- Boost-users - lists.preview.boost.org

missing relaxed constexpr macros
by Gero Peterhoff 09 Dec '22

09 Dec '22

Hi John, i miss these for C++17/20/23 in boost/config. Can you retrofit this? thx Gero It could look like this: see https://en.cppreference.com/w/cpp/feature_test#__cpp_constexpr #if (__cpp_constexpr < 201703L) #define BOOST_NO_CXX17_CONSTEXPR #endif #if defined(BOOST_NO_CXX17_CONSTEXPR) #define BOOST_CXX17_CONSTEXPR #else #define BOOST_CXX17_CONSTEXPR constexpr #endif #if (__cpp_constexpr < 202002L) #define BOOST_NO_CXX20_CONSTEXPR #endif #if defined(BOOST_NO_CXX20_CONSTEXPR) #define BOOST_CXX20_CONSTEXPR #else #define BOOST_CXX20_CONSTEXPR constexpr #endif #if (__cpp_constexpr < 202207L) #define BOOST_NO_CXX23_CONSTEXPR #endif #if defined(BOOST_NO_CXX23_CONSTEXPR) #define BOOST_CXX23_CONSTEXPR #else #define BOOST_CXX23_CONSTEXPR constexpr #endif

1 0

Does Boost.Multiprecision have a forwards declaration header?
by Ivan Matek 04 Dec '22

04 Dec '22

I presume not since I could not find it in docs, but decided to double check. My motivation is that I would like to limit heavy includes in my header(since intrinsics for avx512 files are 20k+ LOC on my system), but I would still like to use proper types in signature of my functions (since none of them take .args by value fwd declarations would be enough). So I was hoping for something like iosfwd. If it matters what part of library I am using: cpp_int.hpp P.S. I checked if telling compiler that I do not want arch support for avx512 would help, but it looks like avx512 intrinsics headers are included regardless, and even if it was not I would still like to not get into issues when compiling for HW that supports avx512. :) regards, Ivan

2 2

[release] Boost 1.81.0 beta 1 released
by Marshall Clow 17 Nov '22

17 Nov '22

Boost release 1.81.0 beta 1 is now available at: https://boostorg.jfrog.io/artifactory/main/beta/1.81.0.beta1/source/ The SHA256 checksums are: 135f03965b50d05baae45f49e4b7f2f3c545ff956b4500342f8fb328b8207a90 boost_1_81_0_b1.tar.gz 736b041748d04c573d145613161cf2e54beb210c5511b67e6fac5510b82b6646 boost_1_81_0_b1.7z 941c568e7ac79aa448ac28c98a5ec391fd1317170953c487bcf977c6ee6061ce boost_1_81_0_b1.tar.bz2 c58d50acf58a98d4f5a7d891257d154efa50ca90312ca446586c7b4f9b7d3f14 boost_1_81_0_b1.zip Release notes (still need updates): https://www.boost.org/users/history/version_1_81_0.html Please try the beta and report any problems you encounter. Thanks to everyone who contributed to the release. -- Boost Release Team

1 0

Partial serialization of a program
by Markus Lavin 06 Nov '22

06 Nov '22

I have a simulator type program where I want to be able to save (and load) the current simulation state using Boost serialization. The program, as with most programs, consists of a part that is static during run-time and a part that is dynamic. Ideally I would like to serialize only the dynamic part but this however raises the following problem. If a serialized object in the dynamic part contains a pointer to a non-seralized object in the static part then how do I deal with that. Clearly the address cannot be serialized since it may change between builds and even between runs (ASLR). The program does contain a 'registry' though where all relevant static objects are registered with a unique identifier. So question is would it somehow be possible to have Boost serialization use this registry to serialize the unique identifier instead of trying to serialize the entire object when one of those pointers are serialized? As far as I can tell the documentation does not mention this situation and extensive web searching yields very little as well. Any advice appreciated! -Markus

3 4

Question about boost::multi_index emplace operations
by Pavel Vazharov 03 Nov '22

03 Nov '22

Hi there, Is there a way to emplace (key, value) in boost multi_index hashed index in such a way that the value is not moved/stolen if the key is already present? As far as I checked the implementation for emplace (and emplace_) the node is eagerly created and the key and the value are moved into the node and then the insertion operation takes place. If the insertion fails the node is destroyed but the outside value is lost/stolen in all cases. What I actually need is to create the value only if the key is not present, because the creation is expensive, and I'd like to avoid the find+emplace or insert empty value + modify call. (As far as I checked, the value can't be modified directly from the returned iterator because it's const and will need const_cast to work). I thought to use emplace with some machinery for lazy creation struct creator { operator value_type () { return ...; } // create the value lazily, on demand }; The creator is passed to emplace instead of the actual value. However, this doesn't help because the node is eagerly created and the above operator is always called. Thanks, Pavel.

2 4

Re: [Boost-users] Cannot open boost lib file when building CLI dll
by Michael Pogue 28 Oct '22

28 Oct '22

> > I am trying to build a CLR dll file in VS22 and dotnet 6.0 that references >> boost. When I build, I get the following error. >> LNK1104 cannot open file 'libboost_program_options-vc142-mt-x64-1_76.lib' >> >> - This file does exist. >> - I have restarted to make sure the file isn't locked. >> - I have tried building as 32 and 64 bit. >> - I have tried 3 different versions of boost >> >> The problem is reproduced by this project: >> https://github.com/MajorMajor0/boostRepro. >> Any help would be much appreciated. >> >

1 0

Norms methods
by General Magic - Teodora Bogdan 26 Oct '22

26 Oct '22

Hello, I was using the Boost library and I wanted to call the l2_norm() method from "boost/math/tools/norms.hpp", but is says that the namespace boost::math::tools doesn't contain l2_norm() method. When I looked into it, I saw that the namespaces at the beginning of the file were written"boost::math::tools" instead of "namespace boost { namespace math { namespace tools {...". Can this be fixed? Have a nice day, Teodora Bogdan CONFIDENTIALITY NOTICE: This e-mail and any attachments hereto may contain information that is privileged or confidential, and is intended for use only by the individual or entity to which it is addressed. Any disclosure, copying or distribution of the information by anyone else is strictly prohibited. If you have received this document in error, please notify us promptly by responding to this e-mail. DISCLAIMER: General Magic accepts no liability for the content of this email, or for the consequences of any actions taken on the basis of the information provided, unless that information is subsequently confirmed in writing and duly signed on behalf of General Magic by an authorized officer. Any views or opinions presented in this email are solely those of the author and do not necessarily represent those of General Magic. No employee or agent is authorized to conclude any binding agreement on behalf of General Magic with another party by email without express written confirmation which is duly signed on behalf of General Magic by an authorized officer. Thank you.

3 3

Boost::serialization detection of Cxx Standard
by Georg Gast 14 Oct '22

14 Oct '22

Hi, I developed a patch and a test to serialize std::pmr::(w)strings in all 3 archive types. The test I set the requirement to <cxxstd>17. It works so far but when I compile the library for using it in my application the "hidden" functions behind "BOOST_NO_CXX17_MEMORY_RESOURCE" are not found by the linker as the define is set when the library is compiled. How to work around this? Bye Georg

2 2

Is there any official documentation on shutdown procedure for ASIO SSL socket?
by Ivan Matek 11 Oct '22

11 Oct '22

There is this SO answer https://stackoverflow.com/a/25703699/ with Vinnie comment from 8y ago that it should be added to documentation. :) I can not find anything in the documentation so I presume the documentation/example does not exist, but I decided to double check. regards, Ivan

3 2

Very slow destructors of binary_oarchive and filtering_streambuf
by boost.5nyej＠simplelogin.com 03 Oct '22

03 Oct '22

Hello, I just posted a question on StackOverflow, but I realised that this mailing list might be a better place. The link to the question is: https://stackoverflow.com/questions/73938656/very-slow-destructors-of-binar… Tl;dr; I have a code that serializes a big object by using binary_oarchive and filtering_bufstream. The writing on its own takes around 28 minutes, but the deconstructors of the mentioned classes take 37 and 36 minutes, respectively. It is very strange to me that they would take so much time to run, as the first deconstructor adds only 2KiB of data, and the second deconstructor doesn't change anything. Could you help me find the root cause of those execution times?

1 0