October 2017 - Boost-users - lists.boost.org

PolyCollection requiring copyable
by d3fault 05 Oct '17

05 Oct '17

I'm impressed with PolyCollection. I use the list<interface*> strategy tons in my projects, and I didn't think NOT using a pointer there was even possible... but somehow you managed. It also makes sense why PolyCollection is faster to execute etc, the tutorial was well written. The only 'con' is that it requires a copy constructor. You mention later that some copy traits thing can be used to move the "copy code" outside of the class, but I am questioning the need for copying altogether. Is "copying" of the polymorphic types necessary for: a) Technical requirement otherwise it wouldn't compile b) Design decision because you think it's superior and helpful c) To facilitate copying of the PolyCollection object itself ....? If (b) or (c), can there be a slightly altered version of PolyCollection which does not allow/require copying? I'm completely fine (and even happier) _not_ being able to copy the PolyCollection. When I use the list<interface*> strategy, I tend to consider the list to be the owner of it's objects (and delete them when the list is deleted). There should be PolyCollection::emplace and the poly instances should "live" in the PolyCollection. Accessing the poly instances should be by [const] reference only (isn't this already the case? what benefit does the copy constructor bring?!?). I didn't examine the source so maybe it's (a) after all, in which case I apologize for wasting precious time, bandwidth, and disk space. d3fault

5 11

boost::process
by Chris Rodgers 05 Oct '17

05 Oct '17

Hi, I'm trying to use Boost 1.65.1's boost::process::child in a Windows console program. At the moment, the child process is launched in the same console window as the parent. Is there an option to make a new console window for the child? The pre-boost-incorporation version of boost::process seemed to allow manipulation of the startupInfo struct directly (e.g. https://stackoverflow.com/questions/43582022/boostprocess-hide-console-on-w…) and then a win32_launch call. But this seems to have been removed now? Any tips gratefully received. Thanks, Chris.

1 0

looking for a memory leak in an asio based application
by David Medine 05 Oct '17

05 Oct '17

I know that the answer to this question is /always/ 'no'. But is it possible, just possible that one can leak memory out of boost::asio? Let me explain. I work on an open source client/server data streaming library for multi-modal data streaming called labstreaminglayer (https://github.com/sccn/labstreaminglayer) At the core of the data transmission architecture is boost::asio. Now, to avoid calling collisions, the original author of the library shipped the source code with a name-mangled version of boost. Over the years this has yet to be updated so we are still relying on boost 1.57. The memory leak starts on the server side after a client is connected to it. This went undetected for years because it is very very slow and it depends on the sampling rate of the data stream. LSL (labstreaminglayer) is widely used in the neurophysiological research community (for whom the tool was intended), but normally in that field you are not sending samples faster than 500 Hz or so. But some devices are now available that offer relatively high sampling rates (e.g. Brain Products' actiCHamp can go as fast as 100kHz). The amount of memory that leaks seems to depend on how often data is sent, not how much data is send at a time. So, when the data is flowing quickly, the leak is severe. I am 99.99% sure that the allocation of memory for all the data I am sending is getting deallocated correctly. I don't see any leaking memory according to Visual Studio's diagnostic tools, nor can Deleaker spot anything. However, when I look at a server process in task manager, I can see the memory rising. I tried the asio::io_service::work trick (https://stackoverflow.com/questions/32258890/c-boost-asio-async-send-to-mem…) on my io_services, but this didn't have any effect. I have not gone through the whole pipeline yet, but I've been at this for two harrowing days already. So I just thought I would ask the list if anyone has seen anything like this before or if there has been any slow-leak fixes in boost (particularly asio) since 1.57 (I didn't see anything in the logs, but maybe I missed something). Due to the state of the project at this moment, it is non-trivial to update the name-mangled boost to another version, but it is far less trivial than the needle-in-a-haystack game I am playing right now. So, if someone can say with authority a boost update will fix everything I'll just do that and be happy. It's a long-shot, I know.

2 1

Bost Logging
by Mário Costa 04 Oct '17

04 Oct '17

Hi, I'm havving a bad time trying to figure out how to add and log custom keywords using boost/log. I havent been able to understand how exactly attribute keywords (BOOST_LOG_ATTRIBUTE_KEYWORD) bind with logging e.g. using BOOST_LOG_STREAM_WITH_PARAMS(logger, (??) (??)). For instance, if I define a keyword line_number, BOOST_LOG_ATTRIBUTE_KEYWORD(line_number, "LineNumber", unsigned int) and add it to the formmater logging::formatter formatter = expr::stream << "[" << line_number << "]" How do I use BOOST_LOG_STREAM_WITH_PARAMS(logger, (??)) Tanking an exemple from the severities, I should use a keyword parameter defining something like BOOST_PARAMETER_KEYWORD(sometag, line_number) then I should be able to use BOOST_LOG_STREAM_WITH_PARAMS(logger, (line_number = 10)) But those things just don add up together. I for sure missing the logics of the logging ... Thanks for any help! Cheers, mc

1 0

[geometry] Get points within certain radius?
by Florian Lindner 03 Oct '17

03 Oct '17

Hello, first of all thanks for the repliest to my previous geometry question! What is the best, especially fastest way to get all points within a certain radius of a point? Using the iterator method with nearest(pt, X) highly depends on X. Since I don't know this number, I have to guess it rather high. for (auto it = rtree.qbegin(bgi::nearest(point, results)) ; it != rtree.qend() ; ++it) { std::ignore = it; break; } results = 10 => time = 3.0s results = 100 => time = 22.6s results = 200 => time = 71.4s results = 300 => time = 153.2s though I always use just one result. The loop is inside another loop, not depending on results, so of course, the times are multiples of the actual times for that query. But clearly, choosing high numbers for results is not feasible. Another idea is to use the within() query with a circle of a radius. But bg seems to have no circle or sphere model. A workaround would be to use a box instead of a sphere. That means I have to test the results again, whether they are actually contained in th sphere. Alternatively, I could use the satisfies predicate. But wouldn't I give up all spatial information? This predicate needs to compare all vertices, don't it? It does include the knowledge, that when one point is discarded, all other points with a greater distance will discarded too. According to https://stackoverflow.com/questions/22909171/boostgeometry-nearest-neighbor…, the way to go seems to be to combine within(boundingBox) and satisfies(). Any other ideas? Thanks, Florian

2 1

[test] Is standalone use possible?
by Adam Nielsen 02 Oct '17

02 Oct '17

Hi all, I'm in the process of trying to remove the Boost dependency from my project, as it is causing too many difficulties for non-developers trying to run my code. This has been mostly solved by switching to C++11 and using the std:: implementations instead, however rather than rewrite all my unit tests I am hoping I can use a standalone version of Boost.Test, and distribute the code alongside my own. Unfortunately I am not quite sure how to extract the Test library out of the main Boost distribution. I have compiled "bcp" and used it to copy "test" into a new folder, but it seems to copy a huge chunk of Boost as well, including a bunch of libraries (system, thread, date_time, etc.) and I end up with 29 MB of source files, which seems wrong (especially considering my own codebase is barely 500 kB). Am I doing something wrong? I was hoping to use the header-only version of Boost.Test and this does seem to work when I #include the system-wide version of Boost (and don't link to any libraries), so it would seem that I don't need all those libraries that bcp copies. Is there any other way to access a slimmed-down standalone version of Boost.Test? Many thanks, Adam.

6 12