- Boost - lists.preview.boost.org

Re: [boost] [asio-users] [http] Formal review of Boost.Http
by Lee Clagett 08 Aug '15

08 Aug '15

Sorry for any duplicates and top posting, but I didn't reply-all, so boost got dropped (which was probably more important). And this likely ruins some mail clients (again sorry!). On Fri, Aug 7, 2015 at 11:10 PM, Lee Clagett <forum(a)leeclagett.com> wrote: > On Fri, Aug 7, 2015 at 6:29 PM, Vinícius dos Santos Oliveira < > vinipsmaker(a)users.sf.net> wrote: > >> 2015-08-07 12:53 GMT-03:00 Niall Douglas <s_sourceforge(a)nedprod.com>: >> >>> I do not believe Http is ready for peer review for these reasons, and >>> I personally would urge you withdraw it until you have fixed these >>> issues and then come back to us. If you do not withdraw it, I will >>> vote for rejection. >>> >> ...[large snip]... >> > >>> * I don't understand why you cannot issue more than one async read at >>> a time nor async write at a time. In something like pipelined HTTP >>> that seems like a very common pattern. I smell a potential design >>> flaw, and while the docs mention a "queue_socket" I can't see such a >>> thing in the reference section. >>> >> >> Too much synchronization and buffering. >> >> The pipeline has the following requests: >> >> A => B => C >> >> If I were to allow replies to request B, the design would be more >> complex, increasing the difficult to implement alternative backends and >> everyone's life. Also, the reply to request B cannot actually be sent while >> the reply to request A is on hold. This means that the whole response to >> request B would need buffering and if response B is a video live stream, >> this means an implicit (the user cannot know) crash after an unfortunate >> memory exhaustion. I couldn't claim a **lightweight** server if this >> excessive buffering was to occur. I couldn't claim the project is >> appropriate for embedded devices if I had chosen the buffering approach. >> It's a trade-off. I should write about it in the "design choice" of the >> documentation. >> >> HTTP/2.0 does have proper multiplexing and it can be more useful to >> respond several requests without "head of line blocking". >> >> A queue socket is a concept, I don't provide one. I can look into how >> make the documentation less confusing by adding a page exclusively >> dedicated to explain the problem it solves. Would that work for you? >> >> > FWIW it should be possible to do client and server side pipelining with > this design right now. The reads and writes of the HTTP Socket concept > should be (and currently are by `basic_socket`) handled independently. > Since the notification for a write completion is based on the lower layer > write completion, and NOT whether a HTTP response was received, multiple > (pipelined) requests should be possible by a client. Similarly, this design > should allow servers to handle pipelined requests in parrallel. Providing a > framework for client or server pipelining will require more work, and > should probably not be a part of the Socket concept. In fact - > > I think the ServerSocket concept should be removed entirely. The > `http::basic_socket<...>` models both concepts, so theres lots of server > specific code in it, which seems confusing to me (is this really a "basic" > socket?). I think standalone functions for typical client and server > operations can be implemented as "composed" operations similar to > `asio::async_write`. The documentation will have to be explicit on what the > composed function is doing so that users know whether the Socket concept > has outstanding reads or writes (and therefore cannot do certain operations > until the handler is invoked). In fact, if there is a server operation that > _cannot_ be done with the Socket concept, then the concept probably needs > to be re-designed. > > Lee >

2 1

Is master open for release 1.59?
by Vicente J. Botet Escriba 08 Aug '15

08 Aug '15

Hi, I was on vacations. Is the master branch open for merge? Best, Vicente

3 4

[Property Tree] Comments in info format
by Gavin McNiff 05 Aug '15

05 Aug '15

First time posting, so hopefully no grave faux pas! I have found the Property Tree info format very useful for config files. However, the in-built comment ";" can be a little limiting. In particular, I miss not being able to comment out an entire block. To this end I have being modifying the relevant header to file to allow ; // and -- as inline comments and /* */ as a block comment. These are familiar to many users. Naturally, with every Boost release I have to re-patch this file, so I thought maybe the info format could be updated to allow comments in these styles. Since the format is not any kind of official format and adding new comment styles should not break existing files, I would hope there should be no major objections. I have taken the liberty of attaching the updated header file, as well as the adapted test file. The _gmcn suffix is just for identification purposes. The changes in the code are marked with !!! GMcN !!!. The changes pass in the built-in tests and my own usage tests. Best regards Gavin McNiff

2 2

boost::asio and I2P connection handling.
by Janusz Zradomia 05 Aug '15

05 Aug '15

Hello. I would like to extend the ASIO (in boost::asio) by adding support for new virtual protocols. For example i2p, it is similar to Tor but it works with equivalent of .onion hidden servers, and is not used to connect to "normal Internet" usually. It works like this: - the end-points, equivalent of the ip4/ip6 address, is a very long string (around 500-1000 characters, it is technically the public key of data) - the end-point address is often expressed as shorter version of the address, that is around 50 characters afair, in two possible formats (b32 and b64) - user can create new i2p listening socket for TCP-like data (for stream of bytes), or for UDP-like data (message oriented). On creation of the listening socket you can get a new own "address" as above. Or, you can on creation/before creation tell it to reuse an previously created address by providing the std::string with private key of it (that you saved previously) - user can connect to an i2p address. When connecting, he is getting a new endpoint from which he connects, or he can reuse old one by loading the private key as above. - there exists naming lookup system, that allows to convert domain-name strings like "zzz.i2p" into proper i2p address Can I add such new class of protocol on top of ASIO? Which classes I need to extend (create child class)? Where exactly would each of the listed above functions go?

1 0

[http] Formal review starts on Friday
by Bjorn Reese 05 Aug '15

05 Aug '15

This is a short notice that the formal review for the Boost.Http library starts on Friday and will last for ten days: from August 7th until August 16th. A proper announcement will be made on Friday, which will include links to the source code and the documentation to be reviewed. As a courtesy to the ASIO community we encourage your participation in the formal review as well.

2 1

[1.59] SVG's not showing up in beta
by John Maddock 05 Aug '15

05 Aug '15

SVG graphics used in documentation are still absent from the (beta) website, see for example: http://www.boost.org/doc/libs/1_59_0_b1/libs/math/doc/html/math_toolkit/sf_… I assume a script somewhere needs updating? Thanks, John.

2 2

Trying out QVM with multiple library implementations
by Andrew Hundt 05 Aug '15

05 Aug '15

I'd really like to try out QVM because I have to switch between a half dozen matrix, vector and quaternion types constantly and would prefer not to write functions for each conversion and operations between them. Examples of the formats I need to handle are: *Eigen <http://eigen.tuxfamily.org/index.php?title=Main_Page>:* Eigen::Affine3d Eigen::Quaterniond + rotation and translation versions *vrep <http://www.coppeliarobotics.com/helpFiles/en/apiFunctions.htm#simBuildMatri…>:* float[3] (vrep vector) float[4] (vrep quaternion) float[12] (vrep homogeneous transform matrix) *cisstVector <https://github.com/jhu-cisst/cisst/tree/master/cisstVector>:* vctDynamicVector <https://github.com/jhu-cisst/cisst/blob/master/cisstVector/vctDynamicVector…> vctQuaternion <https://github.com/jhu-cisst/cisst/blob/master/cisstVector/vctQuaternion.h> vctDynamicMatrix <https://github.com/jhu-cisst/cisst/blob/master/cisstVector/vctDynamicMatrix…> There are also fixed versions of the "Dynamic" objects. + rotation & translations Is there any particular concerns I need to take regarding alignment, ordering (row vs column major) etc? For example, the ordering of the quaternion data components differ between eigen and rep: // 0123 // vrep quat is ordered xyzw, // eigen quat is ordered wxyz Any advice on how to go about this would be appreciated so I can try QVM out! Cheers! Andrew Hundt

2 1

Broken links in 1.58.0 release notes
by Jonathan Wakely 04 Aug '15

04 Aug '15

At http://www.boost.org/users/history/version_1_58_0.html the release notes on the breaking change for Variant say "See get.hpp and polymorphic_get.hpp descriptions for more info" but both hyperlinks are broken.

2 1

Review Request: QVM (generic Quaternion, Vector and Matrix operations library)
by Emil Dotchevski 04 Aug '15

04 Aug '15

Boost QVM defines a set of generic functions and operator overloads for working with quaternions, vectors and matrices of static size. The library also defines vector and matrix data types, however it allows users to introduce their own types by specializing the q_traits, v_traits and m_traits templates. Online documentation and a link to download the source code under the Boost license is available at http://www.revergestudios.com/boost-qvm/. I would like to thank Adam Wulkiewicz for volunteering to be the review manager for QVM and for his valuable initial feedback. Thanks, Emil Dotchevski

13 30

BOOST_NO_CXX11_CONSTEXPR and msvc-14.0
by Peter Dimov 04 Aug '15

04 Aug '15

I see that BOOST_NO_CXX11_CONSTEXPR is no longer being defined for msvc-14.0. Even the RTM version has spotty support for constexpr though; the techniques in my metaprogramming article, for example, don't work. One simple example that fails is #include <type_traits> template< int I > struct X { }; template<class... T> constexpr int f( T... t ) { return 1; } template< class... T > struct Y: X< f(std::is_same<T, void>::value...) > { }; Y<void, void, void> y; int main() { } Another is #include <type_traits> template< class... T > struct X { static constexpr bool v[] = { std::is_same<T, void>::value... }; }; int main() { }

15 32

[serialization] Errors when testing withe mingw-64/gcc on Windows
by Edward Diener 04 Aug '15

04 Aug '15

When running serialization tests with mingw-64/gcc on Windows I get repeated errors of the form: ..\..\../boost/archive/detail/basic_oarchive.hpp:64:5: error: function 'boost::archive::detail::helper_collection& boost::archive::detail::basic_oarchive::get_helper_collection()' definition is marked dllimport get_helper_collection(){ Is this a known problem ? Testing with VC++ does not give any errors.

2 5

Boost 1_59_0_b1_rc1 is available for testing
by Marshall Clow 02 Aug '15

02 Aug '15

Release candidate files for 1.59.0 beta 1 are available at http://boost.cowic.de/rc/ As always, the release managers would appreciate it if you download the candidate of your choice and give building it a try. Please report both success and failure, and anything else that is noteworthy. This helps ensure the candidates build OK before we push them out to SourceForge. The files (and associated md5s) are: MD5 (boost_1_59_0_b1_rc1.7z) = f54a97e5ee7f8e3d04cd9af8396e30f9 MD5 (boost_1_59_0_b1_rc1.tar.bz2) = 9804305aae0c9de9f8cfc02e3de75167 MD5 (boost_1_59_0_b1_rc1.tar.gz) = e584770bd76885c123a0426d24466fe2 MD5 (boost_1_59_0_b1_rc1.zip) = ee34f223aa789a6b3f727cb9480651ed Thanks! -- The release managers

13 14

[Log, Filesystem] Build errors using B2 and Jamfiles distributed with Boost
by Norbert Wenzel 02 Aug '15

02 Aug '15

I'm a lazy guy so I try to use Boost.Build whenever I can. That way I can simply download the Boost tarball, extract it into my project's directory and use the Jamfiles distributed with Boost. Recently I switched to 1.58 and my project setup failed to build, complaining that Filesystem would be built twice with different settings* (actually it would be built one time without any specific settings and one time with a sane set for my platform). I managed to create a "simple" project setup to reproduce this error. I created a library (testlib) with a single extern C function that fills a user provided buffer with the current working directoy. This library uses Boost.Filesystem internally. Therefore the lib's Jamfile also references the Jamfile distributed with Filesystem. I created an executable (test) that calls the lib function and also uses Filesystem to get the parent directory of the current working directory (as reported by the lib function). This setup builds fine as long as I use cout to write the paths in the test exe. When I replace cout with Boost.Log (and therefore also reference the boost_log and boost_log_setup projects) this setup only builds with Boost 1.57. From 1.58 on (also tested with 1.59.b1) the project does not build because Filesystem would be built twice with different settings. Since the error can be triggered by changing the Boost version I reckon this is due to a change in one of the Jamfiles (probably Boost.Log) and not a problem of Boost.Build per se. But for reference I use Boost.Jam Version 2014.03 (the one distributed with 1.58). Since the "simple" setup consists of a few files and directories I've made it available via Github: https://github.com/norbertwenzel/boost-build-problem I've posted this here and not to Boost.Build ML since I suspect this is due to a change in one of the two libraries mentioned and not a bug or misuse in/of Boost.Build. Any help appreciated. thanks, Norbert *) The exact error message I got: error: Name clash for '<p/tmp/b2_problem/external/boost_1_58_0/bin.v2/libs/filesystem/build/gcc-5.1.0/debug>libboost_filesystem.so.1.58.0' error: error: Tried to build the target twice, with property sets having error: these incompatible properties: error: error: - none error: - <abi>sysv <binary-format>elf <log-api>generic <preserve-test-targets>on <python-debugging>off <python>2.7 <testing.execute>on <threadapi>pthread error: error: Please make sure to have consistent requirements for these error: properties everywhere in your project, especially for install error: targets.

2 3

[http] Scheduling a review for Boost.Http
by Bjorn Reese 01 Aug '15

01 Aug '15

I am about to schedule a formal review for Boost.Http (see blincubator for links.) If there are no objections I suggest that the review is held from August 17 to 30.

3 4

Boost 1.53 Libraries for Android
by Ian Calegory 31 Jul '15

31 Jul '15

I've compiled and made available for download the Boost C++ 1.53 Libraries for Android. This enables cross platform code in C++ that uses the Boost libraries. It is especially useful for Embarcadero's FireMonkey environment in RAD Studio, where cross platform code written in C++ can run on Windows, Android, Mac OS X, and iOS. Conceivably, cross platform code can be written using the Boost libraries that can run on all of the aforementioned environments. (The other platform libraries would have to be available, of course, and RAD Studio comes with Boost pre-compiled for Windows; Boost of course is widely available for Mac OS X.) Here is my web page where the include files and .a library files (one file for each of the four main Android architectures) is available for download: http://www.mandelmania.net/boost.html Any comments welcome. logo final small 11 Ian Calegory Software Engineer <mailto:ianc102002@yahoo.com> ianc102002(a)yahoo.com <http://www.mandelmania.net/> http://www.mandelmania.net

1 0

devector feedback request
by Benedek Thaler 30 Jul '15

30 Jul '15

Dear Boost, I'm happy to announce the new devector container developed as part of the GSoC programme reached a presentable state. I invite you to take a look and ask for feedback. Main features: - every std::vector feature - front operations with amortized linear complexity - optional small buffer optimization - customizable growth policy (see the docs) - unsafe methods for the brave and performance addicts Documentation: - Intro: http://erenon.hu/container/container/non_standard_containers.html#container… - Usage: http://erenon.hu/container/container/examples.html#container.examples.devec… - API: http://erenon.hu/container/boost/container/devector.html#idp86138944 Code: - Implementation: https://github.com/erenon/container/blob/devector/include/boost/container/d… - Tests: https://github.com/erenon/container/blob/devector/test/devector_test.cpp Thanks, Benedek

6 11

[Log] Running in a separate thread
by Andrew Hundt 29 Jul '15

29 Jul '15

I really only need some of the functionality of BOOST_LOG_TRIVIAL, perhaps plus loving to a file in addition to the console. However, my only true need is that all the work (or as much as possible) must be done in a separate thread so as not to slow the current thread down. Right now when I use BOOST_LOG_TRIVIAL as is it causes latency in my network communications which have soft realtime requirements so I've had to comment all of my logs out. Is there a straightforward non-intrusive solution to having as much of the work as possible for boost.log be done in a separate thread? Thanks for your help! Cheers! Andrew Hundt

3 4

Boost.DLL formal review result
by Vladimir Prus 28 Jul '15

28 Jul '15

The Boost.DLL library is accepted. Congratulations to Antony, and thanks for everybody who reviewed the library! The votes were as follows: - Ralf Globisch - Yes - Klaim - Joël Lamotte - Yes - Niall Douglas - Yes, conditionally - John P Fletcher - Yes - Bjorn Reese - Yes - Edward Diener - Yes - Rodrigo Madera - Yes - Richard - Yes I'd like to also thank Thomas Trummer, Rob Stewart and Ion Gaztañaga for their contribution to review. The key suggestions from the review are listed below. None is critical enough to become a formal prerequisite, or require a mini-review. The library can be added when Antony feels its ready. Naming ----------- There was a concern that "DLL" might sound windows specific, and search for best terminology. This does not seem all that important, as soon as documentation introduces the terminology it uses, and is consistent. OSX support ------------------- The library was found not to work on OSX; I believe this issue was already fixed off-list with help from Rodrigo. Documentation ---------------------- Several people requested improvements. Bjorn and John, in particular, pointed out specific issues. This seems to be the post important change prior to adding the library. Dependencies on Boost and system headers ---------------------------------------------------------------- It was suggested that the library does not depend on other Boost C++ Libraries. While it might be helpful to some users, I find it logically impossible for a Boost formal review to request that dependencies on Boost libraries be removed. It was also mentioned that Boost.DLL is header-only library and includes system headers that can bring a lot of symbols and maybe macros. I don't think an actual problem was reported, though, and there is no way to avoid the problem while staying header-only. It's up to the author to consider these suggestions and do something, or don't do anything. Continuous integration -------------------------------- There was a recommendation from Niall to improve CI of the library, including static analysis, valgrind, Windows CI, showing CI status in documentation and so forth. These are good suggestions. At the same time, they are not specific to Boost.DLL, and it's not the goal of a formal review to create new generic development process requirements. Many existing libraries of indisputably high quality do not have these requested mechanisms. Therefore, the author can act on these suggestions as he sees fit. Thanks, -- Vladimir Prus http://vladimirprus.com

4 3

BGL pull requests
by Mads Jensen 27 Jul '15

27 Jul '15

Hi, There are many quite trivial pull requests on Github for BGL. Can the project responsible please have a look at them? Most of them just fix typos and add more clarity to the documentation. Thank you, -- Med venlig hilsen / Kind regards, Mads Jensen Gunnery Sergeant Hartman: What's your excuse? Private Cowboy: Sir, excuse for what, sir? Gunnery Sergeant Hartman: I'm asking the fucking questions here, private! Do you understand? -- Full Metal Jacket

6 8

[mpl] Problem with VC++14 and boost::mpl::size_t
by Edward Diener 27 Jul '15

27 Jul '15

The just released VC++14 has a problem with the boost::mpl::size_t template. After preprocessing it sees: namespace boost { namespace mpl { template< std::size_t N > struct size_t { static const std::size_t value = N; typedef size_t type; typedef std::size_t value_type; typedef integral_c_tag tag; typedef boost::mpl::size_t< static_cast<std::size_t>((value + 1)) > next; typedef boost::mpl::size_t< static_cast<std::size_t>((value - 1)) > prior; operator std::size_t() const { return static_cast<std::size_t>(this->value); } }; template< std::size_t N > std::size_t const boost::mpl::size_t< N >::value; }} and gives the error message: size_t.cpp(nnn): error C2872: 'size_t': ambiguous symbol predefined C++ types (compiler internal)(16): note: could be 'unsigned int size_t' size_t.cpp(nnn): note: or 'boost::mpl::size_t' pointing to the line after it encounters 'struct size_t' as far as I can make out ( you need to preprocess the size_t.cpp test, comment out the #line directives and run the test against the preprocessed result to see what the compiler is objecting to ). It therefore fails a number of mpl tests which all prior versions of VC++ pass. It flags this error in 5 mpl tests which use the boost::mpl::size_t integral wrapper, among which the easiest to analyze is the size_t.cpp test itself. Anybody have any idea what is going on here with the compiler, or whether it can be fixed somehow in VC++14 for the mpl code ?

9 21

[config] BOOST_NO_CXX11_HDR_CODECVT is not defined for VC14
by Andrey Semashev 27 Jul '15

27 Jul '15

Hi, I see that BOOST_NO_CXX11_HDR_CODECVT is not defined for MSVC-14, although it is known that the codecvt facets are unusable because of linkin errors: https://connect.microsoft.com/VisualStudio/feedback/details/1348277/link-er… You can also see STL mentioning DevDiv#1060849 here: http://blogs.msdn.com/b/vcblog/archive/2014/11/12/visual-studio-2015-previe… and I can see that the problem is not fixed in RTM. This also concerns std::codecvt specializations for char16_t and char32_t. Can we define the BOOST_NO_CXX11_HDR_CODECVT macro for MSVC?

3 6

[predef] Fails on Intel/win
by John Maddock 27 Jul '15

27 Jul '15

BOOST_COMP_INTEL is set to 0, which means that [ predef-require "BOOST_COMP_INTEL" ] fails. __INTEL_COMPILER is set to 1500. BOOST_COMP_INTEL is set to ( (((0)%100)*10000000) + (((0)%100)*100000) + ((0)%100000) ) Thanks, John.

6 17

__NR_eventfd syscall
by Alexey Brodkin 27 Jul '15

27 Jul '15

Hi, When building Boost for ARC architecture we're seeing this failure: --------------------->8--------------------- ./boost/asio/detail/impl/eventfd_select_interrupter.ipp: In member function 'void boost::asio::detail::eventfd_select_interrupter::open_descriptors()': ./boost/asio/detail/impl/eventfd_select_interrupter.ipp:50:50: error: '__NR_eventfd' was not declared in this scope write_descriptor_ = read_descriptor_ = syscall(__NR_eventfd, 0); --------------------->8--------------------- This happens clearly because ARC following Linux kernel's UAPI doesn't define __NR_eventfd syscall. If you look here https://github.com/torvalds/linux/blob/master/include/uapi/asm-generic/unis… you may notice that "eventfd" syscall with some others resides in "#ifdef__ARCH_WANT_SYSCALL_NO_FLAGS". And from what I may see only "score" architecture has this __ARCH_WANT_SYSCALL_NO_FLAGS defined. Most of other arches that don't use unistd.h from "include/uapi/asm-genereic" folder have "eventfd" syscall defined. For example it's true for ARM (both 32 and 64 versions). So I'd say it's not the best way to use directly a syscall that's being obsoleted in "Boost". And so real fix would be to modify Boost's "eventfd_select_interrupter::open_descriptors()" for example with use of really generic "eventfd2" or even better use libc's eventfd() call instead of direct syscall. Any thoughts what would be the best fix here? -Alexey

1 1

[ublas] [1.5.9.0b1] Merge patch for broken boost::numeric::ublas::shallow_array_adaptor since 1.58
by Hugues Malphettes 24 Jul '15

24 Jul '15

Hi everyone, Does somebody knows how to proceed to merge the patch filed here: https://svn.boost.org/trac/boost/ticket/11207 I am forced to remain on 1.57 because of this issue. At the minimum there are 4 projects affected: myself, Mark Moll who filed the issue on track and nasosi and vh4n who filed the same patch here: https://github.com/uBLAS/ublas/pull/30/files I had convinced myself that because it had been merged in github we would be good to go for 1.59. I just tested 1.59.0b1 and it has not been merged. Many thanks for your help! Hugues

1 0

Test Runner Update to vc14.0 - CTP 6
by Tom Kent 23 Jul '15

23 Jul '15

Just a heads up, I have just updated the test runner teeks99-08f-win2012R2-64on64 from vc14.0 - CTP 4 to vc14.0 - CTP 6. The change should show up in the next run of that runner in a few hours. I think some people were waiting to see if MS was fixing some things, so I thought I'd send a message letting people know these results would be available. Tom

7 22