.UI is a rather thin layer with C++11/Boost support. > (And what is that cryptic "ui_main()" function ? :-) ) ui_main() function call required for GUI initialization, deinitialization, cleanup. It is required for underlying toolkits. Actually you can call some functions directly from main() function, but it is at your own risk. Regards, Kolya.

2 4

[1.65] [process] Library version outdated
by Andrey Semashev 28 Sep '17

28 Sep '17

Hi, It looks like something is broken in the release procedure, which resulted in all Boost 1.64.0, 1.65.0 and 1.65.1 to contain the same revision of Boost.Process 4075371cbe78dab8f43578c655747c831b554318 (you can see that by git tags). This revision is dated Wed Apr 5 11:32:27 2017, which is more like when 1.64.0 was being prepared. I know that 1.65 has long sailed, but can the release managers double check that this doesn't happen in 1.66? As a possible clue, on the local checkout I didn't have a head for the master branch for Boost.Process for some reason. I surely wasn't doing anythhing special with its repository, but maybe something happened on the remote side that caused this. I solved the problem locally by removing libs/process, .git/modules/process and running `git submodule update --init --recursive`.

3 3

[Pool] no maintainer, and should perhaps be deprecated?
by John Maddock 27 Sep '17

27 Sep '17

Moving this conversation to a new thread: >As an example of this: I recently proposed to add variadic move >construction for nodes to boost::pool (a trivial change). And even though >this has obvious advantages (and would get rid for C++11 capable compilers >of the use of M4 to generate the necessary code), this was received with: >file a PR, including tests (?) and change to documentation. Enough to put >me off, so I just changed it locally. Unfortunately the pool lib is not a good example - as previously stated it has no current maintainer - but as (probably) the last person to try and do some community maintenance on this library, the impression I came away with was that it needs rewriting. Actually more than that.... it probably needs redesigning if many more of the existing bug reports are to be fixed. I've actually said this on the list before, and suggested that Pool.2 might be a good candidate for as someone's first Boost library (since it's not too large), but so far no takers. There is possibly another issue: the performance of most std lib allocators has improved to the point that it's actually questionable whether using the pool lib has much real benefit (but this need quantifying so the cases where it does help are documented). Maybe that's what has put people off, or maybe it's just not sexy enough for anyone to take on, but I still hope that someone will.... Best, John. --- This email has been checked for viruses by AVG. http://www.avg.com

2 1

[multi_index] Taking advantage of C++17 auto template params to simplify key specification
by Joaquin M López Muñoz 27 Sep '17

27 Sep '17

Consider the following: struct element { int x; int f()const{return x;} }; int gf(const element& e){return e.x;} using container=multi_index_container< element, indexed_by< ordered_unique<member<element,int,&element::x>>, ordered_unique<const_mem_fun<element,int,&element::f>>, ordered_unique<global_fun<const element&,int,&gf>> > >; With the advent in C++17 of auto non-type template parameters, it is very easy to provide a boost::multi_index::key class template so that we can write: using container=multi_index_container< element, indexed_by< ordered_unique<key<&element::x>>, ordered_unique<key<&element::f>>, ordered_unique<key<&gf>> > >; (Example at http://coliru.stacked-crooked.com/a/5742c768cc093c03 ). Is there interest in adding this to Boost.MultiIndex? Comments on this or alternative syntaxes? Thank you, Joaquín M López Muñoz

5 9

Re: [boost] [review] The review of Boost.DoubleEnded starts today: September 21 - September 30
by Fletcher, John P 27 Sep '17

27 Sep '17

________________________________________ From: Thorsten Ottosen [tottosen(a)dezide.com] Sent: 26 September 2017 08:28 To: Fletcher, John P Subject: Re: [boost] [review] The review of Boost.DoubleEnded starts today: September 21 - September 30 >> Hi Fletcher, >> >> On another topic, I have not seen in the documentation any discussion of models for what is going on. I am aware of Chris Okasaki's book "Purely Function Data Structures" (1998) and a lot of subsequent literature. Does this work relate at all to previous work? >I'm not familiar with that book. You should think of standard C++ >containers like std::vector and std::deque which models a sequence concept: I should have given a reference. Here is one. http://okasaki.blogspot.co.uk/2008/02/ten-years-of-purely-functional-data.h… This is the authors comments on 10 years from publication. If you put the title into google you will get lots of information. My point is that there is a strong theoretical basis in this area and one of the issues for me is how far this has been taken into account in this new work. I was not until recently familiar with the work. It has helped me a lot understand the underlying problems in this area and I want to know whether the author of the library has looked at it. There is nothing in the documentation that I have seen about the wider issues. John Fletcher

2 1

Please fix executable permissions on header files
by Jonathan Wakely 25 Sep '17

25 Sep '17

The following files have executable permission bits set, which causes warnings for me when packaging Boost for Fedora (and is just wrong, because they're not executables): $ find boost_1_65_1/boost -name '*.hpp' -perm /111 boost_1_65_1/boost/numeric/conversion/detail/old_numeric_cast.hpp boost_1_65_1/boost/python/converter/arg_to_python.hpp boost_1_65_1/boost/python/converter/context_result_converter.hpp boost_1_65_1/boost/python/converter/pytype_function.hpp boost_1_65_1/boost/python/converter/return_from_python.hpp boost_1_65_1/boost/python/detail/convertible.hpp boost_1_65_1/boost/python/detail/copy_ctor_mutates_rhs.hpp boost_1_65_1/boost/python/detail/def_helper_fwd.hpp boost_1_65_1/boost/python/detail/is_shared_ptr.hpp boost_1_65_1/boost/python/detail/is_wrapper.hpp boost_1_65_1/boost/python/detail/nullary_function_adaptor.hpp boost_1_65_1/boost/python/detail/prefix.hpp boost_1_65_1/boost/python/detail/python_type.hpp boost_1_65_1/boost/python/detail/sfinae.hpp boost_1_65_1/boost/python/detail/unwind_type.hpp boost_1_65_1/boost/python/detail/unwrap_type_id.hpp boost_1_65_1/boost/python/detail/unwrap_wrapper.hpp boost_1_65_1/boost/python/detail/value_arg.hpp boost_1_65_1/boost/python/object/function_doc_signature.hpp boost_1_65_1/boost/python/object/inheritance_query.hpp boost_1_65_1/boost/python/object/stl_iterator_core.hpp boost_1_65_1/boost/python/arg_from_python.hpp boost_1_65_1/boost/python/base_type_traits.hpp boost_1_65_1/boost/python/borrowed.hpp boost_1_65_1/boost/python/cast.hpp boost_1_65_1/boost/python/def_visitor.hpp boost_1_65_1/boost/python/docstring_options.hpp boost_1_65_1/boost/python/handle_fwd.hpp boost_1_65_1/boost/python/object.hpp boost_1_65_1/boost/python/object_protocol_core.hpp boost_1_65_1/boost/python/raw_function.hpp boost_1_65_1/boost/python/refcount.hpp boost_1_65_1/boost/python/self.hpp boost_1_65_1/boost/python/stl_iterator.hpp boost_1_65_1/boost/python/wrapper.hpp boost_1_65_1/boost/interprocess/detail/managed_memory_impl.hpp boost_1_65_1/boost/interprocess/detail/managed_multi_shared_memory.hpp boost_1_65_1/boost/interprocess/detail/managed_open_or_create_impl.hpp boost_1_65_1/boost/interprocess/detail/shared_dir_helpers.hpp boost_1_65_1/boost/bimap/container_adaptor/unordered_associative_container_adaptor.hpp boost_1_65_1/boost/parameter/aux_/preprocessor/flatten.hpp boost_1_65_1/boost/parameter/aux_/preprocessor/for_each.hpp boost_1_65_1/boost/parameter/aux_/python/invoker_iterate.hpp boost_1_65_1/boost/parameter/aux_/overloads.hpp boost_1_65_1/boost/parameter/aux_/parameter_requirements.hpp boost_1_65_1/boost/parameter/aux_/result_of0.hpp boost_1_65_1/boost/parameter/aux_/tag.hpp boost_1_65_1/boost/parameter/aux_/template_keyword.hpp boost_1_65_1/boost/parameter/aux_/void.hpp boost_1_65_1/boost/parameter/aux_/yesno.hpp boost_1_65_1/boost/parameter/match.hpp boost_1_65_1/boost/parameter/parameters.hpp boost_1_65_1/boost/parameter/python.hpp boost_1_65_1/boost/parameter.hpp boost_1_65_1/boost/ptr_container/detail/default_deleter.hpp boost_1_65_1/boost/ptr_container/detail/is_convertible.hpp boost_1_65_1/boost/ptr_container/detail/move.hpp boost_1_65_1/boost/ptr_container/detail/scoped_deleter.hpp boost_1_65_1/boost/ptr_container/detail/throw_exception.hpp boost_1_65_1/boost/ptr_container/detail/void_ptr_iterator.hpp boost_1_65_1/boost/ptr_container/exception.hpp boost_1_65_1/boost/ptr_container/nullable.hpp boost_1_65_1/boost/thread/concurrent_queues/queue_base.hpp boost_1_65_1/boost/thread/experimental/parallel/v2/task_region.hpp boost_1_65_1/boost/thread/v2/shared_mutex.hpp boost_1_65_1/boost/iostreams/detail/adapter/device_adapter.hpp boost_1_65_1/boost/iostreams/detail/adapter/filter_adapter.hpp boost_1_65_1/boost/iostreams/detail/adapter/non_blocking_adapter.hpp boost_1_65_1/boost/iostreams/detail/broken_overload_resolution/forward.hpp boost_1_65_1/boost/iostreams/detail/config/enable_warnings.hpp boost_1_65_1/boost/iostreams/detail/config/limits.hpp boost_1_65_1/boost/iostreams/detail/config/windows_posix.hpp boost_1_65_1/boost/iostreams/detail/absolute_path.hpp boost_1_65_1/boost/iostreams/detail/bool_trait_def.hpp boost_1_65_1/boost/iostreams/detail/counted_array.hpp boost_1_65_1/boost/iostreams/detail/template_params.hpp boost_1_65_1/boost/iostreams/restrict.hpp boost_1_65_1/boost/msm/mpl_graph/mpl_graph.hpp boost_1_65_1/boost/typeof/dmc/typeof_impl.hpp boost_1_65_1/boost/typeof/std/bitset.hpp boost_1_65_1/boost/typeof/std/complex.hpp boost_1_65_1/boost/typeof/std/deque.hpp boost_1_65_1/boost/typeof/std/fstream.hpp boost_1_65_1/boost/typeof/std/functional.hpp boost_1_65_1/boost/typeof/std/iostream.hpp boost_1_65_1/boost/typeof/std/istream.hpp boost_1_65_1/boost/typeof/std/iterator.hpp boost_1_65_1/boost/typeof/std/list.hpp boost_1_65_1/boost/typeof/std/locale.hpp boost_1_65_1/boost/typeof/std/map.hpp boost_1_65_1/boost/typeof/std/memory.hpp boost_1_65_1/boost/typeof/std/ostream.hpp boost_1_65_1/boost/typeof/std/queue.hpp boost_1_65_1/boost/typeof/std/set.hpp boost_1_65_1/boost/typeof/std/sstream.hpp boost_1_65_1/boost/typeof/std/stack.hpp boost_1_65_1/boost/typeof/std/streambuf.hpp boost_1_65_1/boost/typeof/std/string.hpp boost_1_65_1/boost/typeof/std/utility.hpp boost_1_65_1/boost/typeof/std/valarray.hpp boost_1_65_1/boost/typeof/std/vector.hpp boost_1_65_1/boost/typeof/encode_decode.hpp boost_1_65_1/boost/typeof/encode_decode_params.hpp boost_1_65_1/boost/typeof/incr_registration_group.hpp boost_1_65_1/boost/typeof/int_encoding.hpp boost_1_65_1/boost/typeof/integral_template_param.hpp boost_1_65_1/boost/typeof/modifiers.hpp boost_1_65_1/boost/typeof/pointers_data_members.hpp boost_1_65_1/boost/typeof/register_mem_functions.hpp boost_1_65_1/boost/typeof/template_encoding.hpp boost_1_65_1/boost/typeof/template_template_param.hpp boost_1_65_1/boost/typeof/type_encoding.hpp boost_1_65_1/boost/typeof/type_template_param.hpp boost_1_65_1/boost/typeof/typeof_impl.hpp boost_1_65_1/boost/typeof/vector.hpp boost_1_65_1/boost/typeof/vector100.hpp boost_1_65_1/boost/typeof/vector150.hpp boost_1_65_1/boost/typeof/vector200.hpp boost_1_65_1/boost/typeof/vector50.hpp

5 7

request git repository for boost.contract
by Lorenzo Caminiti 25 Sep '17

25 Sep '17

Hello all, I need some help... who should I ask to create a git repository "contract" for boost.contract? How do I link that to the main boost repository in git? Please send me link if this info is already documented somewhere. Thank you. --Lorenzo

2 2

[review][Fit] Fit Review Results
by Matt Calabrese 25 Sep '17

25 Sep '17

Hello everyone, It is my privilege to announce that the Boost library count will once again be incremented as Paul Fultz II's library Fit has been accepted into Boost without conditions. Further details about this decision follow, but before that I'd like to thank Paul for all of the hard work he has put into Fit. Thanks also to everyone who participated in the review, with additional thanks to Vicente who was the review manager for the previous Fit review and who also provided a very detailed review and discussion during this review period. //////////////////////////////////////////////////////////////////////////////// Overview //////////////////////////////////////////////////////////////////////////////// Complete Reviews Considered: Zach Laine: Accept Vicente J. Botet Escriba: Accept John P. Fletcher: Accept Robert Ramey: Accept Jason Rice: Accept Barrett Adair: Accept Lee Clagett: Accept Glen Fernandes: Accept Louis Dionne: Accept Partial Reviews/Discussions Considered: Edward Diener Andrea Bocci Further Discussions (Previous Review): Steven Watanabe Peter Dimov Vladimir Batov Gavin Lambert Rob Stewart Bjorn Reese Manu Sanchez Artyom Beilis //////////////////////////////////////////////////////////////////////////////// Summary //////////////////////////////////////////////////////////////////////////////// Though not all responders reviewed every single aspect of the library, the reviews together adequately covered everything that is necessary for a decision to be made. This decision is made with additional confidence coming from usage of the library by some reviewers going back since during or before the initial review in early 2016. //////////////////////////////////////////////////////////////////////////////// Improvements Since Previous Review //////////////////////////////////////////////////////////////////////////////// This time around we received several more complete reviews than in the previous review period making it easier to be confident in a decision. A couple of notable changes since the 2016 review include: * Improved documentation * Default semantics for "capture" no longer try to determine whether to capture by reference or by value based on the argument being an lvalue or an rvalue (users can still get this via basic_capture) * Improved ADL awareness (internal calls are qualified) Throughout all of the discussions in this second review period, there were no explicit rejections of the library as a whole, though there were several recommendations from participants. A selection of those are listed here, though no changes are strictly required for acceptance: //////////////////////////////////////////////////////////////////////////////// Current Concerns //////////////////////////////////////////////////////////////////////////////// ////////// Design concerns: * Inheritance from user-callables has some potential pitfalls * Generated "infix" operators are "gimmicky" and don't add much value/may cause confusion * Placeholders should live in their own namespace ////////// Documentation concerns (not a complete list): * More examples requested * It is not always clear what files need to be #included for certain facilities in early examples * Better rationale for ConstFunctionObjects * Raw code is not linked from the examples ////////// Suggested name changes from reviewers: * Suggested rename of "partial()" to "partial_apply()" or "curry()" * Suggested rename of "conditional()" to "invoke_first()" or "call_first_of()" * Suggested rename of "apply" to "invoke" * Refer to "Callables" instead as "Invocables" (due to C++17 direction) * Library name "Fit" is not indicative of what it provides //////////////////////////////////////////////////////////////////////////////// Regarding the Author //////////////////////////////////////////////////////////////////////////////// Paul has been active in the Boost community for several years now. He is aware of his responsibilities going forward and I am confident that he will be a reliable maintainer for Fit for the foreseeable future (or whatever the library may be called). //////////////////////////////////////////////////////////////////////////////// Comments on the Name Boost.Fit //////////////////////////////////////////////////////////////////////////////// Note that what I say here is my personal opinion as a member of the community and does not necessarily reflect the opinions of others in the community nor of those in the steering committee. A few reviewers voiced that they would like to see a change of the library name "Fit". While multiple people did suggest this, my opinion is that this should be up to the author to choose whether or not he will change the name before it is added to boost and so I do not think it would be appropriate for us to strictly require or encourage such a change. If the author found the arguments against the name compelling, then he is welcome to change the name, but he should not feel obligated to do so. As a quick note on that, we have several high quality and highly regarded libraries in Boost with names that may not directly tell users what the library is for. Based on their popularity, I do not believe that it has significantly negatively affected their usage or ability to be found. A library name also may be important to an author as a creative outlet and I think it would be... poor form us to force such a change in this particular case and in other cases unless there are very strong reasons (i.e. the name is overly or intentionally misleading or there are legal concerns). I am not convinced that the current name is misleading enough to strictly require a change. As for existing precedent, some examples of libraries already in Boost with names that do not directly tell the user what the library is for include some of Boost's most highly regarded libraries such as Boost.Spirit (including Qi and Karma), Boost.Phoenix, Boost.Wave, Boost.Hana, and the recently accepted Boost.Beast. //////////////////////////////////////////////////////////////////////////////// Thanks again to everyone who participated in this review and in the previous review, and once again, congratulations to Paul on his library's acceptance into Boost. -- -Matt Calabrese

4 3