July 2015 - Boost-users - lists.boost.org

Problem with boost.log in a multithread application
by Rajalakshmi Iyer 17 Jul '15

17 Jul '15

Hello, I am using Boost.Log to produce 2 logs for a multi-thread application. There are 2 singleton channel_mt loggers corresponding to 2 files. Several threads continuously write different messages to the two loggers. Log file rotation is enabled on both loggers and in the event of rotation, the logs are collected into a different folder. All of this works as expected for a while, except randomly (but surely at some point) - I start seeing a stream of messages like - [2015-Jul-17 01:11:26.532653] (info): *** Exception: boost::filesystem::last_write_time: No such file or directory: "/var/infinity/rtb/auction_log" [2015-Jul-17 01:11:26.532896] (info): Boost Exception - libs/log/src/text_file_backend.cpp(1160): Throw in function void boost::log::v2_mt_posix::sinks::text_file_backend::consume(const boost::log::v2_mt_posix::record_view&, const string_type&) This goes on for a while and then my application hangs. It is not entirely clear whether this incidence happens only during size based rotation or even otherwise. Any pointers on how to approach this issue would be greatly appreciated. Thanks Raj -- This email and any attachments to it may be confidential and are intended solely for the use of the individual to whom it is addressed. Any views or opinions expressed are solely those of the author and do not necessarily represent those of BlisMedia Ltd, a company registered in England and Wales with registered number 06455773. Its registered office is 3rd Floor, 101 New Cavendish St, London, W1W 6XH, United Kingdom. If you are not the intended recipient of this email, you must neither take any action based upon its contents, nor copy or show it to anyone. Please contact the sender if you believe you have received this email in error.

1 0

Is this a known bug in lexical_cast<double> ?
by TILLMAN, MICHAEL D9 16 Jul '15

16 Jul '15

We're using boost 1.51. Management is reluctant to upgrade because the software is for manned space flight and our current boost version has been stable. Wanted to see if there was a known bug with boost::lexical_cast<double>. We're seeing failures on what appear to be valid float strings. When we catch the exception, and simply do a second cast on the same string, the cast succeeds. The Code fragment: try { dVal = boost::lexical_cast<double>(theValue); } catch(boost::bad_lexical_cast &eX) { ... more code ... firstAttemptFailed = true; } // If the first cast succeeded, don't bother to attempt it again. if(firstAttemptFailed == true) { // Let the API throw the actual exception if the second attempt also fails. dVal = boost::lexical_cast<double>(theValue); } The string sequence this fails on is: ... ... "7104", "0.416666699999999999983", "0.22222224", "4.08233160000000000001", "1.81436960000000000002", ... ... It fails on the last string, "1.81.....", in the actual program, but when compiled into a simple test program, it succeeds. If there is a known bug, we may be able to get a Boost upgrade installed. Thanks, Mike

9 12

Re: [Boost-users] Query regarding interest in my submitting plf::colony to Boost
by Colin Humphries 16 Jul '15

16 Jul '15

> I don't think anything related to multithreading should be done by such kind of container. Disturbed by that comment I just wanted to add for the benefit of anyone just starting out with multi-threading that in my experience containers are entirely safe for multi-threading PROVIDED that you're only reading from them (i.e. they are passed by const &). This is by far the most efficient way to pass common information into parallel threads (all threads read the same data). For containers required to return a result, the approach that works for me is to create a vector of containers and dedicate one to each thread. The result containers must be guaranteed to remain in scope until all threads have completed so I usually make them global. The only challenge is how to efficiently coalesce the partial results. Get this wrong (or use too many threads) and the performance gain from multi-threading quickly evaporates. Colin Humphries

2 1

[Graph] pgRouting and boost.graph
by Vicky Vergara 16 Jul '15

16 Jul '15

Hello, We work on the pgRouting project and we are updating our code to use Boost.Graph. For the moment our intention is to have a V2.1 of pgRouting modifying 3 of our routing functions, and I am working on the documentation. I am making references to Boost.Graph, I would like some one from Boost.Graph team to check if the references I am making are according to your requirements and /or give us some suggestions. The following link takes you to the tentative documentation that mentions Boost.Graph. https://github.com/pgRouting/pgrouting/blob/develop_2_1_0/src/dijkstra/doc/… Thanks in advance. Vicky

2 2

Boost graph parallel issue
by Mayank Bakshi 16 Jul '15

16 Jul '15

2 3

How do I build boost with a specific compiler flag?
by Gagnon, Joseph - 0553 - MITLL 15 Jul '15

15 Jul '15

Hello, I am developing a C++ application on a Linux (CentOS 6.6) platform with gcc 4.8.2. I need to build the Boost libraries I'm using (Regex and others) such that the code is compiled with the "-fPIC" (position-independent code) option. I looked around on the Boost web site and found information about a configure script and something called Boost.Config. I quickly got overwhelmed. What do I need to do to have the Boost build system use this flag to build its libraries? Is it something fairly simple, or do I need to invoke a complicated set of steps? Thanks, Joe Gagnon Engineer II Group 53 - Secure Resilient Systems and Technology MIT Lincoln Laboratory Lexington, MA

2 1

[Test] Did This Pass, Or Didn't it?
by Merrill Cornish 15 Jul '15

15 Jul '15

Here is the Boost Test assertion: BOOST_CHECK_EXCEPTION(dummyOpr->setOperand(emptyValue), DiGSE::fail, fail_msg); Here is the test execution log message: info: check 'incorrect exception DiGSE::fail is caught' has passed I have only one exception: *DiGSE::fail* (derived from *std::exception*). It is the /only /exception that appears in the *BOOST_CHECK_EXCEPTION()* assertions. So, what does "incorrect exception *DiGSE::fail*" mean? If *DiGSE::fail* was caught, then that part is correct. If the message means that some /other/ exception was caught, then this is a Really Poor way of phrasing that information. Or, did the predicate return *false* (which, in my case, means the exception message is wrong)? But if the wrong exception was caught /or/ the predicate returned *false*, then why was the assertion declared "has passed"? Merrill Cornish

2 1

Test: Different versions of doc (raffienficiaud)
by Florian Lindner 15 Jul '15

15 Jul '15

Hello, additionally to the official documentation of boost test at [1], I have found a different version at [2]. This looks like the official documentation but describes much more features and more comprehensive. However, although not the development version at [3]. What does the version from raffienficiaud describes? Is this some own extensions of him? Thanks, Florian [1] http://www.boost.org/doc/libs/1_58_0/libs/test/doc/html/index.html [2] http://raffienficiaud.github.io/test/index.html [3] http://svn.boost.org/svn/boost/trunk/libs/test/doc/html/index.html

3 2

Boost.DLL formal review is ongoing
by Vladimir Prus 15 Jul '15

15 Jul '15

The formal review of Boost.DLL library by Antony Polukhin is nearing the end of the first week. As of today, over 100 people read through the library's tutorial, suggesting quite some interest. However, no reviews were submitted yet. Please take the time to post your thoughts - even if you don't have time to do a full review, or try the library, comments on design and interfaces are still very valuable. The summary of the library features and review checklists are reproduced below. Boost.DLL is a C++98 library for comfortable work with DLL and DSO. Library provides a portable across platforms way to: - load libraries at runtime - import and export any native functions and variables - make alias names for C++ mangled functions and symbols - query libraries/objects and executables for sections and exported symbols - self loading and self querying - getting program and module location by exported symbol The documentation can be found at: http://apolukhin.github.io/Boost.DLL/index.html and the source can be obtained at: https://github.com/apolukhin/Boost.DLL Please post your reviews on the mailing list and if possible, answer the following questions: - Should the library be accepted? - How useful is it? - What's your evaluation of - Design - Implementation - Documentations - Tests - How much effort did you put into your evaluation? - Did you attempt to use the library? On what systems and compilers? Thanks, Volodya _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

5 8

the stability of breadth_first_visit in BGL
by breadbread1984 14 Jul '15

14 Jul '15

I am wondering if breadth_first_visit can always get the same sequence of vertex no matter whether the graph is serialized and deserialized again. Thx

4 3

Subgraph at distance n
by Rossano Schifanella 14 Jul '15

14 Jul '15

Hi all, which is the easiest way to extract from a graph G and a generic node source a subgraph with all the vertices/edges at distance at most n from source? Thanks a lot for your help. Best, Rossano

2 1

Boost.DLL Review is finished
by Vladimir Prus 14 Jul '15

14 Jul '15

Boosters, the formal review of the Boost.DLL library by Antony Polukhin is now finished. The preliminary results are as follows: The following people have provided reviews: - Ralf Globisch - Yes - Klaim - Joël Lamotte - Yes (*) - Niall Douglas - Yes, conditionally - John P Fletcher - Yes - Bjorn Reese - Yes - Edward Diener - Yes Klaim - Joël has asked that his vote is not counted until he's done trying the library for real, and there was no confirmation it's done yet. The following people have contributed, but did not provide a review - Rodrigo Madera - Thomas Trummer - Rob Stewart - Ion Gaztañaga If anybody would like to amend or clarify anything, or plans to provide additional comments during this week, please contact me off-list. I will post final review result after that. Thanks, Volodya

4 5

Platform checks fail on Centos 6 after bootstrap ?
by gabriel nützi 12 Jul '15

12 Jul '15

Hello all :-) Is somebody able to help with the following problem: http://stackoverflow.com/questions/31347026/boost-build-on-centos-6-5-faile… <http://stackoverflow.com/questions/31347026/boost-build-on-centos-6-5-faile…> boost 1.58 used I dont know why platform checks are not done correctly? Thanks a lot!

2 1

[Boost-announce] [boost] [review] Metaparse Review Result
by Christophe Henry 11 Jul '15

11 Jul '15

Hi all, First of all, the final result. Metaparse is formally unconditionally accepted into Boost. We had 11 reviews, all "Yes" votes. 2 "No" reviews based on formality were later reversed to a "Yes". The review was lively, with a general agreement that the library is very useful, the design of high quality and the documentation great. Here are some noticeable quotes which helped me take my decision: - "thank you Abel for submitting such an amazing library. The quality of the documentation and code is wonderful and it has been a pleasure to review Metaparse." (Michael Caisse) - (Design) "Of high quality. Based on the MPL paradigm, but I don't view this as a problem." (Peter Dimov) - "Part of the role of boost is to push compilers to the limit (and sometimes beyond). This library certainly has the capabilities to do so." (Roland Bock) Worth noticing is the review of Gordon Woodhull. I myself wrote a new front-end for MSM using Metaparse, and other use cases were very welcome and Gordon managed to produce one for his compile-time graph library by the end of the review, which I found impressive (for both the reviewer and the library). Thank you for investing time in this review. Another important factor for me for acceptance is how well the author is going to maintain his library. We have too many abandoned libraries. Abel went out of his way to oblige remarks and demands made by reviewers. Long before the reviews he also had obliged my many critics and demands. He also showed himself open for experiments. All this while keeping a very calm and professional tone. This is for me a very good sign of a library which will be well-maintained. Finally, there were some repeated important comments which I see as adressed to the rest of Boost that the library is targeted towards library writers. >From Peter Dimov: > - Are you knowledgeable about the problem domain? "I'm not sure if anybody is. :-)" It all depends on us, library and potential library writers to become knowledgeable about this domain and make this library move from potentially useful to very useful. We need to use it to write more cool libraries. I notice a recurring theme where Boost would be dead or on the way to doom etc etc. Some very popular Boost libraries are now part of the Standard, and so what? We need to make new ones. Metaparse is one of them. Let's build more, using it! Yes Reviews(11): Andrzej Krzemienski Bjorn Reese Gordon Woodhull Lee Clagett Louis Dionne Michael Caisse Paul A. Bristow Peter Dimov Roland Bock Edward Diener Niall Douglas I want to thank all the reviewers who invested their time in this review and those who provided some usefuls comments. I also want to thank Abel for presenting this great library. Christophe Henry Metaparse Review Manager

1 0

Slow decoding in thread
by Gonzalo Garramuno 10 Jul '15

10 Jul '15

I am writing a movie player using boost. I use a condition_variable to control a mutex in one thread that waits for packets of video to be ready for decoding, while another thread feeds those packets. My problem is that the decode thread may end up doing a lot of work when playing backwards and it is not ready to wait on the condition variable when the message comes from the other thread and so it hangs (or so I think). This is my wait loop. I apologize for the use of macros. #define SCOPED_LOCK(mutex) Mutex::scoped_lock lk_##mutex(mutex) #define CONDITION_WAIT( cond, mutex ) cond.wait( lk_##mutex ); void aviImage::wait_image() { mrv::PacketQueue::Mutex& vpm = _video_packets.mutex(); SCOPED_LOCK( vpm ); for(;;) { if ( stopped() || ! _video_packets.empty() ) break; CONDITION_WAIT( _video_packets.cond(), vpm ); } return; } When it returns it tries to decode one or more frames of video taken from _video_packets. The condition variable is notified whenever the program calls push_back() to add a packet. That is: class PacketQueue { .... inline void push_back( AVPacket& pkt ) { Mutex::scoped_lock lk( _mutex ); // assert( pkt.dts != MRV_NOPTS_VALUE ); if ( av_dup_packet(&pkt) < 0 ) { std::cerr << "Could not duplicate packet - not added" << std::endl; return; } _packets.push_back( pkt ); _bytes += pkt.size; _cond.notify_one(); } }; Can somebody suggest an alternative to this scheme? -- Gonzalo Garramuño

1 1

Query: metis_input_exception
by Mayank Bakshi 10 Jul '15

10 Jul '15

1 0

crc updating how?
by U.Mutlu 09 Jul '15

09 Jul '15

Hi, I need to calc the crc chunkwise (ie. calc crc of a block, then continue with the next block etc.). Is this (not) possible with the boost crc? In other words: how can the type "crc_32_type" be initialized with a user supplied uint32_t value? boost::crc_32_type crc; crc = 0x1234; // <-- this brings compile error crc.process_bytes(buf, len); error: no match for ‘operator=’ (operand types are ‘boost::crc_32_type {aka boost::crc_optimal<32ul, 79764919u, 4294967295u, 4294967295u, true, true>}’ and ‘const uint32_t {aka const unsigned int}’) Thx

2 3

[crc] bug in boost::crc
by U.Mutlu 09 Jul '15

09 Jul '15

/* boost-crc-bug-demo.cpp Written by U.Mutlu on 2015-07-09-Th Compile: g++ -O2 -Wall -std=gnu++11 boost-crc-bug-demo.cpp Run: ./a.out Output: crc("The quick brown fox jumps over the lazy dog")=414fa339 crc("The quick brown fox ")=88b075e2 crc("The quick brown fox jumps over the lazy dog")=4ac56d0f ERROR See also: http://www.boost.org/doc/libs/1_58_0/libs/crc/crc.html#usage Conclusion: It doesn't do what it should and also advertises; ie. there's a bug */ #include <boost/crc.hpp> #include <boost/cstdint.hpp> #include <iostream> int main() { uint32_t expected; { const std::string sTxt = "The quick brown fox jumps over the lazy dog"; boost::crc_optimal<32, 0x04C11DB7, 0xFFFFFFFF, 0xFFFFFFFF, true, true> crc; crc.process_bytes(sTxt.data(), sTxt.size()); std::cout << "crc(\"" << sTxt << "\")=" << std::hex << crc.checksum() << std::endl; // save the result for the following test expected = crc.checksum(); } // test: doing the same as above, but now in 2 steps, ie. chunk- or blockwise: std::string sTxt; uint32_t result = 0xFFFFFFFF; for (size_t i = 0; i <= 1; ++i) { sTxt += !i ? "The quick brown fox " : "jumps over the lazy dog"; // using the special ctor for initial remainder stored in "result" boost::crc_optimal<32, 0x04C11DB7, 0xFFFFFFFF, 0xFFFFFFFF, true, true> crc(result); crc.process_bytes(sTxt.data(), sTxt.size()); std::cout << "crc(\"" << sTxt << "\")=" << std::hex << crc.checksum() << std::endl; result = crc.checksum(); } // eval: const std::string sRes = result == expected ? "SUCCESS" : "ERROR"; std::cout << sRes << std::endl; return 0; }

3 4

[Interprocess] posix robust mutex and EOWNERDEAD
by Malko 07 Jul '15

07 Jul '15

Hi, I lurked a bit and found this: http://boost.2283326.n4.nabble.com/boost-interprocess-gt-1-45-robust-mutexe… I didn't feel like necroposting, and it's /dev anyway, so allow me to ask here about updates on the matter. Is there anything, even if only "experimental" or "beta", that you have that would allow me to handle the EWOENERDEAD result? I'd rather use your code before putting my hands on your library, since it would be easier to let my manager (and his manager) deal with a "boost beta" rather than my personal modification to boost. If there is none, I would really need to modify your files for my project to work, and I would appreciate if you could confirm a couple of things for me. In http://stackoverflow.com/a/1179766/4598277 it is mentioned to modify two files: 1. posix/thread_helpers.hpp to set the attribute (first time ever I was happy to get an exception :D) 2. posix/mutex.hpp to handle the EOWNERDEAD return value Can you please confirm if this would be enough? It would seem so to me, but I have only studied parts of your code. Thank you for the assistance. -- View this message in context: http://boost.2283326.n4.nabble.com/Interprocess-posix-robust-mutex-and-EOWN… Sent from the Boost - Users mailing list archive at Nabble.com.

2 3

iterator over boost::tuple
by Uthpal Urubail 06 Jul '15

06 Jul '15

Experts, What is the best way to iterator over boost::tuple and return value? Example: For(int i=0;i<boost::tuples::length<T>::value;i++) { Boost_tuple_example.get<i>(); } Regards, UJ

5 4

Boost python and python sets
by Will Hall 06 Jul '15

06 Jul '15

Hi, I'm struggling to find a way to support python sets with boost. The Python API is well defined, so am unable to switch to Python list type. The API needs to support Python sets for both input and output values. I have output values (from C++ to Python) working as follows with to_python namespace bp = boost::python; template<class T> struct to_python; template<class T> struct to_python<std::set<T> > { to_python() { bp::to_python_converter< std::set<T>, to_python< std::set<T> > >(); } static PyObject* convert(const std::set<T>& set) { bp::list values; for(typename std::set<T>::const_iterator iter=set.begin(); iter!=set.end(); ++iter) { values.append(bp::object(*iter)); } PyObject* result = PySet_New(values.ptr()); return result; } static const PyTypeObject* get_pytype() { return &PySet_Type; } }; BOOST_PYTHON_MODULE(helpers) { to_python< std::set<int> >(); bp:class_<MyClass>() // Define constructor with const std::set& arg // Define method that returns const std::set& arg } But I have no idea how to write an equivalent "from_python" to convert Python set to std::set. Any help would be very appreciated!

1 0

{Serialization] Deserialization of const references
by Mccall, Kurt E. (JSC-EG411) 06 Jul '15

06 Jul '15

As near as I can tell from the documentation http://www.boost.org/doc/libs/1_57_0/libs/serialization/doc/serialization.h… , the following is the way to handle const reference members, which are initialized in the constructor. Serialization seems to be working, from an inspection of the resulting XML file, but after deserialization, the values in the object given by the const reference are wrong and are all zero. Here is my question: Is load_construct_data() below correct? It looks like it is initializing the constant reference from a temporary, but I'm doing the same thing as the example in the documentation. class MyClass1 { friend boost::serialization::access; … }; class MyClass2 { friend boost::serialization::access; const MyClass1 & mc1_; MyClass2 ( const MyClass1 &mc1) : mc1_(mc1) { } ; }; namespace boost { namespace serialization { template<class Archive> inline void save_construct_data( Archive &ar, const MyClass2 *t, const unsigned int file_version) { // mc1_ is a reference and requires special handling -- save it // as a pointer const MyClass1 *mc1 = &t->mc1_ ; ar << boost::serialization::make_nvp("mc1_", mc1); } template<class Archive> inline void load_construct_data( Archive &ar, MyClass2* t, const unsigned int file_version) { MyClass1 *mc1; ar >> boost::serialization::make_nvp("mc1_", mc1); // NOTE: IS THIS INITIALIZING THE REFERENCE FROM A TEMPORARY? ::new(t) MyClass2(*mc1); } Thanks, Kurt

1 0

Problems to use Boost Graph
by Oodini 04 Jul '15

04 Jul '15

Hello, I am trying to understand the boost::dijkstra_shortest_paths() function. I don't understand the second and third arguments. The documentation says they're supposed to be properties map : OUT: predecessor_map(PredecessorMap p_map) The predecessor map records the edges in the shortest path tree, the tree computed by the traversal of the graph. Upon completion of the algorithm, the edges (p[u],u) for all u in V are in the tree. The shortest path from vertex s to each vertex v in the graph consists of the vertices v, p[v], p[p[v]], and so on until s is reached, in reverse order. The tree is not guaranteed to be a minimum spanning tree. If p[u] = u then u is either the source vertex or a vertex that is not reachable from the source. The PredecessorMap type must be a Read/Write Property Map whose key and value types are the same as the vertex descriptor type of the graph. but that doesn't work. I have to use an iterator on the properties map, by using : boost::make_iterator_property_map(pred_map.begin(),vertex_index) Does the documentation need reviewers ?

3 4

Re: [Boost-users] [boost] Boost.DLL formal review is ongoing
by Ralf Globisch 03 Jul '15

03 Jul '15

Hi, Firstly, thanks Anthony for submitting this useful library to boost, and to Vladimir for managing the review. - Should the library be accepted? Yes. I had been meaning to test the library for some time, and with the review reminder finally made some time to do so. My vote of yes is based on my (limited) positive experience with the library, as well as having followed past related discussions on the mailing list, in which many questions are addressed. (http://lists.boost.org/Archives/boost/2014/09/216650.php) I think boost.DLL should be accepted unless someone has a serious/valid reservation as to why this library should not to be accepted. - How useful is it? Very useful. It enabled me to add a test-plugin mechanism to an existing application (Windows and Linux) in a matter of minutes. - What's your evaluation of - Design Looks good to me. Simple to use, with good integration into rest of boost (e.g. file_system). I like the fact that the library is header only and minimal effort was required to integrate plug-in like functionality into an existing application. - Implementation I'm not knowledgeable enough to comment on this. - Documentations Good, the tutorials made it fairly easy to get going and covered my common usage scenarios Code is well commented. - How much effort did you put into your evaluation? About three hours worth of effort, reading through documentation and source code, and using library to create a very simple plug-in mechanism in an existing application. Tested basic plug-in functionality on Windows and Linux. - Did you attempt to use the library? On what systems and compilers? Yes -Linux Ubuntu 14.10 with gcc 4.9.1 with boost 1.57 -Window 7 64-bit with VS 2013 community edition with boost 1.58. Initially made the mistake of not checking the boost version (which had been 1.53 and caused lots of compilation errors.) Minor typo in documentation: Just above the "Full Sources" for tutorial 4 "If we want o make a traditional" should be "If we want to make a traditional" Regards, Ralf >>> Vladimir Prus <vladimir.prus(a)gmail.com> 07/03/15 3:52 PM >>> The formal review of Boost.DLL library by Antony Polukhin is nearing the end of the first week. As of today, over 100 people read through the library's tutorial, suggesting quite some interest. However, no reviews were submitted yet. Please take the time to post your thoughts - even if you don't have time to do a full review, or try the library, comments on design and interfaces are still very valuable. The summary of the library features and review checklists are reproduced below. Boost.DLL is a C++98 library for comfortable work with DLL and DSO. Library provides a portable across platforms way to: - load libraries at runtime - import and export any native functions and variables - make alias names for C++ mangled functions and symbols - query libraries/objects and executables for sections and exported symbols - self loading and self querying - getting program and module location by exported symbol The documentation can be found at: http://apolukhin.github.io/Boost.DLL/index.html and the source can be obtained at: https://github.com/apolukhin/Boost.DLL Please post your reviews on the mailing list and if possible, answer the following questions: - Should the library be accepted? - How useful is it? - What's your evaluation of - Design - Implementation - Documentations - Tests - How much effort did you put into your evaluation? - Did you attempt to use the library? On what systems and compilers? Thanks, Volodya _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. Please consider the environment before printing this email. -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. Please consider the environment before printing this email.

1 0

[asio] async udp socket latency
by Andrew Hundt 03 Jul '15

03 Jul '15

I'm having difficulty with latency when running an async asio udp socket. I don't have the same problems with an identical application that uses synchronous udp. I'm communicating with a physical device which permits a maximum 5ms latency for each UDP packet, and I'm currently getting a latency of 8ms with a std dev of 0.065ms which is surprisingly precise. I'm on Mac OS X 10.10.4 with Xcode 6.3.2 (6D2105). Here are some of the key details about this problem: - Sync UDP test program - works, staying under 5ms latency 100% of the time for over an hour - Code - main file KukaFRIClientDataTest.cpp <https://github.com/ahundt/grl/blob/master/test/KukaFRIClientDataTest.cpp> - UDP socket communication class KukaFriClientData.hpp <https://github.com/ahundt/grl/blob/master/include/grl/KukaFriClientData.hpp> - Key lines - calls socket.receive_from() then socket.send() at the lowest level. - receive_from call <https://github.com/ahundt/grl/blob/master/include/grl/KukaFriClientData.hpp…> - send call <https://github.com/ahundt/grl/blob/master/include/grl/KukaFriClientData.hpp…> - Async UDP test program - buggy, 8ms w/ 0.065 ms std dev, continuously high latency - Code - main file KukaFRITest.cpp <https://github.com/ahundt/grl/blob/master/test/KukaFRITest.cpp> - High level wrapper class KukaFRIThreadSeparator.hpp <https://github.com/ahundt/grl/blob/master/include/grl/KukaFRIThreadSeparato…> - Low level wrapper class KukaFRI.hpp <https://github.com/ahundt/grl/blob/master/include/grl/KukaFRI.hpp> . - Key lines - calls socket_.async_receive_from() then that handler calls socket_.async_send_to(), which calls the final callback with the results. - async_receive_from call <https://github.com/ahundt/grl/blob/master/include/grl/KukaFRI.hpp#L374> - async_send_to_call <https://github.com/ahundt/grl/blob/master/include/grl/KukaFRI.hpp#L432> Does anyone have insight into why this latency issue may be occurring in the Async UDP version? Thanks. Cheers! Andrew Hundt

5 7