- Boost - lists.boost.org

[outcome] Feedback on parts A and B of Tutorial mk III requested
by Niall Douglas 02 Feb '17

02 Feb '17

Dear Boost, I am truly hoping that this will be the last time I ask for feedback on Outcome's tutorial as this is the third complete rewrite. Many thanks for all the feedback so far, this Mk. III tutorial wouldn't look like this without all your help. Part A is 100% on pure expected<T, E> as currently before the standards committee. It has nothing to do with Outcome, but places Outcome in context. Part B is a polemic on why people should not use arbitrary type E in expected<T, E> and instead should restrain themselves to either E = std::error_code or E = std::exception_ptr. I'm requesting feedback on parts A and B of the tutorial which I think are mostly complete now. Specific feedback on these topics is particularly welcomed: 1. I've tried to make some of the code examples in Part A less contrived. Are there any obviously specific additional things I can do (note Parts B and C have more realistic programs)? 2. Are you convinced by the end of Part B that using arbitrary types for type E is a bad idea? If not, what did I fail to do in my explanation? Many thanks in advance. Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/

2 5

filtering_streambuf: Compressing a character buffer
by Adlai Shawareb 02 Feb '17

02 Feb '17

Hi, I have a character buffer ( "buffer" ) of known size ( "size_in_bytes" ), and would like to compress it using gzip_compressor and then get a pointer to the compressed buffer along with the compressed size. I've put together the following, but it says the first counter is 0 and the second counter is 2. Any help that can be offered is appreciated. namespace io = boost::iostreams; boost::asio::streambuf sb; std::ostream os(&sb); os.write( buffer, size_in_bytes ); io::filtering_streambuf<boost::iostreams::output> out; out.push(io::counter()); out.push(io::gzip_compressor()); out.push(io::counter()); out.push(os); io::close(out); // Needed for flushing the data from compressor std::ostream output(&out); // Maybe I can use rdbuf() to get the pointer to the stream?

1 0

[type_erasure] Inefficient extra shared_ptr copy when passing parameter of type _self
by Yuanchen Zhu 01 Feb '17

01 Feb '17

Hi List, I have been using boost::type_erasure a bit. I found that in the case of implementing equality_comparable, when it passes into operator== a placeholder argument corresponding to some any<C, T>, it always wraps it in a param<C, T>, which internally instantiates an any<C, T&>. While this instantiation does not cause the actual content to be copied, it always clone the share_ptr pointing to the dispatch table, causing an extra shared_ptr construct/destruct invocation whenever I do an equality comparison. This source of inefficiency seems unnecessary to me. Is it possible to modify param<C, T> such that it only stores non-ref counted raw reference to the original argument? Of course the functions in type_erasure::detail::access need to be updated accordingly. Will this modification break some grander scheme underlying type_erasure that I am unaware of, being new to the code base? It seems to me that when param<C, T> wraps a temporary, the temporary is guaranteed to live past the lifetime of the invocation. Can people vet this idea before I take a stab at changing the code? Thanks, Yuanchen

1 0

[serialization][uBlas] Serialization still breaking anything that uses uBlas?
by John Maddock 01 Feb '17

01 Feb '17

See for example: http://www.boost.org/development/tests/develop/developer/output/teeks99-02-… I'd really like to get these failures off my radar if possible... Thanks, John.

2 1

[testing] Need optimization help!!!
by Rene Rivera 01 Feb '17

01 Feb '17

If someone is interested in an optimization challenge here's your opportunity. Currently the program(s) that generate the test results take a long time to run. In both the live setup of a 4 core 16 GB ram VM it takes about 40 minutes to generate the reports for master and develop branches. On my 4 core 16 GB ram real hardware SSD HD machine it also takes about 40 minutes. If some brave soul could look into why it's so terribly slow the Boost community would be very grateful. As it would make it possible to move the processing to a CI service and improve results reliability and frequency. Unfortunately the only price I can promise for such a task is to buy you a meal and/or some drinks if I see you :-) If you are interested all that's needed to run the results is: a) A good chunk of disk space.. At least 27GiB. b) A bash shell, and Python, and git, and C++ tools (but you probably already have all that). b) An empty directory to run the reports. c) Download this script < https://github.com/boostorg/regression/blob/develop/reports/src/build_resul… >. d) Comment out the upload commands on lines 220-223. e) Run the script, and wait. And then the fun part of finding out what programs get run and what is slow :-) FYI.. The source for the processing is downloaded by the script. But you can find it here < https://github.com/boostorg/regression/tree/develop/reports>. -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

5 10

Re: [boost] Is there any interest in non-owning pointer-like types?
by Joseph Thomson 01 Feb '17

01 Feb '17

On 1 Feb 2017 7:51 p.m., "Olaf van der Spek" <ml(a)vdspek.org> wrote: On Wed, Feb 1, 2017 at 12:39 PM, Oswin Krause <Oswin.Krause(a)ruhr-uni-bochum.de> wrote: > On 2017-02-01 12:02, Olaf van der Spek wrote: >> >> On Wed, Feb 1, 2017 at 11:24 AM, Oswin Krause >> <Oswin.Krause(a)ruhr-uni-bochum.de> wrote: >>> >>> For me, the biggest problem of the proposal is that observer_ptr<T> is >>> implicitely constructed from T&. In my code I often use: >>> >>> Foo a; >>> Bar b(&a);//&a signals that b only references to a, but does not copy it. >>> >>> Now, when I write >>> >>> Bar b(a); //so is a now copied? >>> >>> While the interface is clearer in documentation, the usage is less clear! >> >> >> IMO using pointers instead of references to indicate something isn't >> copied is bad practice.. the language is C++, not C. > > > > This is not a C vs C++ thing. > > But assume it was. How would you indicate that in C++? It should be clear > without looking at the reference whether it is okay for a to go out of scope > before b or not. There's no way to indicate that, AFAIK. The way to indicate it is to require explicit conversion from `T&` to your pointer-like type. `make_observer` can play the role of a type-safe `&`. I would be happy to remove implicit conversion from `T&` if it is agreed to be undesirable. Joseph

1 0

Re: [boost] Boost XML Serialization Stability across library versions
by Adlai Shawareb 01 Feb '17

01 Feb '17

Thanks Robert, Regarding your suggestion to use text serialization. Since, over time, we will be adding and deleting data from the data set that we serialize, would the following be the right approach to managing the changes in the data that is de/serialized. I'm thinking that we would set the BOOST_CLASS_VERSION before de/serializing. { ar & mInt; If(version >= 2) ar & mFloat; If(version >= 3) ar & mInt2; ar & mFloat2; } Regards, Adlai On 1/25/17 12:04:00, Robert Ramey wrote: > Right. No one as under taken that task. Should be fairly straight > forward using spirit v3. > Robert Ramey

2 3

[container] ::flat_map function force_copy break strict-aliasing on gcc < 4.4
by Владислав Щапов 31 Jan '17

31 Jan '17

Hello! On function: template<class D, class S> BOOST_CONTAINER_FORCEINLINE static D force_copy(S s) { D *vp = reinterpret_cast<D *>(&s); return D(*vp); } gcc 4.3.4 (SLE 11) produced wrong code. It seems that the compiler does not create a copy of an object "s" and returns an object of D(). If I change the signature of function on "BOOST_CONTAINER_FORCEINLINE static D force_copy(const S& s)" the code works and the tests pass. But then does not create a copy of an object "s" in a function call. I'm not sure that this change is correct. Is there any other workaround for this problem with a compiler gcc 4.3 or better disable strict-aliasing? -- Best regards, Vladislav

2 2

C++Now 2017 Registration is Open
by Jon Kalb 31 Jan '17

31 Jan '17

Boost Devs, Registration for C++Now 2017 is now open: http://bit.ly/CppNow2017IsOpen The conference will be in Aspen, Colorado from May 15th to 20th. The deadline for submissions is this Friday: http://cppnow.org/2017-conference/announcements/2017/01/06/call-for-submiss… Jon

1 0

[CallableTraits] Still seeking Review Manager
by Barrett Adair 31 Jan '17

31 Jan '17

Hi, I'm still seeking a review manager for my CallableTraits library. I have spent hundreds of hours on this library, most of which went toward documentation and cross-platform testing. I think a small subset of the features in CallableTraits belong in the C++ standard. Occasionally at work, I have wished this were in Boost already. In short, this library eliminates the need to specialize templates for function signatures, and accounts for the *many* corner cases and cross-platform quirks that most users never care to do thoroughly themselves. I have 9 issues remaining on the GitHub page that would be resolved before an eventual acceptance, and I'd like to write some additional tests and documentation, but I think this library is ready for a review manager. Would anyone like to volunteer? https://github.com/badair/callable_traits Thanks, Barrett

1 0