March 2017 - Boost - lists.boost.org

Re: [boost] [release] 1.64.0 please test master snapshot
by Rene Rivera 22 Mar '17

22 Mar '17

On Tue, Mar 21, 2017 at 12:36 PM, Tom Kent <lists(a)teeks99.com> wrote: > It looks like the boostorg/boost:master commit this snapshot was based off > doesn't have the latest boostorg/config:master submodule commit. I think > updating this is a manual process at this point in the release cycle? (Thus > the autolink file is still looking for vc150). > Grr.. Forgot to do that for config. I did it for build though :-) It's done now.. Just need to wait again for the new archives. > As an aside, I had a difficult time figuring out what the snapshot I > downloaded referred to. I was able to use the boost_1_64_0_snapshot.tar.bz2.json > file to determine that I had the correct boost_1_64_0_snapshot.tar.bz2, but > bintray doesn't make it super clear what git commit this snapshot download > refers to. > > Would it be possible to add a file to snapshots/releases that contains the > git commit hash? > Can't promise I'll do it right this second... But I'll add the commit hash to the json data for the archives. -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

7 14

Boost MSM parallel behavior with delayed self-transitions?
by Christophe Henry 22 Mar '17

22 Mar '17

Hi Danylo, >Hi community, > > I am using Boost MSM (basic and functor front-ends) and am trying to >implement the following state machine (https://i.stack.imgur.com/K79qk.png ): Ok, State1 is itself a state machine (also called composite). See: http://www.boost.org/doc/libs/1_63_0/libs/msm/doc/HTML/ch02s02.html#d0e151 and: http://www.boost.org/doc/libs/1_63_0/libs/msm/doc/HTML/ch03s02.html#d0e529 Then you have 2 active states, which are called orthogonal regions. See: http://www.boost.org/doc/libs/1_63_0/libs/msm/doc/HTML/ch03s02.html#d0e577 >I am using Boost MSM (basic and functor front-ends) and am trying to >implement the following state machine: enter image description here > >In words: < snip > > >I would like to know the way to create this state machine in Boost MSM. The above link has an example. >There are two tricks here which I cannot figure out how to do: > > - Execution in parallel (i.e. running action_A does not stop running >action_B at the same time) It depends what you mean with "in parallel". Having 2 active states means that the actions execute only logically at the same time, meaning within the same process_event call. But as you have just one thread, they will execute after each other. If you mean in different threads, then no, the library does not support this. But my next library, Asynchronous does ( https://github.com/henry-ch/asynchronous) I have examples with state machines, timer, etc. >- Delayed transition (i.e. the transitions which happen after 2 >seconds for state A and after 5 seconds for state B). Neither delay >should be blocking! The transitions should just "fire" after this time. Sorry, this is out of scope of the library as it would involve threads. You can do one of the following: - use your own timer (Boost.Asio offers one) and handle thread safety yourself. - have a look at Asynchronous. Sorry if it looks like advertising, but the library was written for this kind of things. >Thank you very much for helping, > >Danylo. HTH, Christophe

3 4

Re: [boost] [gsoc17] Project1: Static map: Can I implement the competency test in g++-7?
by Vic Luo 22 Mar '17

22 Mar '17

Hi, In the last few days I finished a toy implementation with continuous integration at https://github.com/htfy96/gsoc-static-map, which has been tested on g++5/6 and clang++3.8/3.9. Current test failure is as expected due to g++-5's incomplete constexpr optimization. Besides, my proposal draft has been submitted which includes a very tiny benchmark with boost::hana on compile-time lookup performance on map<int, int> structure. Any kind of suggestion is welcomed at https://gist.github.com/htfy96/0bd9b3200d9a0e67f449edc40bfa43b5*. * Thanks,* * -- --------------------------------- Vic Luo Shanghai Jiaotong University Key fingerprint 0x98809ca08bf5662a

1 0

Re: [boost] [release] 1.64.0 please test master snapshot
by Olaf van der Spek 22 Mar '17

22 Mar '17

On Wed, Mar 22, 2017 at 12:45 PM, Andrey Semashev <andrey.semashev(a)gmail.com> wrote: >> Right, boostorg doesn't have tickets enabled though.. where should I >> open a ticket? > > > Probably, in Trac: I hate Trac. > https://svn.boost.org/trac/boost/newticket?component=algorithm;version=Boos… > > Although since it's fixed in develop and there is a PR for merge to master, > I'm not sure there's much point in a ticket. I meant in general. -- Olaf

1 0

Re: [boost] [release] 1.64.0 please test master snapshot
by Andrey Semashev 22 Mar '17

22 Mar '17

And once again, this went to Olaf directly instead of the list. Re-sending. On 03/22/17 14:45, Andrey Semashev wrote: > On 03/22/17 13:45, Olaf van der Spek wrote: >> On Wed, Mar 22, 2017 at 11:05 AM, Andrey Semashev via Boost >> <boost(a)lists.boost.org> wrote: >>> On 03/22/17 11:59, Olaf van der Spek via Boost wrote: >>>> >>>> VC2015 /std:c++latest: >>>> >>>> struct to_lowerF : public std::unary_function<CharT, CharT> >>>> >>>> boost\algorithm\string\detail\case_conv.hpp(33): error C2143: syntax >>>> error: missing ',' before '<' >>> >>> >>> This seems to be fixed in develop but not merged to master. I've >>> created a >>> PR to do the merge: >>> >>> https://github.com/boostorg/algorithm/pull/34 >>> >>>> https://github.com/mclow/String-Algo/issues/7 >>> >>> >>> That ticket is for a fork, not the boostorg repo. >> >> Right, boostorg doesn't have tickets enabled though.. where should I >> open a ticket? > > Probably, in Trac: > > https://svn.boost.org/trac/boost/newticket?component=algorithm;version=Boos… > > > Although since it's fixed in develop and there is a PR for merge to > master, I'm not sure there's much point in a ticket. >

1 0

Re: [boost] GSoC proposal
by Vissarion Fisikopoulos 22 Mar '17

22 Mar '17

Hi Rasul, thanks for your interest in Boost.Geometry. I guess you have already read this https://svn.boost.org/trac/boost/wiki/SoC2017 There are two possibilities: either you provide a library (with some standards described in the webpage) or you do the test. My opinion is that doing the test (even if not needed) will first strengthen your application and second will give you a good overview of the problem you are going to work on, if selected. >From your repo what is the library that you propose for your application and meets the requirements described in the webpage above (e.g. no coursework, more than 6 months old etc)? The tests are described here: https://svn.boost.org/trac/boost/wiki/SoC2017#a3.Boost.Geometry Depending on what (?) project you prefer to pursue there is a corresponding test. Best, Vissarion ps. I copy the message to boost list so the community can follow and provide feedback if needed On 22 March 2017 at 12:00, Rasul Kerimov <rkerimov(a)std.qu.edu.az> wrote: > Hey Mr. Fysikopoulos, > > I'm really interested in Boost GSoC 2017 project proposal, but I don't > know how to pass Programming competency Test. > > I have a github repo with all my codes that I wrote in C++, but I > don't know if it's enough to pass test. > > Github page: https://github.com/CoderINusE/ > Github repo: https://github.com/CoderINusE/Algorithms-data-structues > > I will be grateful for any response. Thank you!

1 0

[release] 1.64.0 Delayed because of Microsoft (version numbers)
by Rene Rivera 21 Mar '17

21 Mar '17

Sorry to say the beta release of 1.64.0 is hereby going to be delayed for at least another week. The confusion of VS2017 version numbers and the communities inability to address this issue after the beta is to blame. To address the situation the beta release will be published when one of the two tasks below are completed: (A) Boost Config changes to use 1410 as the autolink version to match what the currently working generated libs are taggeg with. Or: (B) The community decides what version number we should be using and the appropriate PRs are filed and applied and tested for all the repos involved. Which means: 1. The community has until 12:00 CDT US March 18 to decide on what the vc version number and tag should be. 2. The community must post PRs for the places where we reference VS2017 version numbers to the new number. Those are at least: <https://github.com/boostorg/config> <https://github.com/boostorg/build> <https://github.com/boostorg/boost> <https://github.com/boostorg/website> People have until 17:00 CDT Monday March 20 to submit those PRs. 3. The community can test the resulting master snapshot when available at the latest 24:00 Monday March 20. 4. If all goes well we can have a beta release on Wednesday March 22 (ie one week late). -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

17 26

[stacktrace] Review
by Peter Dimov 21 Mar '17

21 Mar '17

I think that the Stacktrace library should be ACCEPTED. The design is good. It's unfortunate that to get a stacktrace class that stores the capture internally in a fixed-size array, as the first iteration did, one has to write a special allocator, more so because such an allocator isn't readily available (we don't have one in Boost, for instance). This could easily be addressed by templating basic_stacktrace on a container (default std::vector<frame>) instead, which would allow boost::container::static_vector to be used. The implementation is good. The main problem here, as already discussed in detail, is that under Windows clients of the library that don't make use of the facilities that require windows.h and COM headers to be included still suffer the penalty of this inclusion. Fixing this will be a significant increase in usability. The documentation is adequate, although it could use some editing by a native speaker. The library is potentially very useful. Assert handlers, exceptions and termination handlers can (and should) all make use of it. Wide use under Windows is however hampered by the limitation described above. I tested the library under Windows with msvc-14.0, 14.1, Cygwin g++/clang++. Under Cygwin there is no libbacktrace, but defining BOOST_STACKTRACE_USE_ADDR2LINE worked. I invested a fair chunk of time into my evaluation of the library, and am reasonably knowledgeable about the problem domain.

1 0

Re: [boost] [smart_ptr] Is there any interest in unique_ptr with type erased deleter?
by Andrey Davydov 21 Mar '17

21 Mar '17

On Tue, Mar 21, 2017 at 4:11 PM, Andrey Semashev <andrey.semashev(a)gmail.com> wrote: > On 03/21/17 14:54, Andrey Davydov wrote: > >> On Tue, Mar 21, 2017 at 2:15 PM, Andrey Semashev via Boost >> <boost(a)lists.boost.org <mailto:boost@lists.boost.org>> wrote: >> >> On 03/21/17 12:54, Andrey Davydov via Boost wrote: >> >> On Tue, Mar 21, 2017 at 12:00 PM, Richard Hodges >> <hodges.r(a)gmail.com <mailto:hodges.r@gmail.com>> wrote: >> >> What is required, more than simply creating a type-erased >> deleter? >> >> >> Also deleter must capture pointer, because pointer which was >> passed to >> constructor can differ from the pointer which will be deleted. >> For example, >> >> struct Base { /* ... */ }; >> struct Derived : Base { /* ... */ }; >> >> ptr<Derived> p1(new Derived); >> void * raw_p1 = p1.get(); >> ptr<Base> p2 = std::move(p1); >> void * raw_p2 = p2.get(); >> assert(raw_p1 != raw_p2); >> >> >> The above piece of code strikes me as very unsafe and better be >> avoided in the first place. Designing new tools to handle this case >> is IMHO misguided. >> >> Why this code is very unsafe? If change `ptr` to `shared_ptr` this will >> become perfectly valid and widely used pattern. >> > > Because comparing the void pointers is not safe in general. The above code > in particular will probably work (although not sure if even that's > guaranteed, depending on whether Base and Derived qualify as standard > layout classes), but if the class hierarchy is different, that code will > silently break. I agree, I wrote this code just to demonstrate that deleter must capture pointer because it can be changed due to upcasting. > But, of course, nothing stops you from saving the state in the deleter. > > struct poly_deleter > { > template< typename T > > poly_deleter(T* p) : m_p(p), m_delete(&poly_deleter::do_delete< T >) > { > } > > void operator() (void*) const noexcept > { > m_delete(m_p); > } > > private: > template< typename T > > static void do_delete(void* p) noexcept > { > delete static_cast< T* >(p); > } > > void* m_p; > void (*m_delete)(void*); > }; > > This will work even if you form std::unique_ptr<void, poly_deleter> (i.e. > don't have a universal base class). > Your poly_deleter implementation is the same as I used inside my smart pointer implementation, but there is an issue with std::unique_ptr<T, poly_deleter>, namely it is not constructible from std::unique_ptr<T, std::default_delete<T>> because poly_deleter must capture pointer and so it is not constructible from std::default_delete. -- Andrey Davydov

4 5

[smart_ptr] Is there any interest in unique_ptr with type erased deleter?
by Andrey Davydov 21 Mar '17

21 Mar '17

It is widely known that deleter type is part of std::unique_ptr type. It is the most effective decision, but sometimes not very convenient. The most important use cases (for me) when std::unique_ptr<T, std::default_delete<T>> doesn't work are following: 1. unique_ptr to incomplete class struct MyClass; unique_ptr<MyClass> create(); auto ptr = create(); // compilation error 2. upcast to base without polymorphic destructor struct Base { // ... // no virtual ~Base() here }; struct Derived : Base { // ... }; unique_ptr<Base> ptr = make_unique<Derived>(); // compiles by leaks! 3. unique_ptr with non-default deleter unique_ptr<FILE, int (FILE *)> open_file(std::string const & path) // works but looks ugly, why should the fact that fclose returns `int` be visible from the signature of the function `open_path`? { return { std::fopen(path, "r"), &std::fclose }; } Of course, all this examples could be fixed if unique_ptr would be replaced by shared_ptr, but the semantic of the shared ownership is not desirable often. I have implemented smart pointer with semantics of unique ownership and type erased deleter (https://github.com/AndreyG/unnamed) It has size of 3 pointers and it doesn't require additional memory allocations for the case when deleter is empty class (for example, std::default_delete or non-capturing lambda) and for the case when deleter is known at compile time function (for instance, fclose). The are some simple examples of usage and tests in the repo. Why don't just use unique_ptr<T, function<void (void *>>? The main reason is that due to the small object optimization std::function has rather big size. For instance, sizeof(unique_ptr<T, function<void (void *>>) == 72 for MSVC x64! Results for the other platforms can be found in the repo README. Does it seem useful? -- Andrey Davydov

5 11