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

[thread][build] develop terminally broken?
by John Maddock 07 Oct '17

07 Oct '17

All CI tests are currently failing with: ./b2 headers /home/travis/build/boostorg/boost-root/tools/build/src/build/feature.jam:140: in feature from module feature error: feature already defined: error: in feature declaration: error: feature "threadapi" : "win32" "pthread" : "optional" "propagated" /home/travis/build/boostorg/boost-root/libs/thread/build/threadapi.jam:4: in load from module threadapi /home/travis/build/boostorg/boost-root/tools/build/src/kernel/modules.jam:295: in modules.import from module modules /home/travis/build/boostorg/boost-root/tools/build/src/build/project.jam:1069: in import from module Jamfile</home/travis/build/boostorg/boost-root/libs/thread/build> libs/thread/build/Jamfile.v2:39: in modules.load from module Jamfile</home/travis/build/boostorg/boost-root/libs/thread/build> /home/travis/build/boostorg/boost-root/tools/build/src/build/project.jam:325: in load-jamfile from module project /home/travis/build/boostorg/boost-root/tools/build/src/build/project.jam:64: in load from module project /home/travis/build/boostorg/boost-root/tools/build/src/build/project.jam:89: in load-used-projects from module project /home/travis/build/boostorg/boost-root/tools/build/src/build/project.jam:75: in load from module project /home/travis/build/boostorg/boost-root/tools/build/src/build/project.jam:145: in project.find from module project /home/travis/build/boostorg/boost-root/tools/build/src/build-system.jam:535: in load from module build-system /home/travis/build/boostorg/boost-root/tools/build/src/kernel/modules.jam:295: in import from module modules /home/travis/build/boostorg/boost-root/tools/build/src/kernel/bootstrap.jam:139: in boost-build from module /home/travis/build/boostorg/boost-root/boost-build.jam:17: in module scope from module Any ideas? Thanks, John. --- This email has been checked for viruses by AVG. http://www.avg.com

4 5

[contract] auto_..._error regression tests fail on C++1z
by Lorenzo Caminiti 04 Oct '17

04 Oct '17

Hello all, I asked this question on compl.c++ list as well, given this is more of a general C++1z question than a Boost question. However, this is making some Boost.Contract regression tests fail on C++1z compilers so I wanted to ask it here as well hoping to get some info more quickly... Following code prevents copies outside its friend function f: // File: 05.cpp struct x { x() {} private: x(x const&) {} x& operator=(x const&) { return *this; } friend x f(); }; x f() { return x(); } int main() { auto xx = f(); return 0; } Correctly it does not compile up to C++1y because main cannot access the copy operations: $ clang++ -std=c++1y 05.cpp 05.cpp:14:15: error: calling a private constructor of class 'x' auto xx = f(); ^ 05.cpp:5:5: note: declared private here x(x const&) {} ^ 1 error generated. However, it compiles on C++1z?! $ clang++ -std=c++1z 05.cpp # Not compiler errors?! Why... what changed in C++1z that makes the above compile? How can I prevent copying x outside its friend function f in C++1z? Thanks, --Lorenzo

6 7

[outcome v2] How important is it that an Expected implementation is also provided?
by Niall Douglas 04 Oct '17

04 Oct '17

Dear list, I know it's been a fair few months since any news on Outcome v2. Ordinarily I'd be speaking at CppCon happening right now, but I instead chose Jens' Meeting C++ in November where I'll be speaking on Expected. Outcome v2 has undergone its period of maturation since it was completely rewritten to reflect peer review feedback last May. The new consensus design has fared very well after being put under heavy production use these last few months, and I expect it to return here for a second review before the end of this year. Outcome v2 will also be appearing in the WG21 October mailing, watch out for: * P0762R0: Concerns about `expected<T, E>` from the Boost.Outcome peer review * P0779R0: Proposing `operator try()` Outcome v2 currently implements the not published yet D0323R3 Expected proposal as part of its test suite as a refinement of what boost-dev thought Expected ought to be instead of what it is (ahem!). We run it against the Expected reference test suite, and it passes. So here's the question before any second peer review. Should Outcome v2 provide: 1. No Expected implementation at all, not even under a different name. Not its place. 2. The Expected implementation which the consensus on boost-dev felt the WG21 Expected ought to be (this is the current situation, and no we don't call it `expected<T, E>`). 3. A closely conforming, but not exactly conforming [1], implementation of WG21 Expected. Options 2 and 3 can be combined. But be aware that Peter rolled together an even more conforming Expected implementation using his variant2 library. I had been assuming people would be better advised to use his likely exactly conforming implementation, rather than Outcome v2's not exactly conforming implementation. Less surprise. Thoughts? Niall [1]: Known differences from P0323R3 `expected<T, E>` in this implementation: - `T` and `E` cannot be the same type. - No variant storage is implemented (note that technically speaking, D0323R3 does not require variant storage). -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/

3 5

Boost CMake update
by paul 04 Oct '17

04 Oct '17

Hi all, So I have been working at converting more of boost tests to cmake utilizing the BCM library, here: https://github.com/boost-cmake/boost/tree/1.64 This is a build of Boost 1.64, with a majority of the tests integrated. There is still much more work needed to update the tests, as there are some that still fail. Also, some of the libraries were skipped due to missing headers in the 1.64 tag. You can check the library yourself by running cmake: cd build mkdir build cmake .. cmake --build . --target check This will build all the tests. You can check the tests for one target with: cmake --build . --target check-boost_core Which will run the tests for just `boost_core`. This was all generated using scripts in https://github.com/pfultz2/boost-cmake I plan to update this to a newer version of boost. If anyone has fixes or changes they can send a PR or issue on either repos, and I will try to incorporate them in. Thanks, Paul .

8 35

[beast] Supporting ICY 200 OK
by Vinnie Falco 04 Oct '17

04 Oct '17

Bjorn brought up the point that there are some niche cases where applications do not strictly follow the HTTP message syntax for responses. Specifically, for Shoutcast(?) / Icecast mp3 servers the status-line in the response looks something like this: ICY 200 OK\r\n I have thought about ways to support this directly in Beast but I have decided against it. "ICY 200" is not HTTP according to the RFC, and trying to accommodate this use case places needless technical debt in public interfaces. However, I came up with a very simple technique to make this work with Beast by writing just a tiny bit of extra code. The solution is to implement a stream wrapper which meets the requirements of SyncReadStream[1] and/or AsyncReadStream[2] and forwards the calls to read to the "next layer" object, which will be the actual socket or ssl::stream. The implementation of the icy_stream wrapper can check to see if the first 3 bytes received from the next layer match "ICY" and if so, return a buffer to the caller which replaces those characters with "HTTP/1.1" (or "HTTP/1.0"). With this wrapper, calls to beast::http::read or beast::http::async_read can parse the custom response as a regular HTTP message. The wrapper can embed state information to inform the caller if the "ICY" string was not received. The declaration for such a wrapper might look like this: template<class NextLayer> class icy_stream; Usage for the wrapper: asio::ip::tcp::socket sock{ios}; icy_stream<asio::ip::tcp::socket&> stream{sock}; beast::http::response<beast::http::empty_body> res; beast::multi_buffer buffer; beast::http::read(stream, buffer, res); This technique may of course be generalized to support any desired similar but non-standard HTTP requests or responses in a fashion that is completely transparent to Beast. Thanks [1] <http://www.boost.org/doc/libs/1_65_1/doc/html/boost_asio/reference/SyncRead…> [2] <http://www.boost.org/doc/libs/1_65_1/doc/html/boost_asio/reference/AsyncRea…>

2 4

Re: [boost] [beast] A Networking-TS version of Boost.Asio and Boost.Beast is here!
by Vinnie Falco 04 Oct '17

04 Oct '17

On Tue, Oct 3, 2017 at 9:40 AM, Vinnie Falco <vinnie.falco(a)gmail.com> wrote: > My preview of Boost.Asio version 1.66.0 is here: > > <https://github.com/vinniefalco/boost.asio/commits/net-ts> Upon re-reading the message I realize that it might be confusing, so to clarify - this is NOT an "official" version of Boost.Asio it is just an experimental branch of what Boost.Asio might look like soon, for anyone who wants to try out the latest stand-alone Asio features in the Boost.Asio context. Thanks

2 5

Boost.Buffers, a new library from the makers of Beast!
by Vinnie Falco 03 Oct '17

03 Oct '17

During the formal review of Beast, desires were expressed to use components like the HTTP parser and serializer independently from the rest of the library. The sticking point was the abstraction used to represent a range of non-owning buffers. In particular, Asio imposes certain concrete type requirements in its ConstBufferSequence and MutableBufferSequence concepts. It would seem there is significant interest in making some components of Beast modular, in their own libraries. The main feature of these components would be that they should not require Boost.Asio. Unfortunately, the way Asio's concepts are defined this is not possible. Two solutions present themselves: 1. Create a new buffer sequence concept which retains all the features of Asio's concepts but does not require the concrete types. While this is possible it would create a significant complication. Interfaces which use the new concept would require some kind of adapter at every call site that tries to use a traditional Asio style interface. 2. Create a new library which uses conditional compilation to either include just the Asio header files which fulfill the needs for interacting with buffer sequence concepts, or provide the declarations directly from a copy of the Asio sources. The solution used in Boost.Buffers (a new library I am considering proposing) is choice number two. Someone who wants to create a new library ("Foo") that does no network I/O but wants to interoperate with Asio buffer concepts merely need to do the following #include <boost/buffers.asio.hpp> and then, in their build script for the tests and examples, set: #define BOOST_NET_BUFFER_NO_ASIO 1 This works because Boost.Buffers contains copies of all the necessary header files to work with buffer sequences, and JUST those header files. All of Boost.Asio is not necessary: <https://github.com/vinniefalco/buffers/blob/30ef7031ec0909972a720c0cdd8d6c6…> Now if someone wants to use library Foo they will also be getting a dependency on Boost.Buffers. If that user doesn't need Asio, then they will also set the macro to 1 in their build script. However. if that user wants to use Asio in their program or library, they just set #define BOOST_NET_BUFFER_NO_ASIO 0 and now Boost.Buffers will simply include the relevant Asio header files. This library will allow the creation of new libraries which use Asio buffer concepts, but do not require all of Boost.Asio as a dependency, as well as all the things that Boost.Asio is dependent on, which is non-trivial as can be seen here: <https://github.com/boostorg/beast/blob/develop/.travis.yml#L140> I would like feedback on this library before I move forward with it, as you can imagine it has some controversial aspects to it. Boost.Buffers is here: <https://github.com/vinniefalco/buffers> Thanks

1 0

[beast] The Networking-TS version of Boost.Asio and Boost.Beast is here!
by Vinnie Falco 03 Oct '17

03 Oct '17

I have taken the tip of standalone Asio, from: <https://github.com/chriskohlhoff/asio/commits/asio-1-10-branch> and applied this "boostify.pl" script: <https://github.com/chriskohlhoff/asio/blob/asio-1-10-branch/asio/boostify.pl> plus a bit of manual fixing, to produce a new branch of Boost.Asio which is up-to-date with respect to the stand-alone Asio. This means you can have access to the Networking-TS changes that will be introduced in Boost.Asio (according to the author, this will be in Boost 1.66.0). My preview of Boost.Asio version 1.66.0 is here: <https://github.com/vinniefalco/boost.asio/commits/net-ts> Beast has been ported to this version of Asio, you can try it here: <https://github.com/vinniefalco/beast/commits/net-ts> When Boost.Asio is updated for 1.66.0 to bring it in-line with the standalone Asio (which supports Networking-TS), Beast will simultaneously be updated. Hopefully this will all be done in time for 1.66.0, so that users of the debut version of Beast will have the latest interfaces available for their use. Thanks!

1 0

Modularizing explicit-failures-markup.xml
by Daniel James 03 Oct '17

03 Oct '17

Hi, I'm looking into implementing a mechanism for updating explicit-failures-markup.xml without updating the super project. My basic idea is that modules can have a 'status/explicit-failures-markup.xml' file, and I'll write a script to update the super project file from modules. It will have the same format as the super project file, so can just copy and paste the relevant parts into your module. Does that sound okay? It might be better if the regression scripts handled this themselves, but I'm not sure if anyone would be available to implement that. Daniel

6 12

[asio] Verify SSL certificates using the OS-specific certificate store
by Vinnie Falco 01 Oct '17

01 Oct '17

My understanding of boost::asio::ssl::context is that it is just a small wrapper around OpenSSL. You can set default verify paths, but that assumes the certificates are in a particular directory which they almost never are. For example they might be in some OS-specific type of database which need system calls to access. If a program wants to use the native operating system facilities for verifying certificates, then I believe significant additional code is needed. Is this correct? I am interested in writing a simple function object which will validate a hostname and its accompanying certificate against the operating-system-dependent certificate authorities. Is there some code somewhere that does this? Any resources that I can be pointed to would be of immense value. Thanks

2 1