August 2011 - Boost-users - lists.preview.boost.org

Segmentation fault when using asio and poll_one
by Allan Nielsen 09 Aug '11

09 Aug '11

Hi I'm using Boost 1.46.1 asio for doing serial communication. About 1/1000 times when I run the poll_one I get this segmentation fault: Program received signal SIGSEGV, Segmentation fault. 0x0000000000434a7a in boost::asio::detail::reactor_op::perform (this=0x402010af1010004) at /usr/include/boost-1_46/boost/asio/detail/reactor_op.hpp:40 40 return perform_func_(this); (gdb) bt #0 0x0000000000434a7a in boost::asio::detail::reactor_op::perform (this=0x402010af1010004) at /usr/include/boost-1_46/boost/asio/detail/reactor_op.hpp:40 #1 0x0000000000435da6 in boost::asio::detail::epoll_reactor::run (this=0x6625c0, block=false, ops=...) at /usr/include/boost-1_46/boost/asio/detail/impl/epoll_reactor.ipp:286 #2 0x0000000000436a24 in boost::asio::detail::task_io_service::do_one (this=0x6624e0, lock=..., this_idle_thread=0x0) at /usr/include/boost-1_46/boost/asio/detail/impl/task_io_service.ipp:264 #3 0x0000000000436612 in boost::asio::detail::task_io_service::poll_one (this=0x6624e0, ec=...) at /usr/include/boost-1_46/boost/asio/detail/impl/task_io_service.ipp:188 #4 0x0000000000436e65 in boost::asio::io_service::poll_one (this=0x7fffffffe240, ec=...) at /usr/include/boost-1_46/boost/asio/impl/io_service.ipp:103 #5 0x000000000043b69b in CcTalkPort::receive (this=0x7fffffffe240) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:382 #6 0x000000000043b7dc in CcTalkPort::receive (this=0x7fffffffe240, size=5) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:396 #7 0x000000000043bcb5 in CcTalkPort::_ask (this=0x7fffffffe240, req=...) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:411 #8 0x000000000043ac6e in CcTalkPort::ask (this=0x7fffffffe240, req=...) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:317 #9 0x000000000043daef in CcTalkDevice::ask_hex (this=0x7fffffffe300, v=..., m=..., s=...) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:583 #10 0x000000000043ce69 in CcTalkDevice::request_software_revision (this=0x7fffffffe300) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:507 #11 0x000000000043c57f in CcTalkDevice::coreInfo (this=0x7fffffffe300) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:459 #12 0x00000000004325ff in main (argc=1, argv=0x7fffffffe528) at /home/awn/git/gamblify/hardware/gmbio/software/test/ccTalkScan.cxx:820 Here is the snippet where I call it: while(1) { io_service.reset(); io_service.poll_one(ec); if ( read_result ) { timer.cancel(); return; } else if ( timer_result ) { throw CcTalkTimeOutException(); } } Any hints on how to debug this? -- Allan W. Nielsen

2 1

Working around the restriction in enable_shared_from_this
by kelvSYC 09 Aug '11

09 Aug '11

Suppose I have a class called Group, extending boost::enable_shared_from_this<Group>. Now suppose I have something like this: class Foo { boost::shared_ptr<const Group> G; public: Foo(const Group& G_) : G(G_.shared_from_this()) {} }; This would totally barf if, among other things, G_ was allocated from the stack, say, since I don't have an owning shared_ptr instance. So, it would appear that, at the very least, if you were to allocate a Group on the stack, then you would need a shared_ptr with a null deleter to make use of Foo. This seems inconvenient to whoever may want to use Foo, so is there a way to change this so get around the shared_ptr instance restriction without changing Foo?

1 0

path filesystem
by rudy 09 Aug '11

09 Aug '11

Hi, This is my first post, so please advise me if this is not the correct place to post my question. This is my question: I have the following directory pattern (Dir1 is the parent directory, and the rest are the child files/subdirectories): Dir1 file1 file2 .file3 subdir1 .subdir2 the following code will go through the Dir1, and it will fail in creating the "path" when it reaches ".file3" or ".subdir2" boost::filesystem:ath dir1_path(filepath.c_str()); How should I create a path to a directory that (for example) has the following path? /root/my_docs/pgg/.file1 filesystem::path works just fine if I had to deal with (/root/my_docs/pgg/file1) instead of (/root/my_docs/pgg/.file1) !!! apparently when the file/directory starts with '.', I cannot get filesystem::path to behave properly. please advise, thanks, Rudy

1 0

Boost.Program_options: Parsing arguments like "+a" possible?
by boost_mailinglist＠gmx.de 09 Aug '11

09 Aug '11

Hello, is it possible to parse also arguments like "+a" with the library "Boost.Program_options"? The numbers of arguments are simple to much to use only the "-" or the "--" as a sign for arguments. If it possible, how have I to activate this function? Thank you. -- Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de

4 11

[asio] user defined composed operations and strands
by Tacheon 09 Aug '11

09 Aug '11

Hello, I am trying to implement some new composed operations (like boost::asio::async_read) with asio. However I am a bit concerned about thread-safety when the run-method of the used io_service object ist called from multiple threads and there are parallel read- and write operations on the same socket. The composed operations look like this: class io_port { public: //Ctors and so on... typedef boost::function<void (error_t const &)> completion_handler_t; void async_read_data ( raw_buffer_t & buffer , completion_handler_t const & handler) { //uses a composed operation itself... boost::asio::async_read ( socket , boost::asio::buffer(header) , boost::bind ( &io_port::handle_header_received , this , boost::ref(buffer) , handler , boost::asio::placeholders::error) ); } //first intermediate handler void handle_header_received ( raw_buffer_t & data , completion_handler_type const & handler , boost::system::error_code const & error { if(error) { handler(error_t(error)); } else { prepare_buffer(data, header.perparation_data()); //pseudo-code //uses a composed operation itself... boost::asio::async_read ( socket , boost::asio::buffer(data) , boost::bind ( &io_port::handle_data_received , this , boost::ref(buffer) , handler , boost::asio::placeholders::error) ); } } //second intermediate handler void handle_data_received ( raw_buffer_t & data , completion_handler_type const & handler , boost::system::error_code const & error ) { if(error) { handler(error_t(error)); } else { handler(error_t()); } } void async_write_data ( raw_buffer_t const & buffer , completion_handler_t const & handler) { ... } //quite similar to async_read_data }; usage may be either port.async_read ( buffer , handler ); or port.async_read ( buffer , strand.wrap(handler) ); //how to forward this to asio::async_read? If async_write_data is called and io_service::run is called from multiple threads, strands are should be necessary to avoid raced. The question is: How do I support handlers wrapped in a strand and how do I get this strand forwarded to async_read? As far as I understand this should be somehow possible by using asio_handler_invoke but the documentation is a bit too fuzzy on the topic for me to understand. -- Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de

1 0

Re: [Boost-users] Enums to/from Strings
by Benjamin van den Hout 09 Aug '11

09 Aug '11

Hi John, > For the time being, I've made a function called to_string, which seems to be a common > choice. But what's the best way to make it automatically known to those other features? It's not the most pretty solution, but perhaps you could resort to the Boost Preprocessor library to generate both the enum and the 'toString' method for you. Some sample code can be found here on my blog: http://www.smork.info/blog/index.php?entry=entry110809-105615 Kind regards, Benjamin van den Hout

1 0

Re: [Boost-users] [fusion] adapt std::tuple to be a fusion sequence?
by "Claas H. Köhler" 09 Aug '11

09 Aug '11

I was wondering about the status of the above topic: Does a boost implementation of an adaptor exist or is Sebastian's suggestion the best alternative currently available? His solution did not work out of the box for me, since g++-4.6 complained about an "incomplete type std::tuple_size<const std::tuple<..." However, when I replaced each occurence of std::tuple_size<X> by std::tuple_size< typename std::remove_cv<X>::type > everything seemed to work fine. Regards Claas > No docs. Also, this is just for the MSVC 10 (no-variadics) > >implementation. It was just something I hacked up quickly because I > >needed it for something internal. > > >Anyway, here it is. > > >Sebastian

2 1

Boost::Test and streams
by Jeffrey Walton 09 Aug '11

09 Aug '11

Hi All, I'm working form http://www.boost.org/doc/libs/1_47_0/libs/test/doc/html/index.html. I can't seem to find Boost::Test's streams. The discussion of streams begins at [1]. [1] does list macros such as BOOST_TEST_DONT_PRINT_LOG_VALUE, but I'm interested in a C++-centric solution. For example: // test framework output stream using boost_test::tout; tout << loglevel::error << "My error message" << endl; tout << loglevel::warn << "My warning message" << endl; According to [2], I can set the log level from the environment and from the command line. But would like to set it dynamically. For example: tout << setlevel(warn); // warnings and above or tout.setlelvel(error); // errors and above In addition, I would like to be able to tie a file output to screen output, so I can watch the tests roll by and go to the log file for in depth reading. To be fair, redirection is covered at [3], but its not quite what we are looking for. For example: ofstream logfile(...); ... tout.attach(logfile); or tout.tie(logfile); Though the authors did a great job on the documentation, these topics do not appear to be included. Any ideas how one might go about these sorts of things? Thanks in advance, Jeffrey Walton [1] http://www.boost.org/doc/libs/1_47_0/libs/test/doc/html/utf/user-guide/test… [2] http://www.boost.org/doc/libs/1_47_0/libs/test/doc/html/utf/user-guide/runt… [3] http://www.boost.org/doc/libs/1_47_0/libs/test/doc/html/utf/user-guide/test…

2 1

[boost.asio 1.47] : changing host time
by Mathieu Peyréga 08 Aug '11

08 Aug '11

Hello, I would like to know if there are any implications with the fact of having a call to settimeofday in order to change the system time while using boost asio (asynchronous timers). The point of calling settimeofday (software is running on a linux box with super user privileges) is to synchronize the system clock with a GPS receiver clock at some point while the io service is already running. Best regards, Mathieu Peyréga

2 3

Re: [Boost-users] Asio problems with async_read
by Alex Perry 08 Aug '11

08 Aug '11

Allan Nielsen wrote on Mon, 8 Aug 2011 14:13:08 +0200 > Hi > > I have a funciton where we I am using boost::asio to read from the > serial port of my computer. > > I have started out with an implementation which uses read, and I am > now trying to convert it to async_read because I want a time out. > > When I am trying to use the async_read version, I get timeouts after I > have read one or two charterers, even though I know there are more > charterers to be read on the wire. > Not sure that this is much help since I've never really used boost.asio and reading further down the thread it may be a problem with how you are calling it. However having done a lot of serial comms in the past seeing a couple of characters only whilst more data is on the wire I'd guess that the timeout is as a result of the low level serial hardware interrupt. The UART is probably signalling that its internal FIFO buffer is full and you need to read this promptly - quite how that all pans out and what daemon / service has the responsibility of reading this these days I haven't got the faintest - I was writing TSR's when I knew the gory technical details of this so life was much simpler. I'm actually impressed that you've found a computer that still has a serial port on it! Most seem to have gone over to the ubiquitous USB these days - though that may be the easiest solution - use USB and a usb/serial converter which would avoid having to rediscover the lost art of programming UART latch timings etc and give you a decent amount of low level buffering so you could ignore low level stuff and just read the data. http://www.bb-europe.com/tech_articles/USB_converters.asp?gclid=CPvOgur5v6o… or some such. HTH Alex

2 1