November 2010 - Boost-users - lists.preview.boost.org

asio UDP async receive question - how to do this properly?
by Ákos Maróy 25 Nov '10

25 Nov '10

Hi, I wonder what the proper pattern is to receive messages, store them, and process them in an async manner? What I'm trying to do, and which doesn't work: - receive the packets in an async call - copy the data over - start to receive again after copying - store the data using a producer - consumer pattern, using a mutex - process the data in another thread (which is the consumer in the pattern) but, this doesn't seem to work, as a lot of packets are lost. my assumption is that because the handler function blocks on the mutex, and returns 'late', packets received in the meantime are lost forever. but: how to do this properly? I have a small sample here: http://pastebin.com/NnbK1T7G , which is based on one of the asio examples. the sample code is quite simple: - it allows sync sending of messages (strings) - it allows async reception of such messages - it will put each received message into a vector, which is guarded by a mutex - another thread will process messages from this vector on a regular basis, which processing is also guarded by a mutex. if I compile & run the sample, I get: ./async_asio b sent 20000 messages a received 1981 messages which is a profound loss of packets. as seen in the sample, after the receive handler is called, the recieved message is copied over, and socket::async_receive_from() is called ASAP. I don't see what else would need to be done to have all messages received. If I remove the processing of messages in threa_func(), the results are markedly better: ./async_asio b sent 20000 messages a received 19963 messages increasing the sleeping time at the end does no good - it seems the messages were lost already. Akos

2 4

A question about boost thread
by x57978 25 Nov '10

25 Nov '10

Hi, Does somebody know whether the boost thread can set the affinity like other thread? Thanks. Hongmiao

1 0

[asio] strange behaviour on android
by Igor R 24 Nov '10

24 Nov '10

Hello, Recently I ported my application to Android, and after some testing I encountered the following strange behavior. Long story short: I send text to a peer and sometimes it gets "cut" in the beginning. like this: aaa\r\nbbb\r\nccc\r\nddd\r\n becomes \nccc\r\nddd\r\n I checked with sniffer, and ensured that the problem is really in my application. To send the text I use the following simple approach: * io_service::run is always invoked from one thread * the data comes from another thread and gets posted to the io_service thread: void connection::send(const std::string &str) { io_.post(boost::bind(&connection::doSend, shared_from_this(), str)); } * then it's streamed into asio::streambuf: void connection::doSend(const std::string &str) { std::ostream(&output_).write(str.c_str(), str.length()); } * and finally async_write is invoked, IF it's not in progress: if (!activeWrite_) { activeWrite_ = true; async_write(m_output, boost::bind(&connection::sent, shared_from_this(), asio::placeholders::error)); } (of course, in sent() handler async_write is re-activated, if needed). I understand that I add data to the streambuf, while async_write is in progress, but it's done in the same thread, and seems to be legitimate. The above approach works on Windows and on some other platforms, it was stress-tested, and never showed any problem. OTOH, on Android the above issue happens very frequently (~20% of events). The only difference here is that I compiled my project with BOOST_ASIO_DISABLE_FENCED_BLOCK, because otherwise the linker complains it doesn't see definitions of some gcc synchronization intrinsics. Is it possible that this's the problem? As far as I see, fenced_block is used when a completion handler is called, so I can't see how it can be connected to my problem... Anyway, I'd appreciate any idea. Thanks.

1 1

cross platform high resolution timer
by Sean Farrow 24 Nov '10

24 Nov '10

Hi: Does boost contain a cross platform high resultion timm that allows a function to run every resolution. Something that is equivalent to the win32 multimedia timer. Any help appreciated. Sean.

1 0

[Interprocess] Different outputs under bash and gdb
by Aditya Gadre 24 Nov '10

24 Nov '10

Dear all, I have a question about proper use of Boost.Interprocess for sharing custom C++ classes. I recently started using Boost.Interprocess and have been successful in sharing boost::array, std::vector<float>, std::vector< boost::array<float, 2> > etc using boost::interprocess::managed_shared_memory and boost::interprocess::allocator. Currently, I am trying to share a custom matrix class that is used to hold data in a tabular format. The Matrix class (attached header file Matrix.hpp) is a simple wrapper class that uses std::vector to store objects and provides functions to find items using linear and matrix indexing. I have modified a boost::interprocess example program to try out sharing the Matrix class (attached file test_matrix_shm.cpp). The parent process creates a managed memory segment along with a Matrix in the shared memory. The child process accesses the shared memory and successfully finds the Matrix object. However, when I run the program under bash, the output of the child process shows wrong data. However, if I run the same program under gdb, then every time the child process shows correct Matrix data as was set by the parent process (please find sample outputs below). I am really not sure what would cause this behavior. I would appreciate any input in this regard. * bash output * parent process Matrix size = 2 x 2 0 1 2 3 child process Matrix size = 2 x 2 -1.2522e-41 -1.25486e-41 -1.2571e-41 -1.25949e-41 child returning parent returning * gdb output * GNU gdb (GDB) SUSE (7.1-3.12) (gdb) run Starting program: test_matrix_shm parent process Matrix size = 2 x 2 0 1 2 3 Detaching after fork from child process 2548. child process Matrix size = 2 x 2 0 1 2 3 child returning parent returning Program exited normally. Thanks - Aditya

2 7

[boost::asio] link errors using boost 1.44.0 on 32-bit
by Samir Kharchi 24 Nov '10

24 Nov '10

Hi, I am mainly using boost::asio and I get a link error when I try to build my project on 32-bit release VC8.0 saying (The according 32-Bit libs are correctly included): error LNK2019: Verweis auf nicht aufgelöstes externes Symbol ""class boost::system::error_category const & __cdecl boost::system::system_category(void)" (?system_category@system@boost@@YAABVerror_category@12@XZ)" in Funktion ""class boost::system::error_category const & __cdecl boost::system::get_system_category(void)" (?get_system_category@system@boost@@YAABVerror_category@12@XZ)". error LNK2001: Nicht aufgelöstes externes Symbol ""class boost::system::error_category const & __cdecl boost::system::system_category(void)" (?system_category@system@boost@@YAABVerror_category@12@XZ)". However, compiling my project in 64-Bit release works just fine (according 64-bi libs included) and I don´t get these errors. Anybody knows what´s going on here? :-/ I searched the boost list for previous problems but none of the topics I found helped. Thank you in advance Samir

2 2

undefined references '__sync_fetch_and_add_4 ' for Boost_1_39_0 on Alpha platform
by Yan Wu 24 Nov '10

24 Nov '10

Hi, I got an error when building OpenOffice.org on Alpha EV6 platform, the error log is as following: Compiling: comphelper/unxlngaxp.pro/misc/comphelp4gcc3_version.c Making: comphelper.lib Making: libcomphelp4gcc3.so ../unxlngaxp.pro/slo/enumberablemap.o: In function 'comphelper::EnumerableMap::initialize(com::sun::star::uno::Sequence<com::sun::star::uno::Any> const&)': '__sync_fetch_and_add_4' undefined reference ./unxlngaxp.pro/slo/enumberablemap.o: In function 'comphelper::EnumerableMap::initialize(com::sun::star::uno::Sequence<com::sun::star::uno::Any> const&)': '__sync_fetch_and_add_4'undefined reference ../unxlngaxp.pro/slo/enumberablemap.o: In function 'comphelper::EnumerableMap::initialize(com::sun::star::uno::Sequence<com::sun::star::uno::Any> const&)': '__sync_fetch_and_add_4'undefined reference ../unxlngaxp.pro/slo/enumberablemap.o: In function 'comphelper::EnumerableMap::initialize(com::sun::star::uno::Sequence<com::sun::star::uno::Any> const&)': '__sync_fetch_and_add_4'undefined reference ../unxlngaxp.pro/slo/enumberablemap.o: In function 'comphelper::EnumerableMap::initialize(com::sun::star::uno::Sequence<com::sun::star::uno::Any> const&)': '__sync_fetch_and_add_4'undefined reference ../unxlngaxp.pro/slo/enumberablemap.o: more undefined references to '__sync_fetch_and_add_4' follow collect2: ld returned 1 exit status dmake: Error code 1: while making '../unxlngaxp.pro/lib/libcomphelp4gcc3.so' The <boost/shared_ptr.hpp> of Boost_1_39_0 is used here. Is this a bug on Boost_1_39_0 ? and is there already a fix for it? Best Regards, Yan

4 8

boost thread query
by Sean Farrow 24 Nov '10

24 Nov '10

Hi: I have two threads that are started by a class created by the applications main thread. Can the thread function be a private member of that class or does it need to be public? When the application exits I need the threads to be notified and exit. What is the best way of handling this? I need this to be cross platform but in windows I would have used a waitforSingleObject and signalled the event when the main thread wants to exit. Could I use a condition for this, and do a do while loop waiting on the condition to be notified, if I do this do I need a mutex just for the condition? Any help appreciated. Cheers Sean.

2 3

[boost.test] Floating point comparisons
by Anders Wang Kristensen 24 Nov '10

24 Nov '10

Hi, I'm using boost.test and BOOST_CHECK_CLOSE to verify that some matrices I compute are equal to some reference matices. This works fine as long as no entries are zero in either matrix (reference or computed). However, when an entry is zero, BOOST_CHECK_CLOSE always fails, which makes sense, since relative error is undefined in such cases. If either entry is zero, I can use BOOST_CHECK_SMALL instead with an absolute error: if (ref==0.0 || val==0.0) BOOST_CHECK_SMALL(ref - val, absolute_error); else BOOST_CHECK_CLOSE(ref, val, relative_error); But is this really the best way to do this? And if this is the standard way of doing this, shouldn't there be function in boost.test for automating this, like: BOOST_CHECK_NEAR(ref, value, relative_error, absolute_error); kinds regards, Anders (sorry if this post appears twice)

3 3

[Build] System/integration testing?
by ajb＠spamcop.net 24 Nov '10

24 Nov '10

G'day. Boost.Build has fabulous facilities for unit testing, but I'm curious what is "best practice" for system testing. The simplest thing would be to run a shell script (portability isn't an issue here), passing it the path to all of the built binaries, but I'm sure someone must have already done something better than this. Thanks, Andrew Bromage

1 0