January 2019 - Boost - lists.preview.boost.org

[asio][coroutine] forbid_suspend blocks
by Vinícius dos Santos Oliveira 07 Aug '19

07 Aug '19

I've implemented a dumb linter using Python + clang API that allows you to mark certain blocks as “atomic in respect to coroutines” using C++11 attributes. Usage example follows: void f() { int yield; g(yield); [[bt::forbid_suspend]] { // Comment this call to make linter accept the code g(yield); t(); } } You can change bt::forbid_suspend to another identifier by tuning settings on the beginning of the script. The prefix forbid was inspired by Rust's linters[1] as you cannot disallow the linter from inside the forbid_suspend block. Some limitations follow: - A suspend point is identified by looking references to any id equals to yield in the last argument passed to function calls. It means that yield[ignored_ec] and alike will be correctly identified, but it needs to be named yield and it needs to be the last argument to the function call. - I haven't found a way to extract attributes related to COMPOUND_STMT nodes with clang API, so I manually look for token gaps and do a simple regex. It means you cannot use macros to [[bt::forbid_suspend]] as the macro won't be expanded. - You better call the script with '-x c++' and also ' -I/usr/lib/clang/5.0.1/include'. Like I've said in the beginning of the email, it's a *dumb* linter, which means it doesn't try to handle all cases. Therefore, I don't hope to do many changes to it or maintain it to work against newer releases of clang and I'm only uploading it to gist (not fullblow Github): < https://gist.github.com/vinipsmaker2/d930fbe5b7597432b021effe618da171>. It should also be easy to hack as it is very small. I'm releasing the linter under public domain. Enjoy. I have been using fibers in a daily basis on the last months of my job to solve networking problems on a gateway project. Some of the technologies I use are: - Boost.Asio and its spawn function. - A few custom synchronization primitives for the fibers (e.g. mutex, semaphore). - Eventually I'll have to write a custom spawn function that returns a joinable handle so I can kill the usage of a fiber::barrier in a few places. I've been using the linter for about a week already and looks fine. My use case are: - To prevent iterator invalidation in regions accessing shared variables. - To guarantee events order in algorithms that handle races (cannot yield while task YYY hasn't finished). The project I use it in is not an open source project, so I cannot talk much more about it. Also, I don't claim my solution is original. I wouldn't be surprised if the very same solution exists in the wild (please let me know if you know of a similar approach). This is not a new problem. It is know for some time. For instance, in P0171, Gor Nishanov noted this same problem: “In coroutines the suspend point is clearly marked with await, which tells the reader that something unusual happens in this function and allows the reader, for example, to confirm whether the lifetimes of the objects of interest align with the lifetime of the coroutine or not, whether some locks need to be acquired to protect some concurrently accessed data, and whether some locks need to be released before the execution reaches the suspend point.” — <http://open-std.org/jtc1/sc22/wg21/docs/papers/2015/p0171r0.html>, 2015 However, the solution he presented was to manually examine whole blocks of code without help from the compiler. I argue that the solution he presented is as error-prone as manually looking for exit paths that don't free acquired resources in languages lacking RAII. And I've been spoiled too much by RAII. Furthermore, I take this comment of his with highly suspicious look: “Since we are on the subject of maintenance nightmares, we would like to offer a conjecture that the absence of the await in P0114 is a likely source of many maintenance nightmares. Without a syntactic marker to signal to the person reading the code that something funny is going on, it is impossible to tell whether the following code is correct or not” — http://open-std.org/jtc1/sc22/wg21/docs/papers/2015/p0171r0.html If I had to give a vote to the proposal he was defending or Chris' proposal, I'd probably have chosen Chris'. [1] https://doc.rust-lang.org/reference/attributes.html# lint-check-attributes -- Vinícius dos Santos Oliveira https://vinipsmaker.github.io/

2 3

Non-ASCII Path Names in Repo
by Tom Kent 25 May '19

25 May '19

In the last few days we've had a non-ASCII path name added to the repo: tools/boostdep/test/utf8-test-ßµ™∃ Which came with commit 8893d1c57951a [1] I noticed this because a python 2 script (calling shutil.copytree) I had running on windows that copied the repo around started to fail. I've long needed to update that script to python 3, and am working on this. However, I wanted to point this out to the list so we could discuss if there are any other ramifications of having a non-ascii path. I vaguely remember some discussion about something like this in the past, but couldn't find it via search. If this is already a settled topic, that's great. Tom [1] https://github.com/boostorg/boostdep/commit/8893d1c57951a4a83705ed203005576…

2 1

Re: [boost] [ Interest? ] [ out_ptr ] Tiny C++ Abstraction for C-Style Output Pointers
by Bjorn Reese 12 May '19

12 May '19

On 06/28/18 21:03, ThePhD via Boost wrote: > out_ptr is a small abstraction for making types such as > std::unique_ptr, std::shared_ptr, boost::shared_ptr, and the upcoming > std::retain_ptr > <http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0468r0.html> work > work C-Style output functions as a parameter, addressing the need of > wanting to be able to do "my_c_api_init( foo, bar, &my_smart_ptr );". It's Does out_ptr have other use-cases than C interoperability?

7 14

[thread] compiler error gcc 7.3 on cygwin
by James E. King III 06 May '19

06 May '19

Getting an internal compiler error with gcc 7.3 but only on cygwin 64-bit. Jim@pulsar /cygdrive/c/boost/libs/thread $ ../../b2 --abbreviate-paths toolset=gcc cxxstd=11 define=_POSIX_C_SOURCE=200112L define=__USE_ISOC99 threadapi=pthread address-model=64 variant=release -q ... gcc.compile.c++ ../../bin.v2/libs/thread/test/test_10963_c.test/gcc-7.3.0/rls/cxstd-11-iso/thrd-mlt/vsblt-hdn/test_10963_c.o In file included from ../../boost/thread/detail/invoker.hpp:30:0, from ../../boost/thread/future.hpp:31, from ../../libs/thread/test/test_10963.cpp:13: ../../boost/thread/detail/invoke.hpp: In function ‘decltype (*(forward<A0>)(boost::detail::invoke::a0).*boost::detail::invoke::f) boost::detail::invoke(Fp&&, A0&&) [with Fp = void (boost::executors::basic_thread_pool::*)(); A0 = boost::executors::basic_thread_pool*]’: ../../boost/thread/detail/invoke.hpp:102:43: internal compiler error: in gimplify_expr, at gimplify.c:12006 return (*boost::forward<A0>(a0)).*f; ^ Please submit a full bug report, with preprocessed source if appropriate. See <https://gcc.gnu.org/bugs/> for instructions. "g++" -std=c++11 -fvisibility-inlines-hidden -m64 -mthreads -O3 -finline-functions -Wno-inline -Wall -pedantic -fvisibility=hidden -Wextra -Wno-long-long -Wno-unused-parameter -Wunused-function -pedantic -DBOOST_ALL_NO_LIB=1 -DBOOST_THREAD_THROW_IF_PRECONDITION_NOT_SATISFIED -DNDEBUG -D_POSIX_C_SOURCE=200112L -D__USE_ISOC99 -I"../.." -c -o "../../bin.v2/libs/thread/test/test_10963_c.test/gcc-7.3.0/rls/cxstd-11-iso/thrd-mlt/vsblt-hdn/test_10963_c.o" "../../libs/thread/test/test_10963.cpp" ...failed gcc.compile.c++ ../../bin.v2/libs/thread/test/test_10963_c.test/gcc-7.3.0/rls/cxstd-11-iso/thrd-mlt/vsblt-hdn/test_10963_c.o... If I change the C++ language level to 03 the problem goes away. I also cannot reproduce it on linux with the same version of compiler (but it probably isn't the exact same compiler code). Should I go ahead and file that gcc bug? - Jim

2 1

[test] unit test command-line handling
by Stefan Seefeld 22 Mar '19

22 Mar '19

Hello, this is a question concerning the Boost.Test framework. I'd like to pass command-line arguments to my tests, and I'm unsure of how to proceed. I have found https://www.boost.org/doc/libs/1_68_0/libs/test/doc/html/boost_test/adv_sce…, which shows how to access the command-line programmatically, but that's declared obsolete, and it's also lacking some important info, such as how I should handle the argv vector. Are there any examples that demonstrate this use case ? Specifically: * What (non-deprecated) API can I use to access command-line arguments during test initialization ? * How does my own handling of command-line arguments interact with Boost.Test's own command-line argument handling ? Thanks, Stefan -- ...ich hab' noch einen Koffer in Berlin...

3 8

Integer sorting
by Fenil Mehta 19 Feb '19

19 Feb '19

I have written a sorting algorithm which is way faster than std::sort for all array size. The link is as follows: https://github.com/fenilgmehta/Fastest-Integer-Sort Some guidance would be appreciated on how to improve the project structure and the steps to get it included in boost. Regards, Fenil Mehta

11 24

Feedback desired for reflective enum library
by Nir Friedman 08 Feb '19

08 Feb '19

Hey, I have a reflective enum library available here: https://github.com/quicknir/wise_enum. I've been working on this on and off for a couple of months. I've added features, dealt with various issues, added docs, expanded support through 11-17 and as idiomatically as possible. Overall I think the library is in nice shape (lacks unit tests though, though I do already have a boost license for the code). I feel like a reflective enum is just a very standard requirement in pretty much every codebase I've ever worked on. Simply to be able to reasonably log enums already makes it mandatory. It's also invaluable in configuration, error message alerting, and many other places. There is nothing for this in boost and actually there isn't really much of a standard, standalone solution anyway. The closest is Better Enums ( http://aantron.github.io/better-enums/), but this has some major issues (it doesn't actually create enums but rather enum-like classes, it targets 03). So I think a library like this is needed and my implementation ticks off the major requirements. Feedback on the overall design, any issues, any key features it fails to meet, would be very useful. The library itself is very small and simple so its easy to review. Once it seems like there's some consensus on the core design I'm willing to do the work involved getting it into boost (naming conventions/requirements, directory structure, tests, boost-style docs, etc). Note I'm not on the boost mailing list so please respond directly to this email so I can see! Cheers, Nir

3 4

GSoC 2019
by David Bellot 07 Feb '19

07 Feb '19

Hi Boost community, this year again we applied to the GSoC programm, Google Summer of Code. We hope to be as successfull as usual. I'm taking care of the admin work again this year. I completed an application for Boost on the Google website. It is somehow similar to our successfull applications from previous years. What we need from the community now is the following: 1) if you want to be a backup admin (I will need maybe one or two extra names at most), please let me know and I'll send you an invite. 2) if you wish to be a mentor, please let us know now and if possible on this mailing list so that students will be aware of it too. Don't forget to put the [gsoc2019] tag in the header of your email. You can write your proposals on this wiki: https://github.com/boostorg/boost/wiki/Google-Summer-of-Code:-2019 Ideally, we want to see many proposals before beginning of next week. 3) if you are a student and wants to participate, it's about time to speak loudly on the mailing list to pair with a potential mentor. The sooner the better. You can find ideas from previous years here: https://github.com/boostorg/boost/wiki/Google-Summer-of-Code%3A-Overview Happy coding. David

7 7

find_package(Boost) support
by Peter Dimov 04 Feb '19

04 Feb '19

The current master and develop branches now install the necessary BoostConfig.cmake file (in $prefix/lib/cmake) such that f.ex. the CMake command find_package(Boost REQUIRED filesystem) will find Boost 1.70 using the config file, rather than via FindBoost.cmake (*) (if the Boost installation prefix is in CMake's prefix search path.) Interested parties are invited to test this support and share their experiences. Instead of a Git checkout, you can use the 1.70 release snapshots from https://dl.bintray.com/boostorg/master/ or https://dl.bintray.com/boostorg/develop/, as those better represent what the 1.70 release will contain. (*) Actually, the above command will invoke FindBoost.cmake, but FindBoost will look for a config file as a first step and use that, if found. To disable FindBoost completely, use find_package(Boost CONFIG REQUIRED filesystem) Or, on Windows, assuming the default C:\Boost installation, find_package(Boost REQUIRED filesystem PATHS C:/Boost)

4 4

[thread] Passing function with parameters to boost::barrier
by André Offringa 01 Feb '19

01 Feb '19

Hi Boost developers, I'm having a problem with passing a completion function to a boost::barrier, causing a segfault, and I suspect it is a bug in the barrier code. The following is a minimal example that shows the problem: #include <boost/thread/barrier.hpp> void func(int) { } int main() { boost::barrier b(1, std::bind(func, 42)); b.wait(); } Running inside gdb and doing a bt shows some infinite recursion happening, starting with: #0 0x000055555555e41a in boost::thread_detail::decay_copy<boost::detail::nullary_function<void ()>&>(boost::detail::nullary_function<void ()>&) ( t=<error reading variable: Cannot access memory at address 0x7fffff7feff0>) at /usr/include/boost/thread/detail/move.hpp:355 #1 0x000055555555d814 in boost::detail::nullary_function<void ()>::nullary_function<boost::detail::nullary_function<void ()>&>(boost::detail::nullary_function<void ()>&) ( this=0x7fffff7ff0c0, f=...) at /usr/include/boost/thread/detail/nullary_function.hpp:81 ... When I change "std::bind(func, 42)" to just "func" (and declare func without parameters), it does work and the program exits normally. Passing a lambda to barrier (e.g. barrier(1, [&]() { func(0); } ) ) causes the same segfault, and the same holds when I change std::bind to boost::bind. I'm using Boost 1.67 on Debian Linux, amd64, gcc 8.2.0. Is this a bug? Cheers, André

2 1