- Boost-users - lists.boost.org

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

Crossbuilding for ARM
by Mats Webjörn/WIKAB 30 Mar '18

30 Mar '18

I'm trying to build for ARM, and have some problems in making b2 find user-config.jam when it's located outside boost root dir. First I go to boost/tools/build and run ./b2 install -prefix=/home/administrator/project/myproj/build/boost Then I run ./b2 toolset=gcc-arm --build-dir=/home/administrator/project/myproj/build/boost stage According to http://www.boost.org/build/doc/html/bbv2/overview/configuration.html Boost.Build<http://www.boost.org/build/doc/html/bbv2/overview/configuration.html%20Boos…> will search $HOME and $BOOST_BUILD_PATH, but if I add --debug-configuration to b2 command then I can see that it searches '/home/administrator' = $HOME '/home/administrator/project/myproj/externals/boost/tools/build/src/kernel' '/usr/share/boost-build' '/home/administrator/project/myproj/externals/boost/tools/build/src/kernel' '/home/administrator/project/myproj/externals/boost/tools/build/src/util' '/home/administrator/project/myproj/externals/boost/tools/build/src/build' '/home/administrator/project/myproj/externals/boost/tools/build/src/tools' '/home/administrator/project/myproj/externals/boost/tools/build/src/contrib' '/home/administrator/project/myproj/externals/boost/tools/build/src/.'. Where none of the directories searched is the one I've provided as buid directory. It seems odd that $BOOST_BUILD_PATH should be "the one in the top-level directory of your Boost.Build installation" since http://www.boost.org/build/doc/html/bbv2/overview/invocation.html describes --build-dir as "the build directories for all project roots being built". It seems like the documentation isn't correct. I'm just not sure which one is right and which is wrong. Has anyone crossbuilt for ARM on Linux? Regards Mats Webjörn E-Mail: konsult(a)webjorn.com Webjörn IngenjörsKonsult AB Tel +46 8 584 238 80 Solhemsbackarna 75A Fax +46 8 445 69 21 SE-163 56 SPÅNGA Mobil +46 70 639 16 99 Sweden

1 0

[Boost::Test] short form of catch_system_errors not working
by Konstantin Kivi 30 Mar '18

30 Mar '18

Occasionally I need a core dump if testing goes wrong. Boost test have the following option ------------ Parameter: catch_system_errors Allows to switch between catching and ignoring system errors (signals). Command line formats: --[no_]catch_system_errors[=<boolean value>] -s[ <boolean value>] ------------------- I never managed to get short options to work and has to search for a long option exact name. Neither '-s no', '-s false' , -sno with or without quotes are not accepted. examples ----- $ test_app -s no Boost.Test WARNING: token "no" does not correspond to the Boost.Test argument and should be placed after all Boost.Test arguments and the -- separator. For example: test_app --random -- no ------ $ test_app "-s no" Invalid separator for the parameter catch_system_errors in the argument -s no ----- Is it a known bug? Best regards, Konstantin

2 1

[boost][interprocess] Need clarification on file_lock documentation
by Rob Conde 29 Mar '18

29 Mar '18

The documentation of boost::interprocess::file_lock has some cautions on synchronization limitations: * It's unspecified if a file_lock synchronizes two threads from the same process. * It's unspecified if a process can use two file_lock objects pointing to the same file. But then gives the guidance: * For each file, use a single file_lock object per process. * Use the same thread to lock and unlock a file. I then created the following example: <https://wandbox.org/permlink/dO2rRDCHVWBE3pGr>https://wandbox.org/permlink/3CNfkA3RcMkwbd2v #include <fstream> #include <iostream> #include <cstdlib> #include <boost/interprocess/sync/file_lock.hpp> #include <boost/interprocess/sync/scoped_lock.hpp> #include <boost/filesystem.hpp> #include <mutex> #include <thread> #include <chrono> int main() { std::ofstream("f.txt"); boost::interprocess::file_lock lock("f.txt"); std::mutex mtx; const size_t max = 100; size_t t1Count = 0; std::thread t1 ( [&lock,&t1Count,&mtx,max]() { while(t1Count < max) { //std::lock_guard<std::mutex> mtxLock(mtx); boost::interprocess::scoped_lock<boost::interprocess::file_lock> guard(lock); std::cout << "Number 1" << std::endl; t1Count++; std::this_thread::sleep_for(std::chrono::milliseconds(1)); } } ); size_t t2Count = 0; std::thread t2 ( [&lock,&t2Count,&mtx,max]() { while(t2Count < max) { //std::lock_guard<std::mutex> mtxLock(mtx); boost::interprocess::scoped_lock<boost::interprocess::file_lock> guard(lock); std::cout << "Number 2" << std::endl; t2Count++; std::this_thread::sleep_for(std::chrono::milliseconds(1)); } } ); t1.join(); t2.join(); return 0; } It only works properly if the std::mutex is used in addition to the file_lock. That was originally my thought when I read "It's unspecified if a file_lock synchronizes two threads from the same process.", but it seemed like I was following the guidance "Use the same thread to lock and unlock a file.". If I haven't messed anything up, I would probably update the documentation to say something like: "If utilitizing the file_lock from multiple threads within the same process, synchronize access to the file_lock with an intraprocess synchronization mechanism such as a mutex." Otherwise, is it a bug? Thanks, Rob

1 0

[beast] Custom, application specific header
by DePizzottri 28 Mar '18

28 Mar '18

Hi, I need to send custom application specific header in the response, but http::response operates only with set of predefined http::field values. Does anyone know how can I set custom field without implementing own `fields` class? Or is there any tips or example how can it be implemented (I know about Fields/FieldsWriter concepts)? -- Sent from: http://boost.2283326.n4.nabble.com/Boost-Users-f2553780.html

2 3

boost::dll in cygwin
by Matthew von Arx 27 Mar '18

27 Mar '18

Has anyone tried using the boost::dll library in cygwin? Boost attempts to load <link.h> which is not applicable to a Cygwin system. Is there any way to #define an option so that the boost::dll library will operate in this mixed windows/linux-like environment? Thanks, -- Matthew von Arx Electromechanical Engineer (C) 678-953-1752

2 1

Boost managed memory files
by Assaf Ben-amitai 26 Mar '18

26 Mar '18

Hi, I'm a user of boost interprocess managed memory file. I'm facing the follwing issue - it seems that if a process fails while allocating memory from a managed file, the internal data-structures may corrupt , making the managed file becomes unuseable. Does anyone have any info on efforts / projects to improve on this situation? Thx, Assaf

1 0

boost::asio blocking socket read with timeout with multiple threads
by tom＠quarendon.net 24 Mar '18

24 Mar '18

I've got a boost::asio based server, and generally it's working well. The server runs multiple threads, each simply running the io_service::run method on the same io_service object. It accepts multiple connections and processes multiple concurrent sockets. It uses a strand per connection to ensure that the data for each connection is processed strictly sequentially. For various reasons it uses a mixture of async and sync reads. The trouble I'm having is implementing a timeout on the synchronous (blocking) reads. I want to protect the server in the case of a client sending an erroneous or malicious packet for example, and make sure a read doesn't block forever. There was a post back in 2007 that gave a technique (https://lists.boost.org/Archives/boost/2007/04/120339.php) There's a stack overflow question that references that (https://stackoverflow.com/questions/13126776/asioread-with-timeout) There are some examples in the docs that perform a similar kind of technique. In the examples, the server is all asynchronous. The client is synchronous, but only uses one thread. The example from 2007 also only uses one thread. The examples all revolve around the technique of starting an *async* read, then performing a nested loop of io_service::run_one. However, so far I've been unable to find a form of code that works reliably in a multithreaded environment. The basis for my experimentation is here: https://gist.github.com/anonymous/1160c11f8ed9c29b9184325191a3a63b It starts a server thread, then starts a client that makes a connection and then writes nothing, to simulate a "bad" client and to provoke a "timeout" condition. What I'm finding is this: With one server thread I can make it work OK. The nested loop of "run_one()" operates correctly. With multiple server threads though, I can't make it work reliably. The basis of the loop is: bool timedOut = false; bool readComplete = false; timer.async_wait(strand_.wrap(boost::bind(&handleReadTimeout, &timedOut, &socket_, _1))); async_read(socket_, buffer(b), strand_.wrap(boost::bind(handleReadComplete, &ec, &bytesRead, &readComplete, &timer, _1, _2))); while (socket_.get_io_service().run_one()) { if (timedOut && readComplete) { break; } } handleReadTimeout sets the timedOut flag to true and calls "cancel" on the socket. The handleReadComplete callback sets readComplete to true and calls cancel on the timer. With multiple threads, the handleReadTimeout/handleReadComplete callbacks are run on other threads. So the while loop here just blocks, as there is never anything to run. That's my surmise of what's going on anyway. I've experimented with strands to try and force it all onto the same thread, but so far failed (if the above code is called in the context of the same strand, it just seems to block the handleReadTimeout and handleReadComplete callbacks from being called). Note that if I expend my example to include other clients talking to a simple "echo" service to simulate other traffic, then the while loop above *does* wake, as that other traffic gives it something to do. But that's fragile. The alternative formulation, the one I started with actually, is to do an async wait on the timer, and a normal synchronous read on the socket. The timer callback performs a cancel (or close -- I tried both) on the socket hoping to cause the socket read to error. This is the kind of thing you'd do if you were programming raw sockets. That works fine on Windows, but won't work on Linux. The documentation for cancel does say it cancels any outstanding *async* calls, but I'm surprised calling close doesn't work and cause the read to wake. So I'm stuck. I was hoping someone here could point me in the right direction, as it seems an obvious thing (to me) to want try and do. Is there a form of code that allows me to perform a synchronous read from a socket and time it out, which will work in a multi threaded/single io service environment? This is all with boost 1.64 by the way. Thanks for any help in advance.

6 27

Re: [Boost-users] boost::asio blocking socket read with timeout with multiple threads
by tom＠quarendon.net 23 Mar '18

23 Mar '18

> You're right. That poll_read loop consumes the alarm :/ > > So then i think the fallback has to be a raw Berkley sockets read against > the native_handle(), with a signal to cause the EINTR. Yes, and I've tried basically that method earlier. It works fine, as long as you don't want to use the SSL wrapper, which is what I ultimately want to do. Of course there, you don't have the control to do that. Well, I could write my own socket class variant and wrap that in the SSL wrapper, but that's daft. You can't "test" the socket by using a poll with a timeout, because there may be unprocessed data that the SSL code has read off the socket that it hasn't returned to you yet, so you'd block when actually there was data to return (been there). There is, as far as I can see, no sensible solution to this. The best I've seen involves mutexes and condition variables, but that's just horrible for what should just consist of passing a timeout value to the poll call. It's just that from the socket class you don't have that control. You DO have that control if you use the basic_socket_streambuf class and wrap the socket in an iostream. That's the odd thing. The functionality IS available in some form, so it's clearly recognised as useful at some level, it's just never been pushed down into the socket itself.

1 0

[boost.test] floating point comparison with user defined type
by Georgios Sermaidis 22 Mar '18

22 Mar '18

Hello, I have created a class to represent the euro currency and I am trying to use boost.test to check if two euro amounts are sufficiently close. I was expecting the following to pass the test given the tolerance level I have set, but instead it fails. What is the proper way to compare two user-defined decimals? Thank you for your time. #define BOOST_TEST_MODULE My Test #include <boost/test/included/unit_test.hpp> class eur { long double m_amount; public: eur(long double amount) : m_amount(amount){} bool operator ==(const eur& rhs) const { return m_amount == rhs.m_amount; } bool operator !=(const eur& rhs) const { return m_amount != rhs.m_amount; } friend std::ostream& operator <<(std::ostream& os, const eur& rhs); }; std::ostream& operator <<(std::ostream& os, const eur& rhs) { os << "EUR " << rhs.m_amount; return os; } BOOST_TEST_DECORATOR(* boost::unit_test::tolerance(eur(0.1))) BOOST_AUTO_TEST_CASE(first_test) { eur x(1.1); eur y(1.2); BOOST_TEST(x == y); }

2 1