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

Understanding fibers
by Stephan Menzel 18 Dec '18

18 Dec '18

Hello @ll, I am in the process of transforming a sync server application to an async approach. The framework in use (not asio) instantiates and calls an object representing the call when a request comes in, allows me to do things, fill a response object and then expects me to call a finish() method which will send the response and start the next request. This is not what this question is about but more like my circumstances. All code is pseudocode please... void rpc_work() { const value = do_stuff(request); response.fill_value(value) parent->finish(); } Now some of the calls will imply calling methods, which block on some operations, waiting for external resources to respond. Obviously in an async server with only one thread, this will block everything. So I'm looking for a clever way to structure this. My approach so far (in the prototype) is to go boost::async() for those operations. Like this: void rpc_work() { // prepare work boost::async(launch::any, [request, response, parent]() { const value = do_blocking_stuff(request); response.fill_value(value) parent->finish(); }); } The way I understand boost::async or indeed std::async is, that normally it will just spawn a thread for the operation. The best I can hope for is a thread pool but this is not guaranteed either (on Linux x64 gcc 7.3). So I started looking into fibers but I'm having a hard time understanding them, which is why I post here. From what I gather, what I have here is a prime use case as it matches many things the docs talk about. And yet they also suggest that the fibers are very intrusive and everything underneath do_blocking_stuff() would have to be aware of being in a fiber and yield() when appropriate. Is this correct? What would this mean? Most if not all of the blocking operations are using boost::thread::futures in their blocking things. They would do things like... value_t do_blocking_stuff(request) { boost::future<value_t> fv = retrieve_value_from_somewhere(); fv.wait_for(timeout); return fv.get() } Now, considering I would use fibers in the server, would I have to use 'fiber futures' here to make this work? To make the blocking functions automatically 'fiber aware' and yield() when they would block? This would imply replacing the thread::futures? And if so, how will I know when the value is there and they are ready to continue? And how will I cause this continuation? Do I have to keep track of them and somehow poll() them in a loop or something? I am aware of the broad nature of my question but perhaps someone can chime in and give some hints that allow me to understand them better. They look like they would fit my use case perfectly but I am a bit reluctant of dragging the concept all throughout the code until I realize it won't work. Perhaps someone can eli5 this to me ;-) Cheers, Stephan

4 13

Boost Process fails on Solaris 10 in async mode
by Sanket Sharma 18 Dec '18

18 Dec '18

Hi, I tried installing boost process on my Solaris 10 Sparc machine. On compiling it using gcc-6.3.0, I ran into a static assert failure at compile time at https://github.com/boostorg/process/blob/boost-1.68.0/include/boost/process…. In order to fix that, I changed static_assert(!WIFEXITED(still_active) && !WIFSIGNALED(still_active), "Internal Error"); to static_assert(!WIFEXITED(still_active), "Internal Error"); There is some discussion about this at https://www.openwall.com/lists/musl/2018/08/10/1 Now, the following code works well onCentOS 7(where no change was made to Boost process), but fails on Solaris 10 (where I made the above change as a workaround). On Solaris 10, the return code is 127 and stdout and stderr are empty. #include <iostream>#include <thread>#include <boost/process.hpp>#include <boost/asio.hpp> namespace bp = boost::process ;int main(int argc, char** argv){ std::future<std::string> dataOut; std::future<std::string> dataErr; std::string command = "ls"; boost::asio::io_service ios; bp::child c(command, bp::std_out > dataOut, bp::std_err > dataErr, ios); ios.run(); c.wait(); int result = c.exit_code(); std::string stdOut = dataOut.get(); std::string stdErr = dataErr.get(); std::cout << "Exit Code " << result << std::endl; std::cout << "stdOut " << stdOut << std::endl; std::cout << "stdErr " << stdErr << std::endl; return 0;} Output on CentOS 7 Exit Code 0 stdOut a.out main.cppProcess.cppProcess.h simple.cpp stdErr Output on Solaris 10 Exit Code 127 stdOut stdErr 1. Has my workaround to install Boost process on Solaris 10 broken it? 2. Is Boost-1.68.0 not supported on Solaris? 3. Is there a way I can get Boost Process to work on Solaris 10 in async mode? https://stackoverflow.com/questions/53780273/boost-process-fails-on-solaris… -- Regards, Sanket

1 0

Boost version 1.69.0 already included in list of known versions in FindBoost.cmake module
by Osman Zakir 17 Dec '18

17 Dec '18

I looked in the FindBoost.cmake module and saw that version 1.69.0 is already listed as a known version. But when I configured CMake to build Jinja2Cpp after building Boost 1.69.0, I couldn't get it to find some modules of Boost. There were warning messages about imported targets and dependences or some such. Jinja2Cpp uses an earlier version of Boost from its own submodule in this case (it has a GitHub repository that I cloned from and built), but I'd like it if it used the latest version of Boost instead of an earlier version. My Boost installation is in C:/boost_1_69_0/, by the way. Could this be causing problems, or will it still be fine I configure CMake correctly? What CMake option should I use to find Boost? -DBOOST_ROOT?

3 5

[Spirit] Comment skipper or first class member of grammar
by Michael Powell 17 Dec '18

17 Dec '18

Hello, I am turning a corner in my Protocol Buffers grammar, and it is now time for me to consider how best to approach parsing comments. Chiefly, I am weighing two options, I think: 1) A white-space, comments skipper; if memory serves there has been discussion about such a thing on either here, Boost mailing list, or other forums. 2) Or, as with some compiler environments, I'm thinking .NET Roslyn/C#, in particular, where "comment" elements are first class members of the grammar/tokens, quite literally, I think. I'm not quite sure how that would be expressed in terms of synthesized AST, when you could literally have a comment appearing virtually anywhere between tokens, especially of the C variety, i.e. /* this is a comment which may appear virtually anywhere between tokens, potentially */. Or the single line C++ variety, for example, // this is a comment through the rest of this line. // Yet, there could be multiple lines of these... optional myVar = true; // And can appear at the end of a line Thoughts? Suggestions? My first approach I think would be for (1) the skipper; however, for what I am wanting to accomplish, I can see there potentially being some value in capturing the comments and carrying them forward with the target language adapter. Best regards, Michael Powell

1 1

[process] distinguish normal exit from termination by signal - complete
by Oliver Kowalke 17 Dec '18

17 Dec '18

Hello, I'm evaluating boost.process. It seams to me that it is impossible to distinguish a normal exit from a termination by a signal (for instance SIGBUS etc.). child::exit_code() calls eval_exit_status() that expands on POSIX to: if (WIFEXITED(code)) { return WEXITSTATUS(code); } else if (WIFSIGNALED(code)) { return WTERMSIG(code); } else { return code; } With the code above it is impossible if the returned exit_code is a `exit-code` or the number of a signal. Or do I miss something? regards, Oliver

1 0

[process]distinguish normal exit from signaled not possible?
by Oliver Kowalke 17 Dec '18

17 Dec '18

Hello, I'm evaluating boost.process. It seams to me that it is impossible to distinguish a normal exit from a termination by a signal (for instance SIGBUS etc.). child::exit_code() calls eval_exit_status() that expands on POSIX to: if (WIFEXITED(code)) { return WEXITSTATUS(code); } else if (WIFSIGNALED(code)) { return WTERMSIG(code); } else { return code; }

1 0

[contextual async] generator support added
by Christophe Bailly 16 Dec '18

16 Dec '18

Hello, I am still waiting for some feedbacks. I really think it makes possible to add asynchronous code very easiy, I have added the support of generators which can run with asynchronous.code. You will find the added article in the blog: https://pythonc.home.blog/ The project has been updated and the examples are available if you want to test. Regards, Chris

2 2

[corountine2/context] forced_unwind exception on Windows with TDM-GCC
by Hao Jiang 15 Dec '18

15 Dec '18

Hello all, I am porting my program from Linux to Windows with boost coroutine2. I found 32bit program throwing boost::context::detail::forced_unwind every time when coroutine destructed. But 64bit program works like charm. terminate called after throwing an instance of 'boost::context::detail::forced_unwind' The program works well on Linux(Ubuntu). Then I tried the official example: typedef boost::coroutines2::coroutine<int> coro_t; coro_t::pull_type source( [&](coro_t::push_type& sink){ int first=1,second=1; sink(first); sink(second); for(int i=0;i<8;++i){ int third=first+second; first=second; second=third; sink(third); } }); for(auto i:source) std::cout << i << " "; Same thing happened. Platform: windows 7 64 bit Boost: 1.67/1.69 Compiler: TDM-GCC 5.1 64bit. Best,

3 4

project-config.jam File for building Boost with LLVM toolchain?
by Osman Zakir 15 Dec '18

15 Dec '18

Sorry, but I need to see the project-config.jam file I need to use for building Boost with the LLVM toolchain again. I'm going to build it with the MSVC toolchain first and then again with the LLVM toolchain as well so that I have both the MSVC- and Clang-generated binaries. I downloaded Boost version 1.69.0 and I want to build it now. Thanks in advance.

5 11

[boost] mixing async + sync code easily via boost::context
by Christophe Bailly 15 Dec '18

15 Dec '18

Hello, I have worked on a prototype to avoid the current limitations with async code, the use of sync + async code is really difficult, it does not co-exist easily. If you introduce async code, you have to migrate all your code ... There are alternative solutions (threads..) but here is a suggestion to overcome these limitations. It is currently a very simple prototype but I would be interested to have your feedback. I think this kind of solution can be really powerfull, but I will wait to know what you think about it. Here are the links: https://pythonc.home.blog/ https://github.com/chbailly/5a5 Regards, Chris

1 0