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

[python] to_python_converter , is by value needed?
by MM 04 Dec '23

04 Dec '23

Hello, I have posted a question here some time ago, I thought to try here as well https://stackoverflow.com/questions/77552946/to-python-converter-is-by-valu… The community of experts in boost.python seems to have moved off to other concerns? Rds,

1 0

Could sprintf be replaced with snprintf?
by McGrath, Justin M 30 Nov '23

30 Nov '23

I am using some Boost libraries in a code base that does not want any use of sprintf. An automatic test flags any calls to it or vsprintf. Is it possible to replace all uses of sprintf with snprintf? I really doubt there are actually any security issues here, but I'm hoping this wouldn't be too difficult or cause any problems other than the effort to do it. Cheers, Justin

2 6

[Boost.Test] Saving Context Information for Failed Test Cases
by Bingtao Yin 16 Nov '23

16 Nov '23

I am a user of Boost.Test, and I am currently looking for a way to record some context information when dealing with failed test cases. Specifically, I would like to save certain variables to a file to facilitate debugging Could you please advise on any methods or recommendations within Boost.Test to fulfill this requirement? Are there any best practices or example code snippets that you could share? Thank you very much for your assistance! Best regards,

1 0

Re: [Boost-users] Problem with named_mutex on Windows
by David C. Partridge 15 Nov '23

15 Nov '23

I just realised it's not hanging - it's thrown an interprocess exception. I have revised my code to capture and report on that. I will let you know what it says. David From: David C. Partridge <david.partridge(a)perdrix.co.uk> Sent: 13 November 2023 14:16 To: 'boost-users(a)lists.boost.org' <boost-users(a)lists.boost.org> Subject: Problem with named_mutex on Windows I have a user reporting that the following code in my application's main() hangs unless he uses "Run as Adminstrator": ZTRACE_RUNTIME("Checking Mutex"); bip::named_mutex dssMutex{ bip::open_or_create, "DeepSkyStacker.Mutex.UniqueID.12354687" }; bip::scoped_lock<bip::named_mutex> lk(dssMutex, bip::defer_lock); const bool firstInstance{ lk.try_lock() }; ZTRACE_RUNTIME(" firstInstance: %s", firstInstance ? "true" : "false"); The trace shows "Checking mutex" and then it all stops. Is my code in error and if so what do I need to change? If not, how can I determine what the problem is? On a related issue: I think that this is C++ Exception safe, as the mutex will be unlocked and deleted. However, if a C exception occurs then how should I clean up? Is the named_mutex persistent over reboots? If so, where is it held (in case I need to do a manual cleanup). Many thanks David

1 1

Problem with named_mutex on Windows
by David C. Partridge 13 Nov '23

13 Nov '23

I have a user reporting that the following code in my application's main() hangs unless he uses "Run as Adminstrator": ZTRACE_RUNTIME("Checking Mutex"); bip::named_mutex dssMutex{ bip::open_or_create, "DeepSkyStacker.Mutex.UniqueID.12354687" }; bip::scoped_lock<bip::named_mutex> lk(dssMutex, bip::defer_lock); const bool firstInstance{ lk.try_lock() }; ZTRACE_RUNTIME(" firstInstance: %s", firstInstance ? "true" : "false"); The trace shows "Checking mutex" and then it all stops. Is my code in error and if so what do I need to change? If not, how can I determine what the problem is? On a related issue: I think that this is C++ Exception safe, as the mutex will be unlocked and deleted. However, if a C exception occurs then how should I clean up? Is the named_mutex persistent over reboots? If so, where is it held (in case I need to do a manual cleanup). Many thanks David

1 0

Optimizing serialization of std containers
by Gidi Gal 05 Nov '23

05 Nov '23

Hello, Are there any settings or additional code required to optimize the serialization of std containers ? I serialize a nested container: std::vector<std::deque<int>>. The max size of each deque entry is 20. I include boost/serialization/deque.hpp and boost/serialization/vector.hpp and I serialize the container in the following way: std::ofstream ofs(outFileName); boost::archive::binary_oarchive oa(ofs); oa << data; I created data with 10600 integers in the deque. The size of the serialized output file is 54493. Now, I test an alternate serialization. I serialized the size of the vector and then per each entry I serialize its size, which cannot be more than 20, so char is enough for that, next I serialize the integers in this entry. The size of the serialized output file is 43484. Is there a way to achieve better results for the default serialization ? Thanks, Gidi

1 0

why diagnostic_information is std::exception::what: Unknown exception
by Bernard 30 Oct '23

30 Oct '23

greetings kind regards may i please inquire how not to obtain "std::exception::what: Unknown exception" from diagnosic_inforumation() . my meaning is to inquire how to obtain as per example on boost web page https://www.boost.org/doc/libs/1_83_0/libs/exception/doc/tutorial_diagnosti… output : "std::exception::what: example_io error" id est where did "example_io error" come from as neither boost::exception nor std:::exception accept constructor string arguments . please my code below . thank you kindly #include "stl" // #includes all stl #include files #include "boost\exception\all.hpp" using namespace std; struct ceXception : virtual boost::exception, virtual std::exception { using this_type = ceXception; ceXception() {} auto what() { cerr << __FUNCSIG__ << endl; return string("return from " + string(__FUNCSIG__)); } template<typename charType> friend std::basic_ostream<charType>& operator<<(std::basic_ostream<charType>& _ostream, const this_type&) { cerr << "to cerr from " << __FUNCSIG__ << endl; _ostream << "to ostream from " << __FUNCSIG__ << endl; return _ostream; } }; void foo() { try { BOOST_THROW_EXCEPTION(ceXception()); } catch (ceXception& e) { cerr << boost::diagnostic_information(e); } } int main() { foo(); return 0; }

2 2

Async review scheduled for September 22, 2023 - October 2, 2023
by Mateusz Loskot 07 Sep '23

07 Sep '23

Hi, Following request from Nial and Klemens, we have scheduled new round of the Async review for September 22, 2023 - October 2, 2023 https://www.boost.org/community/review_schedule.html Best regards, -- Mateusz Loskot, on behalf of the Review Wizards

1 0

Error LNK1104 cannot open file 'boost_unit_test_framework-vc143-mt-x64-1_82.lib' cal_data_gain_pwr_test C:\Users\mushtaq.syed\Downloads\uhd-master\host\build\tests\LINK 1
by Mushtaq A. Syed, Ph.D. 28 Aug '23

28 Aug '23

Hi: My project uses the Boost library. I have installed C:\Program Files\Boost\boost_1_82_0. I also built using the bootstrap and .\b2 commands from the command window. This was because all the libboost libraries in C:\Program Files\Boost\boost_1_82_0\stage\lib were missing. Now when I build my code using Visual Studio 2022, I am getting the above error. I searched my Boost root folder and cannot find boost_unit_test_framework-vc143-mt-x64-1_82.lib; however, I do see libboost_unit_test_framework-vc143-mt-x64-1_82.lib and other libboost libraries but not a single boost library. I don’t know why it is looking for boost_unit_test_framework-vc143-mt-x64-1_82.lib and not libboost_unit_test_framework-vc143-mt-x64-1_82.lib. Could you please help resolve this? Thanks a lot! Cheers, Mushtaq ---------------------------- Mushtaq Syed, Ph.D. Principal Digital Signal Processing Engineer Verus Research [cid:image5fd716.JPG@ebfac5a4.47a0da54]<http://> Direct 505-396-7109 Cell 408-514-7821 6100 Uptown Blvd NE, Suite 260 Albuquerque, New Mexico 87110 verusresearch.net ---------------------------- NOTICE: This electronic mail message is intended exclusively for the individual or entity to which it is addressed. This message, together with any attachment, may contain sensitive or privileged information. Any unauthorized review, use, printing, retention, copying, disclosure, or distribution of said email is strictly prohibited. If you have received this message in error, please (1) immediately advise the sender by reply email message and (2) delete all copies of this message

2 3

Quadmath.h not found for float128 arm64 rocky linux
by David Frank 25 Aug '23

25 Aug '23

When compile with boost float128 on arm64 rocky linux, i got a failure of missing quadmath.h Thanks.

3 2

[Compiler] Boost compatibility in a untested compiler version
by Kankanamge, Gimhani 17 Aug '23

17 Aug '23

Hi team, Based on release notes for boost 1.83.0, there is a list of tested compilers (https://www.boost.org/users/history/version_1_83_0.html) But, we are going to use boost 1.83.0 in a different compiler which is not listed here(GCC 8.5.0). In such a case, what is the approach we should use to verify that boost is compatible in our given platform? Thank you. Gimhani Uthpala ------------------------------------------------------------------------------------------------------------ Please read these warnings and restrictions: This e-mail transmission is strictly confidential and intended solely for the ordinary user of the e-mail address to which it was addressed. It may contain legally privileged and/or CONFIDENTIAL information. The unauthorised use, disclosure, distribution and/or copying of this e-mail or any information it contains is prohibited and could, in certain circumstances, constitute a criminal offence. If you have received this e-mail in error or are not an intended recipient please inform London Stock Exchange Group (“LSEG”) immediately by return e-mail or telephone 020 7797 1000. LSEG may collect, process and retain your personal information for its business purposes. For more information please see our Privacy Policy. We advise that in keeping with good computing practice the recipient of this e-mail should ensure that it is virus free. We do not accept responsibility for any virus that may be transferred by way of this e-mail. E-mail may be susceptible to data corruption, interception and unauthorised amendment, and we do not accept liability for any such corruption, interception or amendment or any consequences thereof. Calls to London Stock Exchange Group may be recorded to enable LSEG to carry out its regulatory responsibilities. For more details on the LSEG group of companies click here London Stock Exchange Group plc 10 Paternoster Square London EC4M 7LS Registered in England and Wales No 05369106 ------------------------------------------------------------------------------------------------------------

2 1

JSON and no-alloc
by Marc Oscar Singer 14 Aug '23

14 Aug '23

Greetings, I'm targeting the boost JSON library for an embedded target. There appears to be one place where there could be either a beneficial new feature, or the feature exists and I'm not seeing it. The section about Avoiding Dynamic Allocation is close to what we'd like to see, except that we don't want to depend on a stack buffer for value data. All of the text of the JSON message is present in the string view. It would be desirable to reference the values from there. Let me be a little more explicit. We receive a message in JSON format. The whole of this message must be received before we can pass it to the JSON parser. It's /possible/ to let the JSON parser manage the message as it arrives and manage both the parse and storage of the content. Doing so would require shifting some responsibilities within the firmware that are difficult to change. We parse out the fields we care about and then want to hold onto these values for some time without retaining the memory necessary for the parse itself. Since all of the values and keys are present in the original source buffer passed to parse() as a string_view, we could simply save the string_views for the values and discard the JSON parser. Except, this parser copies the value to another buffer s.t. discarding the parser invalidates the value string_views. Alternatively, if the values themselves were returned as string_views into the source buffer, these string_views would (could) survive destruction of the parser. If we are missing the method by which this can be implemented with the current parser, we'd be delighted to hear it. Cheers -- /Marc Oscar Singer/ *Woollysoft* +1.206.328.1718

3 4

Abwesenheitsnachricht: JSON and no-alloc
by Sören Textor 14 Aug '23

14 Aug '23

Derzeit bin ich nicht im Hause Ihre E-Mail wird nicht weitergeleitet. In dringenden Fällen bitte support(a)ditec-gmbh.de oder Telefonhotline +49-6221-31698-110 oder die Zentrale +49-6221-31698-0 kontaktieren. Mit freundlichen Grüßen Dr. Sören Textor ----------------------------------------------- DiTEC Dr. Siegfried Kahlich & Dierk Langer GmbH Im Breitspiel 19 | 69126 Heidelberg | Germany Handelsregister | Mannheim | HRB 341229 Dieses Dokument ist vertraulich zu behandeln. Die Weitergabe, sowie Vervielfältigung, Verwertung und Mitteilung seines Inhalts ist nur mit unserer ausdrücklichen Genehmigung gestattet. Alle Rechte vorbehalten, insbesondere für den Fall der Schutzrechtsanmeldung. This document has to be treated confidentially. Its contents are not to be passed on, duplicated, exploited or disclosed without our express permission. All rights reserved, especially the right to apply for protective rights.

1 0

Full-Text Search is LIVE!
by Vinnie Falco 10 Jul '23

10 Jul '23

Greetings. I am proud to announce that Mohammad’s excellent work on implementing a full-text search which looks into all the Boost libraries and their documentation is now complete, and has been published to the live website. You can access the search by pressing the Search button in the top right of the page: https://boost.org Please note that this is only the first iteration of this search tool and as such it will probably not give the perfect results that everyone expects every time. But now that we have something in place we can feedback, brainstorm on how to make it better, and iterate on it so that it evolves into something that consistently enhances the Boost user experience. To report problems or suggestions related to the search including the user interface, and the indexed contents, please open an issue in the Boost website repository: https://github.com/boostorg/website/issues There is also a “Report Issue” button in the search UI. Thanks!

1 0

[Boost.TypeErasure] - Issue compiling under Windows with Embarcadero C++ Builder 11.3
by Andy Bell 07 Jul '23

07 Jul '23

Hi I'm using the type erasure library but have an issue (which I can workaround) when compiling in Windows 10 using Embarcadero C++ Builder 11.3. My problem is that C++ Builder brings in the Windows SDK Windows.h include file in every Windows VCL UI project it creates *before* any of my code has compiled and it's causing conflicts in boost header member11.hpp. This code is enough to cause the error: #include <boost/type_erasure/any.hpp> #include <boost/type_erasure/any_cast.hpp> #include <boost/type_erasure/builtin.hpp> #include <boost/type_erasure/operators.hpp> #include <boost/type_erasure/member.hpp> #include <boost/type_erasure/free.hpp> #include <boost/mpl/vector.hpp> namespace mynamespace { BOOST_TYPE_ERASURE_MEMBER(ReadStatus) } The error is because Windows.h defines 'interface' but this code from member11.hpp also uses 'interface': template<class P, template<class ...> class interface, class Sig, class Concept, class Base, class ID> using choose_member_interface = typename ::boost::mpl::if_c< ::boost::is_reference<P>::value, typename ::boost::mpl::if_c< ::boost::type_erasure::detail::is_non_const_ref<P>::value, interface<Sig, Concept, Base, const ID>, Base >::type, interface<Sig, Concept, Base, ID> >::type; and the compiler complains of conflicts. Embarcadero support have told me that I can't prevent Windows.h being included in my project before my boost-using code is compiled and attempts I've made to circumvent the problem, such as adding these lines to codegear.hpp: #ifdef interface #undef interface #endif don't work because C++ Builder is still including Windows.h before my code is compiled. To get around the problem, I changed member11.hpp in my boost installation to template<class P, template<class ...> class interface_, class Sig, class Concept, class Base, class ID> using choose_member_interface = typename ::boost::mpl::if_c< ::boost::is_reference<P>::value, typename ::boost::mpl::if_c< ::boost::type_erasure::detail::is_non_const_ref<P>::value, interface_<Sig, Concept, Base, const ID>, Base >::type, interface_<Sig, Concept, Base, ID> >::type; which I don't think will have side effects as 'interface/interface_' here is self-contained, but it doesn't feel like a good long term solution. Embarcadero support said: "Boost brings windows.h in because boost pulls it, *IF* the application is compiled for multi-threading then boost needs to access the OS threading functions, this is when you compile *without* the VCL." and I have confirmed this by creating a simple, multi-threaded Windows console application in C++ Builder... My question is: Is there a better way to work around this issue? My tests indicate that the Type Erasure library is working properly with my change, but it feels more like a hack than a fix... Thanks, Andy Bell Control Systems & Software Engineer Tel: +44 (0)1903 891700 Fax: +44 (0)1903 893888 Web: www.thermcosystems.com<http://www.thermcosystems.com/> www.csd-epi.com<http://www.csd-epi.com/> [cid:image002.png@01D9ADAA.693C8600] [Logo, company name Description automatically generated] P Please consider the environment before printing this email. This email and any files transmitted with it are confidential and intended solely for the intended recipient. If you are not the named addressee you should not disseminate, distribute, copy or alter this email. Any views or opinions presented in this email are solely those of the author and might not represent those of Tetreon Technologies Limited. Warning: Although Tetreon Technologies Limited has taken reasonable precautions to ensure no viruses are present in this email, the company cannot accept responsibility for any loss or damage arising from the use of this email or attachments. Tetreon Technologies Ltd is a limited company registered in England and Wales. Company Registration No. 4966845 Registered Offices: Highfield, Rock Road, Washington, West Sussex, RH20 3BH, England

2 1

[asio] monitoring io_context load
by Dmitry 07 Jul '23

07 Jul '23

Hello, I would like to have a metric that measures time spent for work (executing completion handlers) vs time spent waiting for work in thread pool executing io_context::run(). It will help to estimate service capacity (i.e. how many more requests per second it can handle). It is similar to the system's CPU usage metric but CPU usage metric does not include waiting for mutex locks or sync io performed from completion handlers. In our performance tests io_context becomes overloaded (the queue grows faster than it's processed) when CPU usage is about 85%. I would like to have a metric that shows 100% in this case. There is BOOST_ASIO_CUSTOM_HANDLER_TRACKING macro that allows to intercept handler invocations and to measure execution time. But in the case of many lightweight handlers measuring time can bring significant overhead. I'm considering adding similar macros before and after waiting for the condition variable and epoll to measure wait time and then calculate execution time. I'm writing to ask if there is a better alternative? If not, would you be interested in a patch that adds "Handler Tracking" macros around wait calls? Thank you, Dmitry

3 6

Please remove me from this list
by Vitaly Budovski 03 Jul '23

03 Jul '23

The web and email unsubscribe methods do not seem to work.

1 0

[python][shared_ptr_from_python.hpp] looks like copy/paste or merge error
by Ran Regev 29 Jun '23

29 Jun '23

Hi, The code in construct() seems to surfer from copy/paste or merge error. It seems as though the following lines are duplicated in the else clause: void *const storage = ((converter::rvalue_from_python_storage<SP<T> >*)data)->storage.bytes; // Deal with the "None" case. if (data->convertible == source) new (storage) SP<T>(); Please take a look at: https://www.boost.org/doc/libs/1_82_0/boost/python/converter/shared_ptr_fro… Ran.

1 0

[multiprecision] why 128 bit numbers isn't used in msvc?
by shar yashuv Giat 29 Jun '23

29 Jun '23

msvc added 128 bit numbers(in the header __msvc_int128.hpp). I know multiprecision uses __128 to improve performance so why doesn't it use msvc version? it's seems easy to change and will probably improve performance.

2 1

assign a large integer to a cpp_int
by Matthew Mawson 27 May '23

27 May '23

I want to assign a large number to a cpp_int variable like this: cpp_int large = 30414093201713378043612608166064768844377641568960512000000000000; and I get this: error: integer literal is too large to be represented in any integer type. Thanks!

2 1

Abwesenheitsnachricht: assign a large integer to a cpp_int
by Sören Textor 27 May '23

27 May '23

Derzeit bin ich nicht im Hause Ihre E-Mail wird nicht weitergeleitet. In dringenden Fällen bitte support(a)ditec-gmbh.de oder Telefonhotline +49-6221-31698-110 oder die Zentrale +49-6221-31698-0 kontaktieren. Mit freundlichen Grüßen Dr. Sören Textor ----------------------------------------------- DiTEC Dr. Siegfried Kahlich & Dierk Langer GmbH Im Breitspiel 19 | 69126 Heidelberg | Germany Handelsregister | Mannheim | HRB 341229 Dieses Dokument ist vertraulich zu behandeln. Die Weitergabe, sowie Vervielfältigung, Verwertung und Mitteilung seines Inhalts ist nur mit unserer ausdrücklichen Genehmigung gestattet. Alle Rechte vorbehalten, insbesondere für den Fall der Schutzrechtsanmeldung. This document has to be treated confidentially. Its contents are not to be passed on, duplicated, exploited or disclosed without our express permission. All rights reserved, especially the right to apply for protective rights.

1 0

[boost::locale::generator] Calling boost::locale::generator results in an exception in EnterCriticalSection
by M L 24 May '23

24 May '23

Hello, calling boost::locale::generator results in an exception in EnterCriticalSection. Does anybody have a hint to the cause of this problem? Every small clue is welcome. What could be a possible workaround? What is usually the reason for an exception in EnterCriticalSection? Code snippet: boost::locale::generator generator; generator.add_messages_path("..."); generator.add_messages_domain(".."); std::string lngCode = "en"; std::locale = generator(lngCode + ".WINDOWS-1252"); // this line crashes -> 0x0000005 EnterCriticalSection Boost Version: (boost_1_60_0) Visual Studio 2013 Some more hints: - it seems that there is a problem reading the translation file, which works fine in other applications. - there are no threads in the code - the problem is project dependent. Same code works in a different project. - the problem only appears if a specific header is included, but nothing in this header is ever executed before the crash. Removing the header/cpp removes the problem. - some C code is present in the code - the problem even appears when moving the snipped in main. Thanks for help.

2 1

Boost ASIO: Problems building a conditional variable using io_context / strand
by Suraaj K S 23 May '23

23 May '23

Hi all, In boost-asio, I realized that there was no easy way to have something that resembles a condition variable. However, I realized that I could get something very similar using stackful coroutines ( https://www.boost.org/doc/libs/1_66_0/doc/html/boost_asio/overview/core/spa…) while using an io_context as a 'queue' for completion tokens. Below is my approach: *#include "asio_types.hpp"#include <atomic>#include <iostream>class async_pending_queue {public: async_pending_queue() : pending_handlers_(0), strand_(pending_queue_), wg_(asio::make_work_guard(pending_queue_)) {} template <typename CompletionToken> auto async_submit( CompletionToken &&token, std::function<void(void)> atomic_action = [] {}) { auto init = [this, &atomic_action](auto completion_handler) { auto posted_lambda = [handler = std::move(completion_handler), this]() mutable { pending_handlers_--; asio_sys_err ec; handler(ec); }; post(strand_,std::move(posted_lambda)); pending_handlers_++; atomic_action(); }; return asio::async_initiate<CompletionToken, void(asio_sys_err)>(init, token); } int pending_count() { return pending_handlers_.load(); } // It may not run 1 and run 0 bool try_run_one() { auto cnt = pending_queue_.poll_one(); std::cout << "completion token result" << cnt << std::endl; bool ret = (cnt == 1); return ret; }private: std::atomic<unsigned int> pending_handlers_; asio_ioctx pending_queue_; asio_ioctx::strand strand_; decltype(asio::make_work_guard(pending_queue_)) wg_;};* Here, one simply uses calls `my_async_pending_queue.async_submit(yield)`, if calling from a stackful coroutine. The coroutine can be continued by calling `my_async_pending_queue.try_run_one()`. Using this, I wanted to build a 'memory checker'. It has two functions -> `request_space` and `free_space`. A coroutine calls `request_space`, which may block if there is no space left. Meanwhile, another thread / coroutine can call `free_space`, which will run blocked coroutines if possible. I built a toy memory checker wrapper as follows: *#ifndef MEM_CHECK_HPP#define MEM_CHECK_HPP#include <cstddef>#include <mutex>#include <queue>#include <boost/asio.hpp>#include <boost/asio/spawn.hpp>namespace asio = boost::asio;using asio_ioctx = asio::io_context;using asio_sys_err = boost::system::error_code;using asio::yield_context;#include "async_pending_queue.hpp"class MemoryChecker {public: using bytes_cnt = size_t; MemoryChecker(asio_ioctx &ioctx, bytes_cnt total_mem = 1024, bytes_cnt initial_fill = 0); // This requests some space, and possibly yields; void request_space(bytes_cnt cnt, yield_context yield); void free_space(bytes_cnt cnt);private: bytes_cnt get_available_mem(); const bytes_cnt total_mem_; bytes_cnt mem_used_; std::queue<bytes_cnt> request_queue_; async_pending_queue request_routines_; std::mutex lock_; asio_ioctx::strand fifo_strand_; asio_ioctx &completion_ioctx_;};inline MemoryChecker::MemoryChecker(asio_ioctx &ioctx, bytes_cnt total_mem, bytes_cnt initial_fill) : total_mem_(total_mem), mem_used_(initial_fill), request_queue_{}, lock_{}, request_routines_{}, completion_ioctx_(ioctx), fifo_strand_(ioctx) {}inline MemoryChecker::bytes_cnt MemoryChecker::get_available_mem() { assert(total_mem_ >= mem_used_); return total_mem_ - mem_used_;}inline void MemoryChecker::request_space(bytes_cnt cnt, yield_context yield) { // if (cnt > total_mem_) throw logic_error lock_.lock(); assert(cnt <= total_mem_); assert(cnt > 0); if (request_queue_.empty()) { assert(request_routines_.pending_count() == 0); if (get_available_mem() >= cnt) { // We bypass the pending queue mem_used_ += cnt; lock_.unlock(); return; } } assert(request_queue_.size() == request_routines_.pending_count()); std::cout << "Pushing " << cnt << std::endl; request_queue_.push(cnt); auto wg = asio::make_work_guard(completion_ioctx_); request_routines_.async_submit(yield, [this] { lock_.unlock(); }); auto oldest_req{request_queue_.front()}; assert(cnt == oldest_req); request_queue_.pop(); mem_used_ += cnt; assert(request_queue_.size() == request_routines_.pending_count()); asio::post(fifo_strand_, yield);}inline void MemoryChecker::free_space(bytes_cnt cnt) { { std::lock_guard<std::mutex> lg{lock_}; mem_used_ -= cnt; // Here, we own the lock, and free as many coroutines as we can while (true) { if (request_queue_.size() == 0) { std::cout << "No pending requests. Bailing" << std::endl; break; } assert(request_queue_.size() == request_routines_.pending_count()); auto oldest_req{request_queue_.front()}; auto available_mem{get_available_mem()}; if (available_mem < oldest_req) { std::cout << "Oldest request is larger than available_mem. Bailing" << std::endl; break; } assert(request_routines_.try_run_one() == true); } }}#endif /* MEM_CHECK_HPP */* Here is a test program that can run it: *#include "mem_check.hpp"#include <thread>#include <unistd.h>constexpr size_t mc_size{4};asio_ioctx ioctx;size_t total{0};MemoryChecker mc{ioctx, mc_size};void requestor_coroutine(size_t rq,yield_context yield) { asio::steady_timer t(ioctx); while (true) { total += rq; mc.request_space(rq, yield); std::cout << "Got requested space "; asio_sys_err ec; }}int main() { asio::spawn(ioctx, [](yield_context yield) { requestor_coroutine(1,yield); }); asio::spawn(ioctx, [](yield_context yield) { requestor_coroutine(2,yield); }); asio::spawn(ioctx, [](yield_context yield) { requestor_coroutine(3,yield); }); asio::spawn(ioctx, [](yield_context yield) { requestor_coroutine(4,yield); }); std::thread t([] { ioctx.run(); }); while (true) { getchar(); std::cout << total << std::endl; if (total > 0) { std::cout << "freeing" << std::endl; mc.free_space(1); total -= 1; } } t.join();}* Finally, the problem we face is as follows. When we run the program, the assertion `total_mem >= mem_used_` fails. On some further investigation, I realized that our completion token was being called even when we do not call `try_run_one`, which was very weird. Finally, somewhat more surprisingly, If I replace *post(stand_,std::move(posted_lambda));* by *post(pending_queue_,std::move(posted_lambda));*, things seem to work. However, the asio documentation says that only strands guarantee a FIFO execution order. I am not sure if using a simple `io_context` will work as a FIFO queue (even though it seems to in these examples). Any inputs would be helpful - I am happy to hear the problem in this implementation, as well as other implementations (for example, using a std::queue as a proper queue instead of this io_context hack). This question has also been posted here: https://stackoverflow.com/questions/76310252/c-boost-asio-building-a-condit… . Thanks, Suraaj

3 2

Re: [Boost-users] New Website !
by Vinnie Falco 11 May '23

11 May '23

On Thu, May 11, 2023 at 8:08 AM Damien via Boost-users <boost-users(a)lists.boost.org> wrote: > This looks really good. Congratulations to the team. Thank you very much for the kind words! Regards

1 0

Re: [Boost-users] New Website !
by Damien 11 May '23

11 May '23

This looks really good. Congratulations to the team. Damien

1 0