April 2018 - Boost-users - lists.preview.boost.org

[interprocess] Determine size of managed_shared_memory
by Kai Benndorf 12 Oct '19

12 Oct '19

Hi all, I want to create a managed_shared_memory in exactly the size I needed. Therefore I do something like this: size = sizeof( common_managed_shared_memory ); size += sizeof( MyClassInSharedMem ); managed_shared_memory segment(create_only, "name", size ); MyClassInSharedMem allocates itself some memory by segment->allocate(). I added this also to the size, but get a boost::interprocess::bad_alloc exception on allocating this additional memory. Are there some magic bytes I forgot to add? Many thanks in advance Kai

3 2

boost::filtered_graph filtering on vertex color... if I only had a colormap
by Brian Davis 19 Jul '19

19 Jul '19

Lets say I have a graph g (of type graph_t) and a color map (boy I wish I had a colormap): typedef boost::property_map<graph_t, boost::vertex_color_t>::type color_map_t; or is it? auto indexmap = boost::get(boost::vertex_index, g); auto colors = boost::make_vector_property_map<boost::default_color_type>(indexmap); or maybe? auto indexmap = boost::get(boost::vertex_index, g); auto colors = boost::make_vector_property_map<boost::vertex_color_t>(indexmap); or possibly some goop found on the internet? boost::make_iterator_property_map(color_map.begin(), boost::get(boost::vertex_index, data_->g), color_map[0]) or other internet goop? typedef boost::property_map<GraphType, boost::vertex_color_t>::type color_map_t; color_map_t colorMap; boost::color_map(colorMap)) or other internet goop?? std::vector<int> colorMap(num_vertices(g), boost::white_color); so confused... (certainly not surprised why bgl is not a part of stl) Can someone help explain the various types and ways of specifying colormaps and the logic for each way and means? and a function (if I could get that far without VS syntax highlighter going bonkers and gobs of compiler errors.. you know the type): make_pretty(g, colormap ) which colors the graph colormap and a filter struct for boost::filtered_graph<Graph, EdgePredicate, VertexPredicate> (where I wonder why if I only want to filter vertices I have to set boost::keep_all for EdgePredicate... sigh... I know the answer... and also know how long it found me to find/figure out boost::keep_all - and if you are looking for boost::keep_all and found it here... your welcome ;-) ) and modifying (probably incorrectly) boost::filtered_graph example to: template <typename ColorMap> struct vertex_color_filter_eq { vertex_color_filter_eq() { } vertex_color_filter_eq(ColorMap colormap/*, boost::default_color_type c*/) : m_colormap(colormap)/*, m_c(c)*/ { } template <typename Vertex> bool operator()(const Vertex& v) const { return 0 < get(m_colormap, v); //return c < get(m_colormap, v); } ColorMap m_colormap; //boost::default_color_type m_c; }; auto indexmap = boost::get(boost::vertex_index, g); auto colors = boost::make_vector_property_map<boost::default_color_type>(indexmap); // some where here I would call make_pretty(g, colors) typedef property_map<graph_t, boost::vertex_color_t>::type ColorMap_t; vertex_color_filter<ColorMap_t> filter(colors); // boost::filtered_graph<graph_t, boost::keep_all, VertexPredicate> boost::filtered_graph<graph_t, boost::keep_all, vertex_color_filter> fg(g, filter); // Later say if I could get it compile and better if it compiled and worked I could move on and auto color = boost::red_color; typedef property_map<graph_t, boost::vertex_color_t>::type ColorMap_t; vertex_color_filter<ColorMap_t> filter(colors, color); boost::filtered_graph<graph_t, boost::keep_all, vertex_color_filter> fg(g, filter); Maybe if I color them yellow, I can then reach the wizard pull back the curtain and ask how to create a color map... If I only had a color map.

4 6

[Test] Test executable crashes always
by Nate Widmyer 26 Jun '18

26 Jun '18

Hello, I've made some unit tests with boost test before without issue, but on my latest rather ordinary test, I've been able to get Boost test to crash the executable in debug and release mode when boost test is making a report before terminating. test_exe.exe!boost::unit_test::framework::get(unsigned long id, boost::unit_test::test_unit_type t) Line 1388 C++ test_exe.exe!boost::unit_test::framework::get<boost::unit_test::test_unit>(unsigned long id) Line 208 C++ test_exe.exe!boost::unit_test::results_reporter::make_report(boost::unit_test::report_level l, unsigned long id) Line 175 C++ test_exe.exe!boost::unit_test::unit_test_main(boost::unit_test::test_suite *(*)(int, char * *) init_func, int argc, char * * argv) Line 235 C++ test_exe.exe!main(int argc, char * * argv) Line 287 C++ That's the default report_level of 'confirm' If I use report_level=short or detailed, I get a similar crash: test_exe.exe!boost::unit_test::framework::get(unsigned long id, boost::unit_test::test_unit_type t) Line 1388 C++ test_exe.exe!boost::unit_test::framework::get<boost::unit_test::test_suite>(unsigned long id) Line 207 C++ test_exe.exe!boost::unit_test::traverse_test_tree(unsigned long id, boost::unit_test::test_tree_visitor & V, bool ignore_status) Line 365 C++ test_exe.exe!boost::unit_test::results_reporter::make_report(boost::unit_test::report_level l, unsigned long id) Line 185 C++ test_exe.exe!boost::unit_test::unit_test_main(boost::unit_test::test_suite *(*)(int, char * *) init_func, int argc, char * * argv) Line 235 C++ test_exe.exe!main(int argc, char * * argv) Line 287 C++ Here's the get() method: test_unit& get( test_unit_id id, test_unit_type t ) { test_unit* res = impl::s_frk_state().m_test_units[id]; BOOST_TEST_I_ASSRT( (res->p_type & t) != 0, internal_error( "Invalid test unit type" ) ); return *res; } Crash occurs on line 2 of the function. I logged 'res' each time the function is called and many times its valid, but when it runs this one test (by itself or with others), 'res' always ends up being NULL. id seems to always be '1', and t is TUT_ANY (17) if generating a confirmation report or TUT_SUITE (16) if generating a short or detailed report. If I use the option '--report_level=no', then execution does complete fine. I followed through on some process of elimination and it seems to be related to using a fixture? Having no test body still causes the crash. It wasn't until I removed the fixture (and its usage in the body of the test) did the executable not crash. I ensured that I was using the standard fixture model found here: http://www.boost.org/doc/libs/1_64_0/libs/test/doc/html/boost_test/tests_or… OS is Win 7 w/ VS2015 boost is 1.64.0 and was built with: b2 toolset=msvc-14.0 link=static runtime-link=static variant=debug address-model=64 --without-atomic --without-container --without-context --without-coroutine --without-coroutine2 --without-graph --without-graph_parallel --without-iostreams --without-locale --without-math --without-mpi --without-serialization --without-signals --without-type_erasure --without-wave -j4 --abbreviate-paths --without-python -d2 Any help would of something to look at or change would be appreciated. Thanks for your time, Nathan

2 4

help
by Radekkostra＠seznam.cz 30 Apr '18

30 Apr '18

hello, i am trying to build boost, but i have problems. First is in log(bootstrap.log - could not find "vswhere") and second is when compiling boost(this command :.\b2 --toolset=msvc link=shared,static address -model=64 --with-system --with-filesystem -j 4(5 failed,12skipped,79 updated).

2 2

[Process] Help needed with boost::process::extend::on_*
by Günter Berner 26 Apr '18

26 Apr '18

I am using boost::process::child to create a child process: boost::process::child child("foo"); This works fine so far (on Linux + Windows). Now i want to use a handler which is executed during the fork process: boost::process::child child("foo", boost::process::extend::on_setup = [](auto& exec) { // ... }); This also works fine - as long as a more or less up-to-date C++ compiler or C++ standard. Since g++-4.8.5 (which I have to use) does not support auto template deduction for lambdas, i have to replace the lambda parameter type with a concrete type. boost::process::child child("foo", boost::process::extend::on_setup = [](<WHAT_GOES_HERE?>& exec) { // ... }); Boost.Process documentation only uses the modern approach with auto template deduction. I simply cannot find an example/documentation where the concrete types are shown. Thanks.

1 0

failed coversion
by Markel, Joseph 25 Apr '18

25 Apr '18

Is there a way to get the int value by performing division on boost::any variables? #include <boost/any.hpp> int main(void) { boost::any a = 10; std::cout << boost::any_cast<int>(a) << std::endl; boost::any s = 2.5; // scaling factor // scale the value a = boost::any_cast<int>(a)/boost::any_cast<float>(s); std::cout << boost::any_cast<int>(a) << std::endl; } This throws the error boost::bad_any_cast: failed conversion using boost::any_cast Thanks CONFIDENTIALITY NOTICE: This e-mail communication and any attachments may contain proprietary and privileged information for the use of the designated recipients named above. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message.

2 1

[multiprecision] static_cast<cpp_bin_float_100> gives strange values.
by Vögl, Christian 20 Apr '18

20 Apr '18

Hi everyone. The following Situation: I tried to improve the accuracy of a numerical function by replacing all double variables with cpp_bin_float_100. The parameters of this function contain two 3-element double arrays, which I convert to cpp_bin_float_100 by the following function: static void equal(cpp_bin_float_100* a, double* b) { for(int i=0; i<3;i++) { a[i]=static_cast<cpp_bin_float_100>(b[i]); } } the third element seems to be converted correctly, the first and second are totally off. (e. g. 100 -> 1.5625)

2 1

[asio] Using struct ip_mreqn to joing multicast group
by Klebsch, Mario 20 Apr '18

20 Apr '18

Hi, I need to receive multicasts on multiple interfaces and handle them differently, depending on the network interface they arrived on. For each network interface, I create an UDP socket and I use Socket.set_option(boost::asio::ip::multicast::join_group(MCastReqAddr.to_v4(), LocalAddr.to_v4() ), ec); to join the multicast group. LocalAddr is the local IPv4 address of the interface. Unfortunately, I receive each incoming multicast message on each socket. I found, that boost::asio::ip::multicast::join_group uses struct ip_mreq, where the multicast address and the local address can be specified. On linux the man page on IP_ADD_MEMBERSHIP tells to use struct ip_mreqn. This structure had an additional member imr_ifindex. The documentation is not very clear about the difference between imr_address and imr_ifindex. So, I want to try it out and wonder, whether boost::asio supports using struct ip_mreqn for IP_ADD_MEMBERSHIP. If not, does anybody know, how to write my own class to be used with Socket.set_option? 73, Mario -- Mario Klebsch Actia I+ME GmbH Mario.klebsch(a)ime-actia.de<mailto:Mario.klebsch@ime-actia.de> Dresdenstrasse 17/18 Fon: +49 531 38 701 716 38124 Braunschweig Fax: +49 531 38 701 88 Germany

1 0

Boost.Range and initializer_list
by Robert Jones 18 Apr '18

18 Apr '18

Hi All Does the Boost.Range library support initializer_list as a range type? I thought anything with begin() and end() methods was fair game, but this code doesn't work, unless I'm doing it wrong. #include <iostream> #include <vector> #include <boost/range/adaptor/reversed.hpp> int main( ) { using namespace std; using namespace boost::adaptors; vector<int> v { 1, 2, 3 }; for( int i : v | reversed ) { cout << i << " "; } // Ok for( int i : { 1, 2, 3 } | reversed ) { cout << i << " "; } // Bad } Kind Regards Rob.

3 3

[Boost.Range] combine, native range-for, and destructuring
by John M. Dlugosz 18 Apr '18

18 Apr '18

I want to go through two vectors at the same time, and the way to do this is to "zip" the items from each into a tuple. boost::combine is a function for that. for (auto&[def,residue] : boost::combine(playdef, residues)) { gives me an error 1>e:\scratch\c++code\scratchproject2\fizzbuzz\fizzbuzz.cpp(66): error C2440: 'initializing': cannot convert from 'boost::tuples::cons<std::pair<int,std::string_view> &,boost::tuples::cons<int &,boost::fusion::detail::build_tuple_cons<boost::fusion::single_view_iterator<Sequence,boost::mpl::int_<1>>,Last,true>::type>>' to 'boost::tuples::cons<std::pair<int,std::string_view> &,boost::tuples::cons<int &,boost::fusion::detail::build_tuple_cons<boost::fusion::single_view_iterator<Sequence,boost::mpl::int_<1>>,Last,true>::type>> &' The only difference is in the final `&` character. Trying to explore the behavior in pieces, I try: for (auto row : boost::combine(playdef, residues)) { auto&[def, residue] = row; // line 1 row.get<1>() = 17; // line 2 residue= 42; // line 3 Now line2 does work, and I note that assignment goes through and updates the original `residues` vector. The destructuring on line 1 does compile now, being on a separate line. But line 3 gives me an error about not being able to assign to a crazy complicated type. What's going on? Is Boost.Range zipped ranges simply not compatible with the native language features?

4 5

Turning off BOOST_IOSTREAMS_HAS_DINKUMWARE_FPOS for MSVC++
by Billy O'Neal (VC LIBS) 17 Apr '18

17 Apr '18

Hi folks! As part of a recent cleanup of some of our iostreams bits I removed a double-seek inside std::filebuf. (Previously, std::filebuf didn't take advantage that we know ::fsetpos just accepts a 64 bit number position) Moreover, in previous releases of MSVC++, that still would have been correct; filebuf was getting to the same place in a very indirect way. I know just casting std::fpos to a streamoff works to get the position going back to at least VS 2015, and it probably works a decade earlier than that (though I have not tested). Shame on us for adding a pretty name to std::fpos not depicted in the standard. Would you accept a patch to $\boost\iostreams\detail\config\fpos.hpp adding this guard before attempting to define BOOST_IOSTREAMS_HAS_DINKUMWARE_FPOS? !(_MSVC_STL_VERSION >= 141) I submitted a patch to do that here: https://github.com/boostorg/iostreams/pull/57 Thanks folks! Billy O'Neal Visual C++ Libraries

3 2

Boost.Range and result_type
by Robert Jones 17 Apr '18

17 Apr '18

Hi All I have a suspicion this is all ancient history now, but I was surprised this code worked. #include <iostream> #include <vector> #include <boost/range/adaptor/transformed.hpp> struct TimesTwo { // using result_type = int; int operator( )( int i ) const { return i * 2; } }; int main( ) { using namespace std; using namespace boost::adaptors; vector<int> v { 0, 1, 2, 3, 4, 5 }; for( int i : v | transformed( TimesTwo( ) ) ) { cout << i << " "; } cout << endl; } Due to result_type being commented out. Am I right in thinking this is what BOOST_RESULT_OF_USE_DECLTYPE was all about, and everything now just works because modern compilers are implemtenting decltype now? Kind Regards Rob.

2 1

[DateTime] localtime_r() issue
by Latimerius 16 Apr '18

16 Apr '18

Hi, I'm trying to build a project that uses Boost.Log (which in turn uses Date_Time) from Boost 1.65.1 on a platform that uses Clang with (apparently) Dinkumware for a standard library implementation. Compilation fails in date_time/c_time.hpp on line 67 due to a use of undeclared identifier 'localtime_r'. The standard library implementation indeed doesn't seem to have localtime_r(), just localtime_s(). Compilation of the failing code in c_time.hpp is enabled by the preprocessor symbol BOOST_DATE_TIME_HAS_REENTRANT_STD_FUNCTIONS having been #define'd which happens in date_time/compiler_config.hpp based on BOOST_HAS_THREADS being #define'd and a bunch of compiler checks that probably don't get it quite right. Could someone give me some advice how to deal with this? Thanks in advance!

1 0

[release] Boost 1.67.0
by Daniel James 15 Apr '18

15 Apr '18

Release 1.67.0 of the Boost C++ Libraries is now available. These open-source libraries work well with the C++ Standard Library, and are usable across a broad spectrum of applications. The Boost license encourages both commercial and non-commercial use. This release contains 2 new libraries and numerous enhancements and bug fixes for existing libraries. For details, including download links, see <https://www.boost.org/users/news/version_1.67.0> You can also download directly from: <https://dl.bintray.com/boostorg/release/1.67.0/> To install this release on your system, see <https://www.boost.org/doc/libs/release/more/getting_started/index.html> Thanks to everyone who participated in this release. Thanks, -- The Boost Release Team

3 3

Convert boost::shared_ptr to std::shared_ptr for boost::dll::import
by David Demelier 13 Apr '18

13 Apr '18

Hello, In my application I use std::shared_ptr in the public API for loading plugins. I wanted to use boost::dll to import dynamically plugins as well. It looks like for some reasons that it still use boost::shared_ptr instead of std::shared_ptr even though C++11 is 7 years old. I wanted to convert the boost shared_ptr to std with the following function: std::shared_ptr<plugin> load(boost::shared_ptr<plugin> ptr) { return std::shared_ptr<plugin>(ptr.get(), [ptr] (auto) {}); } It works at perfectly at runtime, however my application segfaults at end, when destroying objects. I suspect that it's because the internal boost::dll shared_library is destroyed **before** my own shared_ptr. The backtrace: (gdb) bt #0 0x00007ffff3f054d8 in ?? () #1 0x00007fffffffdb80 in ?? () #2 0x00007ffff4109400 in ?? () #3 0x00007ffff4add6f8 in _nl_current_default_domain () from /lib64/libc.so.6 #4 0x00007ffff47641d8 in __run_exit_handlers () from /lib64/libc.so.6 #5 0x00007ffff476422a in exit () from /lib64/libc.so.6 #6 0x00007ffff474df31 in __libc_start_main () from /lib64/libc.so.6 #7 0x0000000000413c1a in _start () Is there any chance I can safely convert the smart pointer? Regards, -- David

8 12

Credits to Boost.Asio
by David Demelier 13 Apr '18

13 Apr '18

Hello, I'm the author of irccd [0], a C++ IRC bot extensible with plugins. Up to the 3.0.0 version current in development, I had my network code written by myself (being used of lots of poll(2) usage in C). I have switched to Boost.Asio a while ago to remove my own code and facilitate further development. Even though I think it has its learning curve, I enjoyed a lot the switch and tested my IRC bot to connect 100 times in the same instance to my own local IRC server. It took less than 2 seconds to join all requested channels, taking only 2MB of RAM. This is just awesome. Congratulations for this powerful module :) [0]: http://projects.malikania.fr/irccd -- David

1 0

boost::rational does not maintain its invariant in case of overflow
by Frédéric 11 Apr '18

11 Apr '18

With the program shown below, I obtained the following output: std::numeric_limits<long>::max() == 9223372036854775807 r == boost::rational<long>(4294967296, 1853020188851841) == 4294967296/1853020188851841 r * r == 0/8733086111712066817 Clearly there are 2 overflows but this leads to the invariant of the class not being maintained, i.e. the faction is not simplified to 0/1 (gcd(num, den) != 1). F #include <boost/rational.hpp> #include <iostream> int main() { std::cout << "std::numeric_limits<long>::max() == " << std::numeric_limits<long>::max() << '\n'; boost::rational<long> r(4294967296, 1853020188851841); std::cout << "r == boost::rational<long>(4294967296, 1853020188851841) == " << r << '\n'; std::cout << "r * r == " << (r * r) << '\n'; return 0; }

2 2

How do I "run the configure tests"?
by Chris Stankevitz 10 Apr '18

10 Apr '18

Can't find anything about it in the docs. "Unknown compiler version - please run the configure tests and report the results" [1] Thank you, Chris [1] From https://www.boost.org/doc/libs/1_65_1/boost/config/compiler/visualc.hpp

4 5

Toolset version for Visual Studio 2017 15.6.3
by Chris Stankevitz 10 Apr '18

10 Apr '18

What version should I pass to b2 --toolset=msvc-x.y for "Visual Studio 2017 15.6.3"? [1] says [2] will have an up-to-date list with an answer, but [2] is not up-to-date. Thank you, Chris [1] https://www.boost.org/doc/libs/1_65_1/more/getting_started/windows.html#ide… [2] https://www.boost.org/build/doc/html/bbv2/reference/tools.html

4 4

[asio] SSL async_shutdown() never finishes if called when async_read_some() is in progress
by Alexander D. 08 Apr '18

08 Apr '18

Hello, I'm writing a middleware which basically adds an SSL layer to the connection between a client and a server. The connection is bidirectional, so I need to read from each side and write to the other. The client can also close the connection at any time, in which case async_shutdown() must be called on the server's stream. The problem is, often the shutdown happens when the program is waiting on the server through async_read_some(), and the shutdown will never finish. The following code illustrates this issue: #include <iostream> #include <boost/asio.hpp> #include <boost/asio/ssl.hpp> int main(int argc, char* argv[]) { try { boost::asio::io_context io_context; boost::asio::streambuf buffer; boost::asio::ssl::context ctx(boost::asio::ssl::context::tlsv12); ctx.set_verify_mode(boost::asio::ssl::verify_none); boost::asio::ssl::stream<boost::asio::ip::tcp::socket> sock(io_context, ctx); sock.next_layer().connect(boost::asio::ip::tcp::endpoint( boost::asio::ip::make_address("127.0.0.1"), 4433)); sock.handshake(boost::asio::ssl::stream_base::client); /* "async_shutdown" won't appear unless async_read_some() * is commented out */ sock.async_read_some(buffer.prepare(1024), [](boost::system::error_code, size_t) {std::cout << "async_read_some" << std::endl;}); sock.async_shutdown([](boost::system::error_code) {std::cout << "async_shutdown" << std::endl;}); io_context.run(); } catch (std::exception& e) { std::cerr << "Exception: " << e.what() << "\n"; } return 0; } To reproduce, start up an SSL server: openssl genrsa -out key.pem 2048 openssl req -new -sha256 -key key.pem -out csr.csr openssl req -x509 -sha256 -days 365 -key key.pem -in csr.csr -out server.pem openssl s_server -key key.pem -cert server.pem and run the program. Notice that "async_shutdown" does not finish even if the server is killed. In practice, this leaks the memory of the shutdown handler and whatever is bound to it. The only solution I can think of would be to cancel the read operation, but there's no portable way in asio to do this. Can someone point out how to solve this? Cheers, Alexander

5 11

[release] Boost 1.67.0 Beta 1
by Daniel James 07 Apr '18

07 Apr '18

Boost release 1.67.0 beta 1 is now available at: <https://dl.bintray.com/boostorg/beta/1.67.0.beta.1/source/> The SHA256 checksums are as follows: 154cf490da858fd9c2307bb904b07a42cbe8355ecf99d5477f0d843c6d03cdda boost_1_67_0_b1.7z 0d854b208fefaeb5837d6c417f505b189275e32d017f6226001fb165a5ac51ed boost_1_67_0_b1.tar.bz2 6eb3f9ca8db355765f41361e1e26976228f0c2dff400aade8a5e97aec6aa59f4 boost_1_67_0_b1.tar.gz 678a6459f2517932536ec97297be1de46ea073f2e2ee17ec8397b80aa0953f49 boost_1_67_0_b1.zip For details of what's in the release, see <http://www.boost.org/users/history/version_1_67_0.html>. Please download the beta, give it a try, and report any problems you encounter. Thanks, -- The Boost Release Team

9 14

Pointer Container Library changes in 1.66.0
by Kevin Hardman 06 Apr '18

06 Apr '18

I am seeing changes in the Pointer Container Library that are preventing my existing application from compile. I see no notice of changes in the 1.66.0 release notes <http://www.boost.org/users/history/version_1_66_0.html <http://www.boost.org/users/history/version_1_66_0.html>>. Is there documentation on the changes and what is required to migrate existing users to the new library? —kevin

5 10

[log] Unable to use own severity struct encapsulated in namespace
by Sebastian Messerschmidt 06 Apr '18

06 Apr '18

Hi folks, When taking this example [1] and modifying the enum 'severity_level' like this: namespace x { struct test { enum severity_level { normal, notification, warning, error, critical }; }; } and replacing all references to severity_level with the appropriate x::test::severity_level the 'operator<<' overload for the severity_level is not issued. Can anyone help me with this one? I can supply the complete modified example if needed Cheers Sebastian [1] https://www.boost.org/doc/libs/1_58_0/libs/log/example/doc/tutorial_attribu…

2 4

[ptr_map] Warning about returning address of local variable
by Bill Buklis 05 Apr '18

05 Apr '18

I'm not sure if this a potential bug/problem or just a benign warning, but when using the map_values range adaptor with a ptr_map container I get a warning about returning an address of a local variable. It seems benign as the expected processing happens, but this type of warning generally doesn't sound like a good thing. The following useless code will trigger it. #include <boost/ptr_container/ptr_map.hpp> #include <boost/range/algorithm.hpp> #include <boost/range/adaptors.hpp> struct sample { int value; }; typedef boost::ptr_map<int, sample> themap; int main() { themap data; int key = 5; data.insert(key, new sample); boost::for_each(data | boost::adaptors::map_values, [](sample* s) { s->value = 10; }); return 0; } I'm using Visual Studio 2017. I also have conformance mode (/permissive-) turned on for stricter language compliance in case it makes a difference. As an aside, I was a little surprised when trying out this little sample that the inclusion of the above headers is enough to require linking with the boost regex library. I figure it's probably a reference in the adaptors somewhere. It's not a big deal. These lib files are normally available anyway. -- Bill

2 1

boost::asio::async_read_until fills streambuf less than actually received
by David Demelier 03 Apr '18

03 Apr '18

Hello, I'm using Asio to read IRC messages asynchronously from a server. I use boost::asio::async_read_until with "\r\n" as delimiter and a boost::asio::streambuf as output buffer. I mostly implemented the code like described on my original question/answer here: https://stackoverflow.com/questions/40561097/read-until-a-string-delimi ter-in-boostasiostreambuf I sometimes get the bytes_transferred argument bigger that what is filled into the buffer. My code looks like this: void connection::do_recv(handler_t handler) { boost::asio::async_read_until(socket_, buffer_, "\r\n", [this, handler] (auto code, auto xfer) { if (code || xfer == 0U) handler(std::move(code), message()); else { //assert(buffer_.size() > xfer); std::cout << "xfer: " << xfer << std::endl; std::cout << "size: " << buffer_.size() << std::endl; std::string str( boost::asio::buffers_begin(buffer_.data()), boost::asio::buffers_begin(buffer_.data()) + xfer - 2 ); buffer_.consume(xfer); handler(std::move(code), message::parse(str)); } }); } void connection::rflush() { if (input_.empty()) return; do_recv([this] (auto code, auto message) { auto h = input_.front(); // Delete first, in case of exceptions. input_.pop_front(); if (h) h(code, std::move(message)); if (!code) rflush(); }); } void connection::recv(handler_t handler) { auto in_progress = !input_.empty(); input_.push_back(std::move(handler)); if (!in_progress) rflush(); } For some reasons, after several calls, I get an assertion failure in the std::string construction. Because buffer_.size() is actually smaller than xfer. I could not understand why. Example of output: ... ... xfer: 87 size: 141 Note: I've implemented the recv() public function like what I can see for send in some tutorials. I wanted a queue of handlers so I can safely call recv() multiple times before first functions finished completely. Regards, -- David

2 1